Le calcul scientifique sur processeurs graphiques (GPU) est en plein essor depuis un certain temps, en particulier dans le domaine de l'apprentissage machine. Cette thèse présente les efforts pour établir une structure de données de table au multidimensionnel de manière efficace sur GPU. Nous commençons par faire une revue de ce qui est actuellement similaire dans le domaine et des désavantages d'avoir une multitude d'approches. Nous nous intéresserons particulièrement aux calculs fait à partir du langage Python. Nous décrirons des techniques intéressantes telles que la réduction d'ordre et le calcul asynchrone automatique. Pour terminer nous présenterons l'utilisation du module développé dans le cadre de cette thèse. / Scientific computing on GPU (graphical processing units) is on the rise, specifically in machine learning. This thesis presents the implementation of an efficient multidimensional array on the GPU. We will begin by a review of what currently implements similar functionality and the disadvantage of a fragmented approach. We will focus on packages that have a Python interface. We will explain techniques to optimize execution such as order reduction and automatic asynchronous computations. Finally, we will present the functionality of the module developed for this thesis. / Le code source de la libraire développée accompagne ce dépôt dans l'état où il était à ce moment. Il est possible de trouver une version plus à jour sur github (http://github.com/abergeron).

Die Komplexität heutiger Geschäftsabläufe und die Menge der zu verwaltenden Daten stellen hohe Anforderungen an die Entwicklung und Wartung von Geschäftsanwendungen. Ihr Umfang entsteht unter anderem aus der Vielzahl von Modellentitäten und zugehörigen Nutzeroberflächen zur Bearbeitung und Analyse der Daten. Dieser Bericht präsentiert neuartige Konzepte und deren Umsetzung zur Vereinfachung der Entwicklung solcher umfangreichen Geschäftsanwendungen. Erstens: Wir schlagen vor, die Datenbank und die Laufzeitumgebung einer dynamischen objektorientierten Programmiersprache zu vereinen. Hierzu organisieren wir die Speicherstruktur von Objekten auf die Weise einer spaltenorientierten Hauptspeicherdatenbank und integrieren darauf aufbauend Transaktionen sowie eine deklarative Anfragesprache nahtlos in dieselbe Laufzeitumgebung. Somit können transaktionale und analytische Anfragen in derselben objektorientierten Hochsprache implementiert werden, und dennoch nah an den Daten ausgeführt werden. Zweitens: Wir beschreiben Programmiersprachkonstrukte, welche es erlauben, Nutzeroberflächen sowie Nutzerinteraktionen generisch und unabhängig von konkreten Modellentitäten zu beschreiben. Um diese abstrakte Beschreibung nutzen zu können, reichert man die Domänenmodelle um vormals implizite Informationen an. Neue Modelle müssen nur um einige Informationen erweitert werden um bereits vorhandene Nutzeroberflächen und -interaktionen auch für sie verwenden zu können. Anpassungen, die nur für ein Modell gelten sollen, können unabhängig vom Standardverhalten, inkrementell, definiert werden. Drittens: Wir ermöglichen mit einem weiteren Programmiersprachkonstrukt die zusammenhängende Beschreibung von Abläufen der Anwendung, wie z.B. Bestellprozesse. Unser Programmierkonzept kapselt Nutzerinteraktionen in synchrone Funktionsaufrufe und macht somit Prozesse als zusammenhängende Folge von Berechnungen und Interaktionen darstellbar. Viertens: Wir demonstrieren ein Konzept, wie Endnutzer komplexe analytische Anfragen intuitiver formulieren können. Es basiert auf der Idee, dass Endnutzer Anfragen als Konfiguration eines Diagramms sehen. Entsprechend beschreibt ein Nutzer eine Anfrage, indem er beschreibt, was sein Diagramm darstellen soll. Nach diesem Konzept beschriebene Diagramme enthalten ausreichend Informationen, um daraus eine Anfrage generieren zu können. Hinsichtlich der Ausführungsdauer sind die generierten Anfragen äquivalent zu Anfragen, die mit konventionellen Anfragesprachen formuliert sind. Das Anfragemodell setzen wir in einem Prototypen um, der auf den zuvor eingeführten Konzepten aufsetzt.

