京都大学 / 新制・課程博士 / 博士(工学) / 甲第24814号 / 工博第5157号 / 新制||工||1985(附属図書館) / 京都大学大学院工学研究科高分子化学専攻 / (主査)教授 山子 茂, 教授 辻井 敬亘, 教授 大内 誠 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

Die vier in der vorliegenden Dissertation enthaltenen Studien beschäftigen sich vorwiegend mit dem dynamischen Verhalten makroökonomischer Zeitreihen. Diese Dynamiken werden sowohl im Kontext eines einfachen DSGE Modells, als auch aus der Sichtweise reiner Zeitreihenmodelle untersucht. / The four studies of this thesis are concerned predominantly with the dynamics of macroeconomic time series, both in the context of a simple DSGE model, as well as from a pure time series modeling perspective.

Cette thèse travaille principalement sur trois sujets. Le premier concentre sur le clustering en ligne dans lequel nous présentons un nouvel algorithme stochastique adaptatif pour regrouper des ensembles de données en ligne. Cet algorithme repose sur l'approche quasi-bayésienne, avec une estimation dynamique (i.e., dépendant du temps) du nombre de clusters. Nous prouvons que cet algorithme atteint une borne de regret de l'ordre et que cette borne est asymptotiquement minimax sous la contrainte sur le nombre de clusters. Nous proposons aussi une implémentation par RJMCMC. Le deuxième sujet est lié à l'apprentissage séquentiel des courbes principales qui cherche à résumer une séquence des données par une courbe continue. Pour ce faire, nous présentons une procédure basée sur une approche maximum a posteriori pour le quasi-posteriori de Gibbs. Nous montrons que la borne de regret de cet algorithme et celui de sa version adaptative est sous-linéaire en l'horizon temporel T. En outre, nous proposons une implémentation par un algorithme glouton local qui intègre des éléments de sleeping experts et de bandit à plusieurs bras. Le troisième concerne les travaux qui visent à accomplir des tâches pratiques au sein d'iAdvize, l'entreprise qui soutient cette thèse. Il inclut l'analyse des sentiments pour les messages textuels et l'implémentation de chatbot dans lesquels la première est réalisé par les méthodes classiques dans la fouille de textes et les statistiques et la seconde repose sur le traitement du langage naturel et les réseaux de neurones artificiels. / This thesis works mainly on three subjects. The first one is online clustering in which we introduce a new and adaptive stochastic algorithm to cluster online dataset. It relies on a quasi-Bayesian approach, with a dynamic (i.e., time-dependent) estimation of the (unknown and changing) number of clusters. We prove that this algorithm has a regret bound of the order of and is asymptotically minimax under the constraint on the number of clusters. A RJMCMC-flavored implementation is also proposed. The second subject is related to the sequential learning of principal curves which seeks to represent a sequence of data by a continuous polygonal curve. To this aim, we introduce a procedure based on the MAP of Gibbs-posterior that can give polygonal lines whose number of segments can be chosen automatically. We also show that our procedure is supported by regret bounds with sublinear remainder terms. In addition, a greedy local search implementation that incorporates both sleeping experts and multi-armed bandit ingredients is presented. The third one concerns about the work which aims to fulfilling practical tasks within iAdvize, the company which supports this thesis. It includes sentiment analysis for textual messages by using methods in both text mining and statistics, and implementation of chatbot based on nature language processing and neural networks.

No description available.

