Thesis abstract: Geopolitical Rivalry in Central Asia and Turkmenistan's power as a weak state Charles University, Institute of International Studies, MAS programme 2019 Emma Welsink English Abstract Research on weak states and their position in the international system has been a widely discussed topic for decades, especially in relation to the Cold War bipolar system that revolved around Russia and the United States as great powers. Yet what is lacking is a contemporary analysis that concentrates on the current multipolar system in which the importance of economic alliances and power surpass military power, and how this affects weak states' foreign policy behavior and trade opportunities. This research therefore offers an empirical analysis on great power geopolitical competition between Russia and China over Turkmenistan's economic allegiance, specifically regarding its natural gas export, and seeks to explain how this geopolitical competition has affected Turkmenistan's foreign energy policy behavior as a weak state vis-á-vis these great powers. Additionally, this research seeks to examine how Turkmenistan has even instrumentalized this competition to further its national interest of energy export diversification. More specifically, this research shows that great power competition has played a...

The dissertation thesis addresses the topic of mathematical tasks processing in students with autism spectrum disorder (ASD) and problematic points of their solution. The aim of this work is to delineate methods that ASD students use to approach these tasks and solve them. We focus primarily on whether it is possible to identify different procedures than those used by typically developing students, and whether these procedures show specifics corresponding to distinct cognitive processes. In the theoretical part we provide a theoretical background for the study. Firstly, we create a frame to understand the complexity of autism, furthermore we deal with psychological theories examining the topic, and describe specifics of development of ASD students during their adolescence. Thereafter, we reflect on studies addressing mathematical skills of ASD students. Mathematics is considered as their great strength, however is not sufficiently scrutinized, as highlighted by various international studies (for example Oswald et al., 2016). The empiric part is based on qualitative methodology and pursues comparisons of mathematical tasks processing between six ASD students and typically developing peers. We strive to document differences and common points in solutions as well as to analyze cognitive processes of...

Geokunststoffe sind beständige Produkte aus Polymeren, die in Boden eingelegt werden um geotechnische Probleme zu lösen. Eine Kategorie der Geokunststoffe sind Geogitter, bei denen die Längs- und Querelemente eine offene, gitterförmige Struktur bilden und die meist zur Bodenbewehrung eingesetzt werden. Ihre Verwendung zur Bewehrung von gering tragfähigen Böden für kunststoffbewehrte Erd- und Stützbauwerke kann enorme ökologische und ökonomische Vorteile bieten. Trotz der Tatsache, dass die Verwendung von Geogittern in den letzten Jahren stark zugenommen hat, sind die Grundlagen zum Bau und zur Bemessung limitiert. Besonders die Schlüsselfrage, wie Geogitter und verschiedene Lockergesteine in unterschiedlichen Belastungssituationen interagieren, ist ingenieursmäßig nur lückenhaft untersucht worden. Im Rahmen dieser Dissertation wurde deshalb in über 250 großmaßstäblichen Scher-, Herauszieh- und Reibungsversuchen das Verbundverhalten Geokunststoff/Locker¬gestein eingehend untersucht. Dazu wurden 14 handelsübliche Geogitter, die teilweise auch modifiziert wurden, in dem Interaktionsprüfgerät des Institutes für Geotechnik der TU Bergakademie Freiberg mit verschiedenen Böden systematisch getestet. Die vielen Versuchsanordnungen und die aus ihnen abgeleiteten Überlegungen und gewonnenen Erkenntnisse belegten, dass die wichtigsten Parameter interagieren, und zwar