Ces travaux sont consacrés à l'étude d'un gaz d'excitons bidimensionnel créé dans un double puits quantique de semi-conducteurs. Grâce à un champ électrique appliqué perpendiculairement au plan des puits, les deux composants de l'exciton, l'électron et le trou, sont chacun confinés dans deux puits séparés. Ceci donne lieu à des excitons avec un fort moment dipolaire qui sont piégés et refroidis à très basse température. Dans ce manuscrit, nous étudions les propriétés du gaz d'excitons à travers la photoluminescence émise lors de la recombinaison des paires électron-trou, que nous analysons grâce à des mesures spectroscopiques et interférométriques. Ainsi, nous avons étudié la thermodynamique du gaz d'excitons dans un piège harmonique et établi son équation d'état. Nous avons dévoilé l'universalité de l'équation d'état ce qui a rendu possible la première démonstration de l'équilibre thermodynamique du gaz et a aussi permis d'évaluer la force des interactions entre excitons. En associant des mesures de cohérence spatiale et l'étude de la distribution de défauts de densité dans la photoluminescence des excitons, nous avons montré que dans une zone centrale du piège de rayon R < 3 µm, il s'opère conjointement une augmentation du niveau de cohérence et une diminution de la densité de défauts. Des travaux précédents ayant montré que ces défauts pouvaient constituer des défauts topologiques appelés vortex, nos observations constituent ainsi un ensemble de preuves de l'établissement d'un état superfluide dans la partie centrale du piège, en accord avec la théorie prédite par Berezinskii-Kosterlitz et Thouless. / This work is devoted to the study of a two dimensional exciton gas created in a semiconductordouble quantum wells. Thanks to an electric field applied perpendicular to these layers,the exciton constituents, namely, the electron and the hole, are confined each in one quantumwell so that excitons inherit a strong electric dipole. In this manuscript, we study the propertiesof such exciton gas trapped at very low temperature, through the photoluminescence emittedby electron-hole recombinations. We particularly study the emitted light using spectroscopic andinterferometric techniques. Thus, we explore the thermodynamics of a trapped gas and quantifyits equation of state. We unveil an important feature: the universality of the equation of state,which allows the first demonstration of thermodynamic equilibrium for excitons, and we realizea direct measurement of the excitons dipolar interaction strength. By probing both the spatialcoherence and the density of defects of the excitons photoluminescence, we show that it exists acentral region (R < 3 μm) in an exciton trap where a high degree of spatial coherence is conditionedby a decrease of the defects density. Our experiments thus provide a set of evidences pointingtowards a Berezinskii-Kosterlitz-Thouless transition for trapped 2D exciton gas.

Cette thèse s'intéresse à l'étude de la supraconductivité bidimensionnelle à l'interface entre les oxydes LaAlO3 et SrTiO3 contrôlée par effet de champ électrique. Lorsqu'on fait croître une couche mince de quelques mailles atomiques de LaAlO3 sur un substrat de SrTiO3, l'interface devient conductrice, et même supraconductrice au-dessous de 300mK, bien que ces deux oxydes de structure pérovskite soient des isolants. Il se forme ainsi un gaz bidimensionnel d'électrons de haute mobilité, dont les propriétés - supraconductivité et fort couplage spin-orbite de type Rashba - peuvent être contrôlées par effet de champ électrique à l'aide d'une Back Gate. Nous avons étudié cette supraconductivité bidimensionnelle par trois approches expérimentales différentes : l'étude de la transition supraconductrice en température à l'aide du modèle de Berezinskii-Kosterlitz-Thouless incluant une distribution inhomogène de rigidité ; l'analyse par le groupe de renormalisation de la transition de phase quantique supraconducteur-isolant induite par un champ magnétique perpendiculaire à l'interface selon le modèle de Spivak, Oreto et Kivelson ; et enfin l'étude de l'hystérèse du courant critique ainsi que de sa nature probabiliste dans le cadre du modèle RCSJ. Nous proposons donc de considérer l'interface comme un réseau de flaques supraconductrices couplées par effet Josephson à travers un gaz 2D métallique, dont la transition est régie par le modèle XY des fluctuations de phase. Enfin, nous avons démontré pour la première fois la possibilité de contrôler les propriétés du gaz 2D à l'aide d'une Top Gate, et comparé les effets des deux grilles (Top ou Back Gate). / In this PhD work, we study the field-effect modulated two-dimensional superconductivity at the LaAlO3/SrTiO3 oxides hetero-interface. When one grows epitaxially a few unit cells thin film of LaAlO3 on a SrTiO3 substrate, the interface becomes conducting, and even superconducting below 300mK, although these two perovskite oxides are insulators. The properties of this high-mobility two-dimensional electron gas – superconductivity and strong Rashba-type spin-orbit coupling - can be field-effect modulated by the mean of a Back Gate. We have investigated this two-dimensional superconductivity through three different experimental approaches : the temperature-driven transition with the Berezinskii-Kosterlitz-Thouless model including an inhomogeneous distribution of rigidity ; the finite-size scaling analysis of the superconductor-to-insulator quantum phase transition induced by a perpendicular magnetic field using the model developed by Spivak, Oreto and Kivelson ; and finally the measure of the hysteretic and stochastic properties of the critical current in the framework of the RCSJ model. We hence propose to consider this 2D electron gas as an inhomogeneous network of superconducting puddles coupled to one another by Josephson effect through a normal metallic matrix, which transition is dominated by the XY model of the phase fluctuations. Finally, we demonstrated for the first time the possibility of Top Gated-control of this interface, and we investigated the compared effects of Top and Back Gating.