Chemnitz, Techn. Univ., Diplomarb., 2008.

Stuttgart, Univ., Studienarb., 2003.

Στην παρούσα διπλωματική εργασία παρουσιάζεται ο έλεγχος του εργαλείου NLTK (Natural Language Toolkit) της Python. Συγκεκριμένα, το nltk είναι μια ανοιχτού κώδικα βιβλιοθήκη συναρτήσεων για επεξεργασία φυσικής γλώσσας και ανάπτυξη ανάλογων εφαρμογών. Έχει αναπτυχθεί σε γλώσσα Python με στόχο την ανάλυση και ανάπτυξη εφαρμογών κυρίως για την Αγγλική γλώσσα. Αντικείμενο αυτής της εργασίας είναι η συστηματική μελέτη και ο έλεγχος των συναρτήσεων του nltk για την Ελληνική γλώσσα, καθώς υπάρχουν ενδείξεις ότι σημαντικό μέρος αυτών δουλεύει σωστά. Αρχικά, έγινε η μελέτη για εισαγωγή ελληνικών κειμένων, καθώς και κατάλληλη επεξεργασία αυτών, ώστε να είναι σε επεξεργάσιμη μορφή από το εργαλείο. Έπειτα, ελέγχθησαν όλες οι εντολές και κατηγοριοποιήθηκαν με βάση τη λειτουργία τους. Τέλος, παρατηρώντας τα συγκεντρωτικά αποτελέσματα, εξάγεται το συμπέρασμα ότι οι υποψίες για σωστή λειτουργία μεγάλου αριθμού εντολών επαληθεύονται, καθώς το 87,9 % των εντολών φαίνεται να λειτουργεί σωστά. / This diploma dissertation presents the examination of Python NLTK (Natural Language Toolkit) tool. Particularly, nltk is an open source function library suitable for natural language processing and the development of respective applications. It has been developed into Python language in order to analyse and develop applications mostly for the English language. The present dissertation is concerned with the systematic study and the examination of nltk functions for the Greek language, given that there is evidence of the correct operation of some. At first, research for the input of Greek texts as well as their appropriate processing was conducted as a way of presenting these texts in a processable by the tool form. Thereupon, all functions were tested and categorised in terms of their operation. Finally, the observation of concentrated results leads to the conclusion that the initial hypothesis for the correct operation of a great number of order is confirmed, as 87,9% of the functions appears to be operating correctly.

