Doctor of Philosophy(PhD) / Snapshot Isolation (SI) is a multiversion concurrency control that has been implemented by open source and commercial database systems such as PostgreSQL and Oracle. The main feature of SI is that a read operation does not block a write operation and vice versa, which allows higher degree of concurrency than traditional two-phase locking. SI prevents many anomalies that appear in other isolation levels, but it still can result in non-serializable execution, in which database integrity constraints can be violated. Several techniques have been proposed to ensure serializable execution with engines running SI; these techniques are based on modifying the applications by introducing conflicting SQL statements. However, with each of these techniques the DBA has to make a difficult choice among possible transactions to modify. This thesis helps the DBA’s to choose between these different techniques and choices by understanding how the choices affect system performance. It also proposes a novel technique called ’External Lock Manager’ (ELM) which introduces conflicts in a separate lock-manager object so that every execution will be serializable. We build a prototype system for ELM and we run experiments to demonstrate the robustness of the new technique compare to the previous techniques. Experiments show that modifying the application code for some transactions has a high impact on performance for some choices, which makes it very hard for DBA’s to choose wisely. However, ELM has peak performance which is similar to SI, no matter which transactions are chosen for modification. Thus we say that ELM is a robust technique for ensure serializable execution.

Doctor of Philosophy(PhD) / Snapshot Isolation (SI) is a multiversion concurrency control that has been implemented by open source and commercial database systems such as PostgreSQL and Oracle. The main feature of SI is that a read operation does not block a write operation and vice versa, which allows higher degree of concurrency than traditional two-phase locking. SI prevents many anomalies that appear in other isolation levels, but it still can result in non-serializable execution, in which database integrity constraints can be violated. Several techniques have been proposed to ensure serializable execution with engines running SI; these techniques are based on modifying the applications by introducing conflicting SQL statements. However, with each of these techniques the DBA has to make a difficult choice among possible transactions to modify. This thesis helps the DBA’s to choose between these different techniques and choices by understanding how the choices affect system performance. It also proposes a novel technique called ’External Lock Manager’ (ELM) which introduces conflicts in a separate lock-manager object so that every execution will be serializable. We build a prototype system for ELM and we run experiments to demonstrate the robustness of the new technique compare to the previous techniques. Experiments show that modifying the application code for some transactions has a high impact on performance for some choices, which makes it very hard for DBA’s to choose wisely. However, ELM has peak performance which is similar to SI, no matter which transactions are chosen for modification. Thus we say that ELM is a robust technique for ensure serializable execution.

ALe travail réalisé dans le cadre de cette thèse en co-tutelle est à cheval entre deux domaines d’étude de l’interférence de sillage d’obstacle non profilé (soit un cylindre ou deux cylindres en tandem) et celle des zones de recirculations dans un sluice artisanal (un appareil de tri d’or). Notre étude contribue à l’étude des écoulements à surface libre par combinaison de quelques méthodes expérimentales. Nous avons couplé quelques techniques entre autres la visualisation de la topologie du sillage proche par la caméra CCD embarquée, les mesures des champs de vitesse par PIV et des efforts hydrodynamiques. Ainsi, le cylindre animé d’un mouvement uniformément accéléré/décéléré est caractérisé par les forces de traînée et de portance, les enveloppes des maximas, la longueur de le zone de recirculation et le nombre de Strouhal local. Dont ce dernier a une valeur avoisinant de 0.4 près de la surface libre. Et entre deux pics de la portance, le déferlement de la vague d’accompagnement est observable. L’interférence entre deux structures est simulée à travers deux cylindres et que sur une configuration symétrique, le paramètre de proximité B a tendance à accroître la longueur de la zone de recirculation. Les méthodes et dispositifs expérimentales décrites ci-dessus sont appliqués dans le cas d’écoulement hydrodynamique turbulent au sein du sluice. Les essais sur sites aurifères à Madagascar ont permis la collecte des informations sur les paramètres optimums du tri. Ces paramètres sont pris comme base initiale des travaux de laboratoire entamés sur une maquette à l’échelle 1. L’acquisition par PIV suivi des traitements statistique à multi-variables POD Snapshot nous a permis de créer un modèle expérimental de l’écoulement composé de 4 zones distinctes. L’étude expérimentale a été complétée par une simulation numérique par ANSYS14.5 ; qui nous a permis de conclure que l’approchement des riffles entraîne une diminution de la zone favorable du dépôt des minerais lourds. / Two major areas are focused in the thesis. The first investigates the interference due to a wake-pattern of non-profiled obstacles (such as one or two cylinder in tandem). The second characterizes the zone of recirculation inside an artisanal sluice-boxe, which is a device for gold extraction. The work contributes to the study of the flow of a free-surface using experimental methods. Also, a few approaches is combined for the investigation. In this regard, the near-wake pattern of the flow is examined using an embedded CCD camera, correlated to a PIV measurement of the velocity fields and the hydrodynamic forces. We found that a cylinder of uniform motion, accelerated or decelerated, is characterized by: the drag and the lift forces, the envelopes of maxima, the length of the recirculation zone and the local Strouhal number. The Strouhal number approaches the value of 0.4 next to a free-surface and a breaking wave is observed in-between two peaks of the lift force. We modeled the interference between two structures using a succession of two cylinders. The proximity parameter B tends to an increase of the recirculation length for a symmetrical arrangement.Our methods and the experimental procedures are applied for the examination of a turbulent hydrodynamic flow inside a sluice. Tests were performed on gold sites in Madagascar. This permitted the collection of information concerning the optimum parameters for an extraction. The obtained values form the basis of our laboratory work and are applied to a scaled model for validation. An experimental flow model, made of four distinct zones, is derived from a monitored PIV data and a statistical analysis of a multi-variable POD snapshot. The results are validated through simulations using the package ANSYS 14.5. The investigation shows that a closer riffles reduces the region for heavy minerals deposition.