in verschiedenen Abhängigkeiten wie weitere Variationen der Bodenparameter (Kornverteilung, Wassergehalte, Verdichtung, Bindemittelzusatz etc.), der unterschiedlichen Geogitterparameter (Oberflächenbeschaffenheit, Struktur, etc.) und der Belastung gezeigt haben. Darüber hinaus wurde das Interaktionsprüfgerät für zukünftige Versuche weiterentwickelt (Entkopplung des Versuchseinbaus vom Versuchsgerät usw.). Der Fokus der Forschungsarbeit lag dabei in der Schaffung von Grundlagenkenntnissen für den gemeinsamen Einsatz von bindigen Lockergesteinen und Geokunststoffen. Bei rolligen Böden zeigte sich, dass die Rautiefe sich stärker (positiv) auf die Widerstände auswirkt als bei den bindigen Böden. Hingegen waren bei bindigen Böden die Auswirkungen durch eine Erhöhung der Querelemente deutlicher. Darüber hinaus konnte gezeigt werden, dass dreidimensionale Querelemente das Verbundverhalten verbessern und Tests an Geogittern mit einem größeren Verhältnis von Öffnungsweite zu Maschenweite zeigten höhere Adhäsionswerte. Die Reibungsversuche hatten einen gleichmäßigeren Verlauf und geringere Streuungen zwischen den unterschiedlichen Geogittern, weshalb aufgrund der deutlicheren Unterschiede zwischen den Produkten bei den Herausziehversuchen detaillierte Schlussfolgerungen gezogen werden konnten. Die vorliegende Arbeit leistet einen Beitrag zur Verbesserung der Bemessungsgrundlagen und ermöglicht ein detaillierteres Verständnis des Verbundverhaltens. Des Weiteren werden Vorschläge ausgearbeitet die im Labor erzielten Ergebnisse auf die Baupraxis zu übertragen, wie zum Beispiel durch das Vermeiden von Trennflächen zwischen Geogitter und Boden bei der Verdichtung. / Geosynthetics are durable polymeric products placed in soil to help solve civil engi-neering problems; one category of these is geogrids. Geogrids are polymers formed into an open, grid-like configuration and function primarily as soil reinforcement. Their use to reinforce soil with weak bearing capacity could have big ecological and economical advantages. Although geogrid use has grown in recent years, there is still limited design information available due to the wide variety of design configurations and soil variability. One of the key factors lacking for more widespread engineering implementation is a greater understanding of the interaction between various geogrids and soil combinations when exposed to different load effects. In more than 250 shear, pull-out, and friction tests, this interaction behavior was examined in detail. Fourteen general geogrids (sometimes additionally modified) using dif-ferent soil types were systematically tested in the Interaction Testing Device at the Geotechnical Institute of the TU Bergakademie Freiberg. Combining the huge quantity of test results with a theoretical analysis, the main parameters which affect soil behavior could be defined. Furthermore the Interaction Testing Device could be improved for test in future. Our analysis showed us that the roughness of the geogrid has more of a beneficial impact on the strength of frictional soils than it does for silty soils. Silty soils however, were more affected by the height of cross elements in the geogrid than other types. It was also observed that 3D elements in the cross machine direction improve the interaction behavior. Tests with geogrids which have a bigger ratio of opening size to mesh size showed higher adhesion values. The friction test results were more regular, had a uniform progress and had smaller differences between geogrid configurations. Therefore, conclusions with higher certainty could be drawn from the pull-out tests with the differential results. The present research contributes to the limited design information to help provide a better understanding of the interaction behavior. Furthermore, suggestions are given to use the laboratory detected results in real-world applications, one of which is the way of compaction to minimize the potential for creating a slip-surface between the geogrid and soil.