Submitted by Automação e Estatística (sst@bczm.ufrn.br) on 2018-06-05T22:22:46Z No. of bitstreams: 1 LuizAugustoDaSilvaFlorencio_DISSERT.pdf: 2581481 bytes, checksum: 01670ae44134ef41cbe2490b4301c203 (MD5) / Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2018-06-11T22:57:58Z (GMT) No. of bitstreams: 1 LuizAugustoDaSilvaFlorencio_DISSERT.pdf: 2581481 bytes, checksum: 01670ae44134ef41cbe2490b4301c203 (MD5) / Made available in DSpace on 2018-06-11T22:57:58Z (GMT). No. of bitstreams: 1 LuizAugustoDaSilvaFlorencio_DISSERT.pdf: 2581481 bytes, checksum: 01670ae44134ef41cbe2490b4301c203 (MD5) Previous issue date: 2018-02-26 / Estruturas geotécnicas estão sujeitas a incertezas decorrentes da variabilidade natural do solo e do conhecimento limitado dos diversos eventos envolvidos. Nesse sentido, a avaliação de incertezas em projetos geotécnicos pode ser feita mediante aplicação da teoria da confiabilidade. Entretanto, apesar da adoção de analises probabilísticas em normas internacionais e pesquisas científicas, pode-se dizer que, no Brasil, sua aplicação ainda é escassa, especialmente, em estruturas de contenção. Desse modo, nesse trabalho apresenta-se um estudo sobre confiabilidade geotécnica de contenções de estacas em balanço executadas em areia. Rotinas de cálculo computacional para análise de confiabilidade foram inicialmente desenvolvidas utilizando-se a linguagem Python. O primeiro grupo de rotinas foi desenvolvido para Estado Limite Último (ELU) através do método FORM (First Order Reliability Method) e da simulação de Monte Carlo. O segundo grupo de rotinas refere-se à análise de confiabilidade para Estado Limite de Serviço (ELS) mediante simulação de Monte Carlo. Uma contenção com características tipicamente executadas na cidade de Natal-RN foi analisada utilizando-se as rotinas desenvolvidas e para a realização de um estudo paramétrico. Para o ELS utilizou-se ainda o software de elementos finitos Plaxis 2D para obter os deslocamentos horizontais da contenção. A contenção analisada apresentou alta confiabilidade para o ELU com probabilidade de falha nula para os dois métodos probabilísticos e alto índice de confiabilidade tanto para o FORM quanto na simulação de Monte Carlo. Para o ELS, o aumento dos deslocamentos horizontais máximos permitidos diminuiu a probabilidade de falha e aumentou o índice de confiabilidade da contenção, em ambos os casos, com tendência linear. As análises de sensibilidade revelaram que o ângulo de atrito do solo foi a variável mais influente no ELU e ELS na grande maioria dos casos. O peso específico seco, por sua vez, teve baixa influência em todas as análises. Os métodos de sensibilidade aplicados juntos com o Plaxis 2D tiveram resultados satisfatórios para as avaliações da função de estado limite realizadas. Em relação à análise paramétrica, o superdimensionamento da ficha da contenção executada diminuiu a influência da variabilidade das propriedades do solo nos resultados. O efeito da variação das características do solo nos resultados só foi expressivo adotando-se valores de ficha menores que o executado. Nesse caso, considerando a profundidade de escavação constante, o aumento da ficha diminuiu a probabilidade de falha usando os dois métodos probabilísticos e os valores dos pontos de projeto fornecidos pelo FORM. A distribuição lognormal diminuiu a probabilidade de falha das análises no ELS. / Geotechnical structures are subject to different uncertainties due to the natural variability of the soil and because of the limited knowledge of the several related events. The evaluation of these uncertainties is possible with the application of the reliability theory. Despite the discussion on this topic presented in international standards and scientific researches, the application of the probabilistic theory in Brazil is still unusual, especially for retaining structures. Accordingly, this work presents a study on geotechnical reliability of cantilever pile walls in sand. First, computational routines were developed for reliability analysis using the Python language. The first series of routines was developed to evaluate the reliability for the Ultimate Limit State (ULS) of the structure using FORM (First Order Reability Method) and Monte Carlo simulation. The second group of routines was devised to evaluate the reliability for the Service Limit State (SLS) using Monte Carlo simulation. A project of a cantilever pile wall embedded in sandy soil, which is typically constructed in the city of Natal/RN, Brazil, was analyzed using the computational routines. A parametric study was conducted based on this specific retaining wall. Particularly for SLS evaluation, the Plaxis 2D finite element software was used to obtain the horizontal displacement values of the cantilever wall. The retaining structure presented high reliability for the ULS with insignificant probability of failure for both probabilistic methods and great reliability index values for the FORM and for the Monte Carlo simulation. For SLS, the increase in maximum allowed horizontal displacements decreased the failure probability and increased the pile wall reliability index, in a linear trend. Sensitivity analysis indicated that the soil friction angle was the most influential variable in ULS and SLS in most cases. On the other hand, soil dry unit weight showed a low influence in all analyzed cases. Sensitivity methods coupled with Plaxis 2D exhibited satisfactory results for the implemented evaluations. Regarding the parametric study, the oversized embedment depth of the structure led to a reduced influence of soil property variability on reliability analyses. The effect of soil properties variations on the results was significant only for adopted embedment depth smaller than that actually executed. For a constant depth of excavation, increasing embedment depths decreased the failure probability found by the two probabilistic methods and increased the design points provided by the FORM. The use of lognormal distribution reduced the failure probability of the SLS analyzes.