京都大学 / 0048 / 新制・課程博士 / 博士(人間・環境学) / 甲第22546号 / 人博第949号 / 新制||人||226(附属図書館) / 2019||人博||949(吉田南総合図書館) / 京都大学大学院人間・環境学研究科相関環境学専攻 / (主査)准教授 木下 俊哉, 教授 吉田 鉄平, 教授 森成 隆夫 / 学位規則第4条第1項該当 / Doctor of Human and Environmental Studies / Kyoto University / DGAM

<div>Today, machine learning models and Deep Neural Networks (DNNs) are prevalent in various areas. Also, educational Artificial Intelligence (AI) is drawing increasing attention</div><div>with the rapid development of online learning platforms. Researchers explore different types of educational AI to improve students’ learning performance and experience in online classes. Educational AIs can be categorized into “interactive” and “predictive.” Interactive AIs answer simple course questions for students, such as the due day of homework and the final project’s minimum page requirement. Predictive educational AIs play a role in predicting students’ learning states. Instructors can adjust the learning content based on the students’ learning states. However, most AIs are not evaluated in an actual class setting. Therefore, we want to evaluate the effects of a state-of-the-art educational AI model, BKT (Bayesian Knowledge Tracing)-LSTM(Long Short-Term Memory), on students’ learning performance in an actual class setting. Data came from the course CNIT 25501, a large introductory Java program?ming class at Purdue University. Participants were randomly separated into the control and experimental groups (AI-group). Weekly quizzes measured participants’ learning performance. Pre-quiz and base quizzes estimated participants’ prior knowledge levels. Using BKT-LSTM, participants in the experimental group had questions from the knowledge that they were most lacking. However, participants in the control group had questions from randomly picked knowledge. The results suggested that both the experimental and control groups had lower scores in review quizzes than in base quizzes. However, the score difference between base quizzes and review quizzes for the experimental group was more often significantly different (three quizzes) compared to the control group (two quizzes), demonstrating the predictive capability of BKT-LSTM to some extent. Initially, we expected that BKT-LSTM would enhance students’ learning performance. However, in post-quiz, participants in the control group had significantly higher scores than those in the experimental group. The result suggested that continuous complex questions may negatively affect students’ learning initiatives. On the contrary, relatively easy questions may improve their learning initiatives.</div>

<div>With the advent of Machine Learning and Deep Learning models, many avenues of development have opened. Today, these technologies are being leveraged to perform a wide variety of tasks that were otherwise not possible with traditional systems. The power of Machine Learning and Artificial Intelligence makes it possible to compute very complicated tasks at near real-time speeds. To provide an example, Machine Learning models are used extensively in the retail industry to predict and analyze critical parameters such as sales, promotions, customer behavior, recommendations, and offers.</div><div><br></div><div><br></div><div>Today, it is increasingly common to observe AI being used across many of the biggest domains such as Health, Environment, Military, and Business. Artificial Intelligence being used in educational settings has thus been a growing field of focus and study. For example, conversational AI being deployed to act as virtual tutors to answer student questions and concerns. Additionally, there is a fill-the-hole type of AIs that will help students learn tasks such as coding by either showing them how to do it or by predicting where the student might go wrong and suggesting preemptive corrective steps. </div><div><br></div><div>As described, a great deal of literature exists about the use of Deep Learning and Machine Learning models in education. However, the existing tools and models act as external appendages that add to the course structure, thereby altering it. This proposed study introduces a Bayesian Knowledge Transfer model based on the Long Short Term Memory structure (BKT-LSTM) utilized in a live STEM (Science, Technology, Engineering, and Mathematics) classroom. The model discovers individual student learning profiles based on past quiz performance and customizes future quizzes based on the learned patterns. The BKT-LSTM model works in tandem with the existing course curriculum and only tests those knowledge items that have already been covered in the classroom. The model does not change the course structure but rather aims to improve the student’s learning experience by focusing on areas of the student's knowledge that require more practice in learning. </div><div><br></div><div><br></div><div>Within a live STEM classroom, the BKT-LSTM model acts as a herald of change in the way students interact with the curriculum, even though no major changes are observed in the course structure. Students interacting with the model are subjected to quizzes with questions that target the individual student’s lack of learning in particular knowledge areas. Thus, students can be expected to perceive the change as unwelcoming due to the increasing difficulty in subsequent quizzes. Regardless, the study focuses on measuring the learning performance of the students. Do the students learn more in the new system? Another focus of the study is the student’s perception of engagement while interacting with the BKT-LSTM model. The effectiveness of the new educational process is determined not only by increased student learning performance, but also by the student’s perception of engagement while interacting with the model. Are the students enjoying the new experience? Do the students feel like they are learning something? Another important factor was also studied, that is learning performance of students interacting with the BKT-LSTM. </div><div><br></div>