We study nonlinear evolutionary partial differential equations that can be viewed as a generalization of the heat equation where the temperature gradient is bounded but the heat flux is apriori only a measure. We consider this system in spatially periodic setting and use higher differentiability techniques to prove the existence and uniqueness of weak solution with integrable heat-flux for all values of the material parameter a. Under some more restrictive assumptions on a, we prove higher integrability of the heat flux. 1

Tesis por compendio / [ES] En esta tesis doctoral se propone y evalúa un sistema para detectar señales débiles (weak signals) relacionadas con cambios futuros trascendentales. Si bien la mayoría de las soluciones conocidas se basan en el uso de datos estructurados, el sistema propuesto detecta cuantitativamente estas señales utilizando información heterogénea y no estructurada de fuentes científicas, periodísticas y de redes sociales. La predicción de nuevas tendencias en un medio tiene muchas aplicaciones. Por ejemplo, empresas y startups se enfrentan a cambios constantes en sus mercados que son muy difíciles de predecir. Por esta razón, el desarrollo de sistemas para detectar automáticamente cambios futuros significativos en una etapa temprana es relevante para que cualquier organización tome decisiones acertadas a tiempo. Este trabajo ha sido diseñado para obtener señales débiles del futuro en cualquier campo dependiendo únicamente del conjunto de datos de entrada de documentos. Se aplican técnicas de minería de textos y procesamiento del lenguaje natural para procesar todos estos documentos. Como resultado, se obtiene un mapa con un ranking de términos, una lista de palabras clave clasificadas automáticamente y una lista de expresiones formadas por múltiples palabras. El sistema completo se ha probado en cuatro sectores diferentes: paneles solares, inteligencia artificial, sensores remotos e imágenes médicas. Este trabajo ha obtenido resultados prometedores, evaluados con dos metodologías diferentes. Como resultado, el sistema ha sido capaz de detectar de forma satisfactoria nuevas tendencias en etapas muy tempranas que se han vuelto cada vez más importantes en la actualidad. La computación cuántica es un nuevo paradigma para una multitud de aplicaciones informáticas. En esta tesis doctoral también se presenta un estudio de las tecnologías disponibles en la actualidad para la implementación física de qubits y puertas cuánticas, estableciendo sus principales ventajas y desventajas, y los marcos disponibles para la programación e implementación de circuitos cuánticos. Con el fin de mejorar la efectividad del sistema, se describe un diseño de un circuito cuántico basado en máquinas de vectores de soporte (SVM) para la resolución de problemas de clasificación. Este circuito está especialmente diseñado para los ruidosos procesadores cuánticos de escala intermedia (NISQ) que están disponibles actualmente. Como experimento, el circuito ha sido probado en un computador cuántico real basado en qubits superconductores por IBM como una mejora para el subsistema de minería de texto en la detección de señales débiles. Los resultados obtenidos con el experimento cuántico muestran también conclusiones interesantes y una mejora en el rendimiento de cerca del 20% sobre los sistemas convencionales, pero a su vez confirman que aún se requiere un desarrollo tecnológico continuo para aprovechar al máximo la computación cuántica. / [CA] En aquesta tesi doctoral es proposa i avalua un sistema per detectar senyals febles (weak signals) relacionats amb canvis futurs transcendentals. Si bé la majoria de solucions conegudes es basen en l'ús de dades estructurades, el sistema proposat detecta quantitativament aquests senyals utilitzant informació heterogènia i no estructurada de fonts científiques, periodístiques i de xarxes socials. La predicció de noves tendències en un medi té moltes aplicacions. Per exemple, empreses i startups s'enfronten a canvis constants als seus mercats que són molt difícils de predir. Per això, el desenvolupament de sistemes per detectar automàticament canvis futurs significatius en una etapa primerenca és rellevant perquè les organitzacions prenguen decisions encertades a temps. Aquest treball ha estat dissenyat per obtenir senyals febles del futur a qualsevol camp depenent únicament del conjunt de dades d'entrada de documents. S'hi apliquen tècniques de mineria de textos i processament del llenguatge natural per processar tots aquests documents. Com a resultat, s'obté un mapa amb un rànquing de termes, un llistat de paraules clau classificades automàticament i un llistat d'expressions formades per múltiples paraules. El sistema complet s'ha provat en quatre sectors diferents: panells solars, intel·ligència artificial, sensors remots i imatges mèdiques. Aquest treball ha obtingut resultats prometedors, avaluats amb dues metodologies diferents. Com a resultat, el sistema ha estat capaç de detectar de manera satisfactòria noves tendències en etapes molt primerenques que s'han