This thesis deals with comparison of rapid and conventional methods used in MRI (Magnetic Resonance Imaging). There is a description of imaging methods such as EPI (Echo Planar Imaging), Ultra-fast GRE, FSE (Fast spin echo) as well as a snapshot-FLASH and FISP (Fast Imaging with Steady Precession). Experimental part of this thesis deals with the rapid FSE (Fast Spin Echo) method. Especially is explained and assembled an algorithm for proper compilation of data from the FSE method. This algorithm allows us to evaluate the images from the FSE method. This method is examined in detail (in terms of impact parameters) and compared with traditional conventional methods. Finally, the individual images are evaluated, and the best parameters for FSE method are identified.

Les bases de données présentent des problèmes de passage à l’échelle. Ceci est principalement dû à la compétition pour les ressources et au coût du contrôle de la concurrence. Une alternative consiste à centraliser les écritures afin d’éviter les conflits. Cependant, cette solution ne présente des performances satisfaisantes que pour les applications effectuant majoritairement des lectures. Une autre solution est d’affaiblir les propriétés transactionnelles mais cela complexifie le travail des développeurs d’applications. Notre solution, Gargamel, répartie les transactions effectuant des écritures sur différentes répliques de la base de données tout en gardant de fortes propriétés transactionnelles. Toutes les répliques de la base de donnée s’exécutent séquentiellement, à plein débit; la synchronisation entre les répliques reste minime. Les évaluations effectuées avec notre prototype montrent que Gargamel permet d’améliorer le temps de réponse et la charge d’un ordre de grandeur quand la compétition est forte (systèmes très chargés avec ressources limitées) et que dans les autres cas le ralentissement est négligeable. / Databases often scale poorly in distributed configurations, due to the cost of concurrency control and to resource contention. The alternative of centralizing writes works well only for read-intensive workloads, whereas weakening transactional properties is problematic for application developers. Our solution spreads non-conflicting update transactions to different replicas, but still provides strong transactional guarantees. In effect, Gargamel partitions the database dynamically according to the update workload. Each database replica runs sequentially, at full bandwidth; mutual synchronisation between replicas remains minimal. Our prototype show that Gargamel improves both response time and load by an order of magnitude when contention is high (highly loaded system with bounded resources), and that otherwise slow-down is negligible.

Predicting student performance is an important topic in academia, especially so in programming context, where identification of struggling students allows teachers to offer early and continuous assistance to help them improve their performance. It is thus essential to analyze student programming behavior to detect those at-risk students. This thesis uses data generated from 220 students in a master’s level programming course at a large European university. The students run unit tests in order to test their code when solving assignments, with a snapshot being taken of each test as it is executed. Unit testing is a method of testing software where individual units of source code are tested for correctness. A data set with simple features is derived from a database of snapshots and labeled with students’ grades. Then, the machine learning models support vector machine (SVM), naive Bayes (NB), random forest, and neural networks with one, two and three hidden layers each are trained, evaluated and performance is compared. The results show that SVM and neural networks models are likely the best performing all-rounders, with a possible naive Bayes selection depending on what goal one has. The thesis contributes by training machine learning models on students’ programming behavior. By arming teacher with models such as these, more students that need assistance can get in-time support and thus improve their performance. Future work can improve the models by using or combining other types of student data as features or use a larger data set. / Att förutsäga studenters prestationer är ett viktigt ämne inom akademin, särskilt i programmeringssammanhang, där identifiering av studenter som kämpar med sina studier gör det möjligt för lärare att erbjuda tidig och kontinuerlig hjälp för att hjälpa dem att förbättra sina prestationer. Det är därför viktigt att analysera studenternas programmeringsbeteende för att upptäcka dessa studenter som är vid risk. Denna uppsats använder data från 220 studenter i en programmeringskurs på masternivå vid ett stort europeiskt universitet. Studenterna kör enhetstester för att testa sin kod när de löser uppgifter, och en snapshot tas av varje test när det körs. Enhetstestning är en metod för att testa programvara där enskilda enheter av källkoden testas för korrekthet. En datamängd med enkla features härleds från en databas med snapshots och märks med studenternas betyg. Därefter tränas och utvärderas maskininlärningsmodellerna support vector machine (SVM), naive Bayes (NB), random forest och neurala nätverk med ett, två och tre dolda lager vardera och deras prestanda jämförs. Resultaten visar att SVM och neurala nätverk sannolikt är de bäst presterande allroundmodellerna, med ett möjligt naivt Bayes-val beroende på vilket mål man har. Uppsatsen bidrar genom att träna maskininlärningsmodeller på studenters programmeringsbeteende. Genom att utrusta lärare med modeller som dessa kan fler studenter som behöver hjälp få stöd i tid och därmed förbättra sina prestationer. Framtida arbete kan förbättra modellerna genom att använda eller kombinera andra typer av studentdata som features eller använda en större datamängd.