Ce travail de thèse rapporte la production tout optique de condensats de Bose-Einstein de 39K. Une étape clé du processus expérimental est l’obtention d’un nuage suffisamment froid permettant le chargement direct d’un piège dipolaire de manière efficace. Notre solution est l’utilisation d’une mélasse fonctionnant dans le bleu de la raie D1 de cet alcalin conduisant à une densité dans l’espace des phases élevée et ainsi au chargement direct d’un grand nombre d’atomes dans un piège à 1550 nm. Le nuage est ensuite polarisé puis comprimé dans un piège dipolaire croisé avant d’entamer un refroidissement évaporatif efficace au voisinage d’une résonance de Feshbach. Ce processus permet la production rapide de condensats de Bose-Einstein toutes les 7 secondes sur notre expérience. Ces nuages dégénérés représentent le point de départ pour la conduite d’expériences visant à étudier les effets du désordre dans les gaz quantiques en dimensions réduites. Nous envisageons l’étude du diagramme de phase du gaz de Bose bidimensionnel désordonné, de la localisation d’Anderson en dimension deux ainsi que l’étude de l’influence du désordre sur un soliton brillant dans une géométrie unidimensionnelle. / This thesis presents the all optical production of 39K Bose-Einstein condensates. A key point in the process is the sub-Doppler cooling that allows for an efficient loading of an optical dipole trap. To this aim we use a gray molasses scheme working on the blue side of the D1 line of this alkali that leads to a high phase space density and a high number of trapped atoms in a 1550 nm optical trap. The cloud is then polarized and compressed in a crossed dipole trap before starting an efficient forced evaporation close to a Feshbach resonance. This process allows us to produce Bose-Einstein condensates every 7 seconds with our experiment. Those degenerate clouds represent the starting point of experiments aiming to study the influence of disorder on quantum gases in low dimensions. We discuss the perspectives to study of the phase diagram of the two-dimensional disordered Bose gas as well as the Anderson localization phenomenon in two dimensions and the behaviour of bright solitons in a disordered potential in a one-dimensional geometry.

Cette thèse s'intéresse aux phénomènes électrostatiques émergents dans les modèles magnétiques toroïdaux bi-dimensionnels à symétrie XY, fournissant ainsi un support pour de plus amples recherches dans le domaine de la transition de phase Berezinskii-Kosterlitz-Thouless (BKT).Dans de nombreux systèmes bi-dimensionnels, dont le modèle bi-dimensionnel XY du magnétisme, la transition BKT contrôle la dissociation thermique de paires de défauts topologiques liés. Le modèle XY est analogue au gaz de Coulomb bi-dimensionnel, à ceci près qu'il peut être simulé sans avoir à modéliser les interactions à longue distance du système Coulombien. Cette thèse élucide ce paradoxe en démontrant que l'approximation de Villain appliquée au modèle XY est strictement équivalente au modèle électrostatique de Maggs-Rossetto (MR) appliqué au système Coulombien bi-dimensionnel.Cette équivalence est utilisée pour sonder la transition BKT par l'application de l'algorithme MR au gaz de Coulomb bi-dimensionel. En simulant le système Coulombien, il est prouvé que les fluctuations dans l'organisation des charges autour du tore sont activées à la température de transition BKT. Ces fluctuations du champ électrique indiquent ainsi la phase de haute température de la transition.Il est ensuite montré que l'exposant critique effectif de la théorie de Bramwell-Holdsworth (BH) peut être mesuré dans les films d'hélium 4 superfluide, qui correspondent à des gaz de Coulomb effectifs dans la limite de systèmes de grandes tailles finies. / This thesis addresses the emergent electrostatics of two-dimensional, toroidal magnetic models that possess XY symmetry, providing a platform for novel investigations into the Berezinskii-Kosterlitz-Thouless (BKT) phase transition.The BKT transition drives the thermal dissociation of bound pairs of topological defects in many two-dimensional systems, including the two-dimensional XY model of magnetism. The XY model is closely analogous to the two-dimensional Coulomb gas, but can be simulated without computing the long-range interactions of the Coulombic system. This thesis elucidates this paradox by showing that Villain's approximation to the XY model is strictly equivalent to the Maggs-Rossetto (MR) electrostatic model when applied to the two-dimensional Coulomb gas.The mapping is used to probe the BKT transition through the application of the MR algorithm to the two-dimensional Coulomb gas. By simulating the Coulombic system, fluctuations in the winding of charges around the torus are shown to turn on at the BKT transition temperature. These topological-sector fluctuations in the electric field therefore signal the high-temperature phase of the transition.It is then shown that the effective critical exponent of Bramwell-Holdsworth (BH) theory can be measured in superfluid 4He films, which correspond to effective Coulomb gases in the limit of large but finite system size. With the Coulombic system taken as the base BKT system, it is inferred that BH theory is a general property of BKT systems.