abstract: Though it is a widespread adaptation in humans and many other animals, parental care comes in a variety of forms and its subtle physiological costs, benefits, and tradeoffs related to offspring are often unknown. Thus, I studied the hydric, respiratory, thermal, and fitness dynamics of maternal egg-brooding behavior in Children's pythons (Antaresia childreni). I demonstrated that tight coiling detrimentally creates a hypoxic developmental environment that is alleviated by periodic postural adjustments. Alternatively, maternal postural adjustments detrimentally elevate rates of egg water loss relative to tight coiling. Despite ventilating postural adjustments, the developmental environment becomes increasingly hypoxic near the end of incubation, which reduces embryonic metabolism. I further demonstrated that brooding-induced hypoxia detrimentally affects offspring size, performance, locomotion, and behavior. Thus, parental care in A. childreni comes at a cost to offspring due to intra-offspring tradeoffs (i.e., those that reflect competing offspring needs, such as water balance and respiration). Next, I showed that, despite being unable to intrinsically produce body heat, A. childreni adjust egg-brooding behavior in response to shifts in nest temperature, which enhances egg temperature (e.g., reduced tight coiling during nest warming facilitated beneficial heat transfer to eggs). Last, I demonstrated that A. childreni adaptively adjust their egg-brooding behaviors due to an interaction between nest temperature and humidity. Specifically, females' behavioral response to nest warming was eliminated during low nest humidity. In combination with other studies, these results show that female pythons sense environmental temperature and humidity and utilize this information at multiple time points (i.e., during gravidity [egg bearing], at oviposition [egg laying], and during egg brooding) to enhance the developmental environment of their offspring. This research demonstrates that maternal behaviors that are simple and subtle, yet easily quantifiable, can balance several critical developmental variables (i.e., thermoregulation, water balance, and respiration). / Dissertation/Thesis / Ph.D. Biology 2011

This thesis describes a timing calibration method for the detector array of the ALTO experiment. ALTO is a project currently at the prototype phase that aims to build a gamma-ray astronomical observatory at high-altitude in the Southern hemisphere. ALTO can be assumed as a hybrid system as each detector consists of a Water Cherenkov Detector (WCD) on top of a Scintillator Detector (SD), providing an increased signal to background discrimination compared to other WCD arrays. ALTO is planned to complement the Very-High-Energy (VHE) observations by the High Altitude Water Cherenkov (HAWC) gamma ray observatory that collects data from the Northern sky. By the time the full array of 1242 detectors is installed to the proposed site, ALTO together with HAWC and the future Cherenkov Telescope Array (CTA) will serve as a state-of-the-art detection system for VHE gamma-rays combining the WCD and the Imaging Atmospheric Cherenkov Telescope (IACT) techniques. When a VHE gamma-ray or cosmic-ray enters the Earth’s atmosphere, it initiates an Extensive Air Shower (EAS). These particles are sampled by the detector array and by checking the arrival times of nearby tanks, the method reveals whether a detector suffers from a time-offset. The data analyzed in this thesis derive from CORSIKA (COsmic Ray SImulation for KAscade) and GEANT4 (GEometry ANd Tracking) simulations of cosmic-ray events within the energy range of 1–1:6TeV, which mainly consist of protons. The high flux of this particular type of cosmic-rays, gives us a tool to statistically evaluate the results generated by the proposed timing calibration method. In the framework of this thesis, I have written code in Python programming language in order to develop the timing calibration method. The method identifies detectors that suffer from time-offsets and improves the reconstruction accuracy of the ALTO detector array. Different Python packages were used to execute different tasks: astropy to read filter-present-write large datasets, numpy (Numerical Python) to make datasets comprehensiveto functions, scipy (Scientific Python) to develop our models, sympy (Symbolic Python) to find geometrical correlations and matplotlib (Mathematical Plotting Library) to draw figures and diagrams. The current version of the method achieves sub-nanosecond accuracy. The next stepis to make the timing calibration more intelligent in order to correct itself. This self correction includes an agile adaptation to the data acquired for long periods of time, in order to make different compromises at different time intervals.