Made available in DSpace on 2016-03-15T19:38:03Z (GMT). No. of bitstreams: 1 RONALDO DE CASTRO CORREA.pdf: 1467426 bytes, checksum: a5e2b011297c7075c278d2dfd4502c98 (MD5) Previous issue date: 2015-11-23 / Cellular automata are discrete dynamic systems that by local action rules, even very simple, can present arbitrarily complex global processing. The reversibility is a property that a few rules have and guarantees the existence of an inverse rule capable of reversing the temporal evolution of the original rule. For a rule be reversible all possible configurations of lattices must have only one predecessor, i.e., only one pre-image. In this context, it had been proposed in the literature the concept of rules may be characterized by its relative partial reversibility, that is, rules can be more or less reversible than others. This notion is represented by the rule pre-image pattern, which is composed of the quantities ordered pre-images of all the possible configurations of lattices up to a maximum size. To sort the rules of the reversible for less reversible, or group them if they had the same reversibility was made lexicographical ordering patterns of pre-images. This paper reviewed the original definition, based on elementary cellular automata, although the results are applicable to any other one-dimensional family rules. Thus, proposed it was a measure of the reversibility level of a rule, also based on its default pre-image, but now from the probability of correctly reverse each possible configurations of lattices up to a given maximum size. This measure allows us to analyze the degree of reversibility of a rule in absolute terms and not relative to other rules. Thus, it becomes possible to individually analyze the reversibility levels of rules, making it possible to infer the degree of rules reversible for lattices arbitrarily larger than calculated, in particular, identifying rules that tend to be reversible when the size of the lattices tends to infinity. It was also possible to define an operation that, from their own state transitions rule, allows to obtain partially rules that have the same level of reversibility without the need of standard calculating preview image, which is extremely expensive computationally. / Autômatos celulares são sistemas dinâmicos discretos que, por meio de regras de ação local, até mesmo muito simples, podem apresentar processamento global arbitrariamente complexo. A reversibilidade é uma propriedade que poucas regras possuem e que garante a existência de uma regra inversa capaz de reverter a evolução temporal da regra original. Para uma regra ser reversível, todas as configurações possíveis de reticulado devem possuir somente uma única configuração antecessora, ou seja, uma única pré-imagem. Nesse contexto, havia sido proposto na literatura o conceito de regras poderem ser caracterizadas por sua reversibilidade parcial relativa, ou seja, regras poderem ser mais ou menos reversíveis que outras. Essa noção é representada por meio do padrão de pré-imagem da regra, que é composto pelas quantidades ordenadas de pré-imagens de todas as configurações possíveis de reticulado, até um tamanho máximo. Para classificar as regras das mais reversíveis para as menos reversíveis, ou agrupá-las caso possuíssem a mesma reversibilidade, era feita a ordenação lexicográfica dos padrões de pré-imagens. Este trabalho reavaliou a definição original, com base nos autômatos celulares elementares, apesar de os resultados serem aplicáveis a qualquer outra família unidimensional de regras. Assim, foi proposta uma grandeza que representa o nível de reversibilidade de uma regra, também baseada em seu padrão de pré-imagem, mas agora a partir da probabilidade de reverter corretamente cada configuração possível de reticulado, até um tamanho máximo dado. Tal medida permite analisar o nível de reversibilidade de uma regra em termos absolutos, e não mais relativamente a outras regras. Dessa forma, torna-se possível analisar individualmente os níveis de reversibilidade das regras, o que possibilitou inferir o nível de reversibilidade de regras para reticulados arbitrariamente maiores que os calculados, em particular, identificando regras que tendem a ser reversíveis conforme o tamanho do reticulado tende a infinito. Também foi possível definir uma operação que, a partir das próprias transições de estado de uma regra, permite obter parcialmente as regras que possuem o mesmo nível de reversibilidade, sem a necessidade do cálculo do padrão de pré-imagem, que é extremamente custoso computacionalmente.

No description available.