Optical coherence tomography systems are used to image the retina in 3D to allow ophthalmologists diagnose ocular disease. These systems yield large data sets that are often labor-intensive to analyze and require significant expertise in order to draw conclusions, especially when used over time to monitor disease progression. Spectral Domain Optical Coherence Tomography (SD-OCT) instantly acquires depth profiles at a single location with a broadband source. These systems require mechanical scanning to generate two- or three-dimensional images. Instead of mechanically scanning, a beamlet array was used to permit multiple depth measurements on the retina with a single snapshot using a 3x 3 beamlet array. This multi-channel system was designed, assembled, and tested using a 1 x 2 beamlet lens array instead of a 3 x 3 beamlet array as a proof of concept prototype. The source was a superluminescent diode centered at 840nm with a 45nm bandwidth. Theoretical axial resolution was 6.92um and depth of focus was 3.45mm. Glass samples of varying thickness ranging from 0.18mm to 1.14mm were measured with the system to validate that correct depth profiles can be acquired for each channel. The results demonstrated the prototype system performed as expected, and is ready to be modified for in vivo applicability.

International Telemetering Conference Proceedings / October 22-25, 2001 / Riviera Hotel and Convention Center, Las Vegas, Nevada / Reducing spacecraft ground system operations costs are a major goal in all missions. The Fast Auroral Snapshot (FAST) flight operations team at the NASA/Goddard Spacecraft Flight Center developed in-house scripts and procedures to automate monitoring of critical spacecraft functions. The initial staffing profile of 16x7 was reduced first to 8x5 and then to “lights out”. Operations functions became an offline review of system performance and the generation of future science plans for subsequent upload to the spacecraft. Lessons learned will be applied to the challenging Triana mission, where 24x7 contact with the spacecraft will be necessary at all times.

This dissertation merges the optical design and computational aspects of imaging systems to create novel devices that solve engineering problems in optical science and attempts to expand the solution space available to the optical designer. This dissertation is divided into two parts: the first discusses a new active illumination depth sensing modality, while the second part discusses a passive illumination system called plenoptic, or lightfield, imaging. The new depth sensing modality introduced in part one is called depth through controlled aberration. This technique illuminates a target with a known, aberrated projected pattern and takes an image using a traditional, unmodified imaging system. Knowing how the added aberration in the projected pattern changes as a function of depth, we are able to quantitatively determine depth of a series of points from the camera. A major advantage this method permits is the ability for illumination and imaging axes to be coincident. Plenoptic cameras capture both spatial and angular data simultaneously. This dissertation present a new set of parameters that permit the design and comparison of plenoptic devices outside the traditionally published plenoptic 1.0 and plenoptic 2.0 configurations. Additionally, a series of engineering advancements are presented, including full system ray traces of raw plenoptic images, Zernike compression techniques of raw image files, and non-uniform lenslet arrays to compensate for plenoptic system aberrations. Finally, a new snapshot imaging spectrometer is proposed based off the plenoptic configuration.

Introduction: Genetic determination of human face is clearly visible in family members. The resemblance between monozygotic twins who are genetically identical is especially remarkable. So far the possibilities of reliable prediction of the complex morphology of facial traits on the basis of genome analysis and the ability to capture the variability of human facial morphology through genotype variability are highly limited. Complete genetic basis of the physiological variability of craniofacial traits remains more or less unknown. This master's thesis was created as a pilot study of the shared project of the Laboratory of 3D Imagining and Analytical Methods and the Laboratory of Molecular Anthropology on Department of Anthropology and Human Genetics. Material and Methods: The specimen collection is composed of DNA samples derived from 30 families (29 with 4 members, 1 with 5 members) who fulfilled required criteria. Nine single nucleotide polymorphisms were chosen based on the available information. Eight of them are linked to normal facial variability and one was chosen based on the assumed function of the gene where the polymorphism is located. There were two methods of genotyping: RFLP method with the use of restriction endonuclease and SNaPshot method. Morphological data were provided by the...