abstract: Parental care provides many benefits to offspring. One widely realized benefit is enhanced regulation of offspring's thermal environment. The developmental thermal environment during development can be optimized behaviorally through nest site selection and brooding, and it can be further enhanced by physiological heat production. In fact, enhancement of the developmental thermal environment has been proposed as the initial driving force for the evolution of endothermy in bird and mammals. I used pythons (Squamata: Pythonidae) to expand existing knowledge of behavioral and physiological parental tactics used to regulate offspring thermal environment. I first demonstrated that brooding behavior in the Children's python (Antaresia childreni) is largely driven by internal mechanisms, similar to solitary birds, suggesting that the early evolution of the parent-offspring association was probably hormonally driven. Two species of python are known to be facultatively thermogenic (i.e., are endothermic during reproduction). I expand current knowledge of thermogenesis in Burmese pythons (Python molurus) by demonstrating that females use their own body temperature to modulate thermogenesis. Although pythons are commonly cited as thermogenic, the actual extent of thermogenesis within the family Pythonidae is unknown. Thus, I assessed the thermogenic capability of five previously unstudied species of python to aid in understanding phylogenetic, morphological, and distributional influences on thermogenesis in pythons. Results suggest that facultative thermogenesis is likely rare among pythons. To understand why it is rare, I used an artificial model to demonstrate that energetic costs to the female likely outweigh thermal benefits to the clutch in species that do not inhabit cooler latitudes or lack large energy reserves. In combination with other studies, these results show that facultative thermogenesis during brooding in pythons likely requires particular ecological and physiological factors for its evolution. / Dissertation/Thesis / Ph.D. Biology 2012

I de fallen där det är möjligt använder gärna byggnadsingenjören sig av standardlösningar med materialval och dimensioner som garanterar specifika U-värden. Standardlösningarna måste dock ibland frångås, vid exempelvis specifika krav på tjocklekar eller material som inte finns som standardlösning hos leverantörerna. Byggnadsingenjörens uppgift blir då tidskrävande gällande beräkningar av U-värden, där många iterativa beräkningar krävs, och att överföra konstruktionen till en 3D-modell. En rationalisering av detta arbetsflöde kan åstadkommas med hjälp av programmering. I denna studie skapas ett datorprogram i Visual Basic för Excel, som gör U-värdeberäkningar på olika väggkonstruktioner med trästomme. Även ett script i insticksprogrammet Dynamo skapas för att överföra den beräknade väggkonstruktionen till BIM-programmet Revit. Resultatet blev ett program som beräknar U-värdet, som även innehåller kontroller som hindrar användaren att ange felaktigt indata. Beräkningen med detta program ger en definitiv effektivisering i tid gentemot handberäkningar. Dock kan det ta tid att skriva ett sådant program för ovana, men med tidigare kännedom om programmering kan detta avhjälpas. Att ta sig an och lära sig Dynamo, hur det fungerar och var man hittar rätt funktioner var detta arbetes främsta svårighet. Dynamo bjuder på branta inlärningströsklar och ingen tidigare forskning som berörde det som gjordes i detta arbete. Ett fungerande script lyckades tillslut skapas, men mycket av det svåra arbetet skulle ha underlättats om tidigare kännedom om programmering funnits. Det kan således vara fördelaktigt för byggnadsingenjören som vill arbeta med Dynamo att lära sig programmering i deras utbildning.