The alarming rate at which we see the decline in biodiversity due to human activity has raised concerns about the well-being of our planet. Butterflies which serve as pollinators are an essential part of many ecosystems and sensitive indicators of environmental changes and can provide valuable insight into how ecosystems function and evolve. This thesis aims to create phylogenetic trees based on DNA sequences from butterflies and compare different nucleotide substitution models and methods in order to better understand butterflies' evolution and genetic relationships. Our approach was to use Markov theory to investigate how the four nucleotides are evolving. In regard to the comparison of models, the General Time Reversible model with more degrees of freedom was found to be better than the K80 model. Although the Maximum Likelihood and Pairwise Distance methods were found to have different transition rate matrices, the tree reconstructions had no registered differences. Interestingly, the Q matrix was found to be similar across butterfly families. These findings can suggest that it is possible to have a standard Q matrix when estimating or inferring evolutionary relationships among butterflies, and probably other animal groups. This should improve the accuracy of estimations within phylogenetics when dealing with small data sets. The information helps with reconstructing evolutionary relationships and species, therefore contributing to preserving biodiversity and thereby the ecosystems to whom they belong - and in addition humankind. / Den oroväckande takten med vilken vi ser en minskning i biologisk mångfald på grund av mänsklig aktivitet har väckt rädsla för vår planets fortsatta välbefinnande. Fjärilar som är en väsentlig del av många ekosystem fungerar som pollinatörer och indikatorer för miljöförändringar, vilket ger värdefull insikt om ekosystemens funktion och utveckling. Fjärilar fungerar som pollinatörer och är en viktig del av många ekosystem. Därmed är de känsliga indikatorer på miljöförändringar och kan ge värdefull insikt om hur ekosystem fungerar samt utvecklas. Detta kandidatexamensarbete syftar till att skapa fylogenetiska träd baserade fjärilars DNA-sekvenser och jämföra olika modeller för nukleotid substitution och metoder för att bättre förstå fjärilars utveckling och genetiska relationer. Vårt tillvägagångssätt var att använda Markovs teori för att undersöka hur de fyra nukleotiderna utvecklas. När det gäller jämförelsen av modeller visade sig General Time Reversible-modellen med fler frihetsgrader vara bättre än K80-modellen. Fastän Maximum Likelihood och Pairwise Distance metoderna visade sig ha olika övergångsmatriser fanns det inga registrerade skillnader i trädrekonstruktionerna. Ytterligare ett intressant resultat var att Q-matrisen var liknande oberoende av fjärilsfamilj. Detta kan tyda på att det är möjligt att ha en standard Q-matris när man uppskattar eller drar slutsatser om evolutionära samband mellan fjärilar och förmodligen andra djurgrupper. Vidare studier behövs men detta skulle förbättra noggrannheten av uppskattningar inom fylogenetiken när man hanterar små datamängder. Sammanfattningsvis hjälper nämnda insikter till att rekonstruera evolutionära relationer bland arter och bidrar därför till att bevara den biologiska mångfalden, tillhörande ekosystem och dessutom mänskligheten.

<p>    </p> <p>We live in an age in which computation touches upon every aspect of our lives in ever increasing ways. To meet the demand for increased computing power and ability, new computation strategies are continually being proposed. In this dissertation, we consider two research projects related to two such cutting edge paradigms. We first consider developing superconducting devices that implement asynchronous reversible ballistic computation. This paradigm was developed to circumvent Landauer’s principle of a minimum energy required per bitwise computation operation. We report the design of a new device, the rotary, which is a critical step towards developing universal computation gates in the scheme of synchronous reversible ballistic computation. Next, we turn to the consideration of anyons which have been predicted to enable topological quantum computing, a quantum computing paradigm that is relatively immune to environmental noise. We consider initial steps in the development of a Bethe ansatz solvable model that will help decipher the many-body properties of Majorana zero modes in superconducting Kitaev wires. </p>

Cette étude porte sur l’application de la réaction de Diels-Alder dans la synthèse de réseaux polymères thermoréversibles ainsi que la compatibilisation de mélanges de polymères initialement immiscibles. Les polymères étudiés ne sont autres que le polyméthacrylate de méthyle et le polyuréthane. Les réseaux thermoréversibles sont obtenus par réaction de Diels-Alder faisant intervenir le couple furanne-maléimide. Les polymères fonctionnalisés furanne ou maléimide sont obtenus soit par modification chimique d’un polymère commercial (cas du PMMA) soit par réaction de polyaddition (cas du polyuréthane). Les réseaux sont ensuite obtenus par réaction de DA avec des agents de couplage maléimide (bis- ou tris-maléimide) pour les polymères fonctionnalisés furanne ou par réaction de DA avec des agents de couplage furanne pour les prépolymères fonctionnalisés maléimide. La thermo-réversibilité a été soulignée par DSC, par des tests de solubilité à chaud ainsi que par rhéologie. Des réseaux interpénétrés thermoréversibles PMMA/PU ont été également obtenu par Diels-Alder et ce grâce à la présence d’interaction supramoléculaire entre les deux phases. / This study focuses on the application of the Diels-Alder reaction in the synthesis of thermoreversible polymer networks and compatibilization of immiscible polymer blends. The studied polymers are polymethyl methacrylate and polyurethane. Thermoreversible networks were obtained by a Diels-Alder involving furan-maleimide couple. The furan or maleimide functionalized polymers were obtained either by chemical modification of a commercial polymer (case of PMMA) or by polyaddition reaction (case of polyurethane). Networks were then obtained by DA reaction with maleimide coupling agents (bis-or tris-maleimide) for furan functionalized polymers or DA reaction with furan coupling agents for maleimide functionalized prepolymers. The thermoreversibility was highlighted by DSC, solubility tests (at high temperature) as well as rheology. Thermoreversible interpenetrating PMMA/PU networks were also obtained by Diels-Alder reaction and this is due to the presence of supramolecular interaction between the two phases

La thèse porte sur la conception et la simulation d'un réservoir de stockage solide de l'hydrogène sous forme d'hydrure de magnésium (MgH2). La particularité du réservoir conçu réside dans sa capacité à stocker l'énergie d'absorption grâce à un matériau de changement de phase (MCP). Afin de pouvoir prouver la viabilité du système, une étude portant sur le comportement de l'hydrure de magnésium compacté lors du cyclage à été effectuée. Celle-ci montre qu'après 100 cycles, les cinétiques de réaction et les taux massiques de stockage d'hydrogène ne sont pas affectés. En revanche, un changement de morphologie important a été observé puisqu'une dilatation ainsi qu'une augmentation importante de la conductivité des matériaux composites ont été relevées. L'étude du MCP révéla l'importance de certains paramètres, en particulier la conductivité thermique et l'enthalpie de fusion. Le MCP sélectionné est un alliage métallique en composition eutectique. Celui ci est bon conducteur de chaleur, présente une enthalpie de fusion élevée et une stabilité de comportement thermique au cyclage. Le réservoir construit contient 10 kg d'hydrure de magnésium co-broyé + 5 % de Graphite Naturel Expansé. Il est capable de stocker 7000 NL d'hydrogène (625 g) en 3h. L'avantage principal du réservoir est son efficacité énergétique, puisque la chaleur stockée par le MCP à l'absorption est refournie lors de la désorption. Afin de pouvoir prédire les comportements thermiques et cinétiques des prochains réservoirs basés sur cette technologie, 2 modèles numériques utilisant Matlab et Fluent ont été développés et validés. / The thesis's subject is about creation and modeling of a solid state hydrogen tank using magnesium hydride (MgH2). The main characteristic of this tank is the ability to store the heat of absorption due to the use of a Phase Change Material (PCM). In order to prove the sustainability of this system, a study, on the magnesium hydride's behavior, has been carried out. On one hand, kinetic properties and the amount of the stored hydrogen do not decrease after 100 cycles. On the other hand, a significant change on material morphology has been noticed. Indeed, a swelling and an increasing of thermal conductivity have been measured. Investigations about the MCP showed the importance of the thermal conductivity and the heat of fusion. That's why a metallic eutectic alloy has been selected. His atomic composition is Mg69Zn28Al3, he is a good thermal conductor, having a high heat of fusion, and presenting a good chemical stability during cycling. The designed tank contains 10 kg of magnesium hydride ball-milled added with Expanded Natural Graphite. It can absorb 7000 NL (625 g) of hydrogen in 3 hours and a half. On one total cycle, the energetic efficiency can be estimated to more than 70 %. At the same time, two numerical modeling have been achieved with Fluent and Matlab softwares, in order to make the design of next generation of tanks easier.