tornat cada cop més importants actualment. La computació quàntica és un paradigma nou per a una multitud d'aplicacions informàtiques. En aquesta tesi doctoral també es presenta un estudi de les tecnologies disponibles actualment per a la implementació física de qubits i portes quàntiques, establint-ne els principals avantatges i desavantatges, i els marcs disponibles per a la programació i implementació de circuits quàntics. Per tal de millorar l'efectivitat del sistema, es descriu un disseny d'un circuit quàntic basat en màquines de vectors de suport (SVM) per resoldre problemes de classificació. Aquest circuit està dissenyat especialment per als sorollosos processadors quàntics d'escala intermèdia (NISQ) que estan disponibles actualment. Com a experiment, el circuit ha estat provat en un ordinador quàntic real basat en qubits superconductors per IBM com una millora per al subsistema de mineria de text. Els resultats obtinguts amb l'experiment quàntic també mostren conclusions interessants i una millora en el rendiment de prop del 20% sobre els sistemes convencionals, però a la vegada confirmen que encara es requereix un desenvolupament tecnològic continu per aprofitar al màxim la computació quàntica. / [EN] In this doctoral thesis, a system to detect weak signals related to future transcendental changes is proposed and tested. While most known solutions are based on the use of structured data, the proposed system quantitatively detects these signals using heterogeneous and unstructured information from scientific, journalistic, and social sources. Predicting new trends in an environment has many applications. For instance, companies and startups face constant changes in their markets that are very difficult to predict. For this reason, developing systems to automatically detect significant future changes at an early stage is relevant for any organization to make right decisions on time. This work has been designed to obtain weak signals of the future in any field depending only on the input dataset of documents. Text mining and natural language processing techniques are applied to process all these documents. As a result, a map of ranked terms, a list of automatically classified keywords and a list of multi-word expressions are obtained. The overall system has been tested in four different sectors: solar panels, artificial intelligence, remote sensing, and medical imaging. This work has obtained promising results that have been evaluated with two different methodologies. As a result, the system was able to successfully detect new trends at a very early stage that have become more and more important today. Quantum computing is a new paradigm for a multitude of computing applications. This doctoral thesis also presents a study of the technologies that are currently available for the physical implementation of qubits and quantum gates, establishing their main advantages and disadvantages and the available frameworks for programming and implementing quantum circuits. In order to improve the effectiveness of the system, a design of a quantum circuit based on support vector machines (SVMs) is described for the resolution of classification problems. This circuit is specially designed for the noisy intermediate-scale quantum (NISQ) computers that are currently available. As an experiment, the circuit has been tested on a real quantum computer based on superconducting qubits by IBM as an improvement for the text mining subsystem in the detection of weak signals. The results obtained with the quantum experiment show interesting outcomes with an improvement of close to 20% better performance than conventional systems, but also confirm that ongoing technological development is still required to take full advantage of quantum computing. / Griol Barres, I. (2022). Modelling of a System for the Detection of Weak Signals Through Text Mining and NLP. Proposal of Improvement by a Quantum Variational Circuit [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/183029 / TESIS / Compendio

[Background]: The advent of bidirectional encoder representation from trans- formers (BERT) language models (Devlin et al., 2018) and MS Marco, a large scale human-annotated dataset for machine reading comprehension (Bajaj et al., 2016) that made publicly available, led the field of information retrieval (IR) to experience a revolution (Lin et al., 2020). The retrieval model based on BERT of Nogueira and Cho (2019), by the time they published their paper, became the top entry in the MS Marco passage-reranking leaderboard, surpassing the previous state of the art by 27% in MRR@10. However, training such neural IR models for different domains than MS Marco is still hard because neural approaches often require a vast amount of training data to perform effectively, which is not always available. To address the problem of the shortage of labelled data a new line of research emerged, training neural models with weak supervision. In weak supervision, given an unlabelled dataset labels are generated automatically using an existing model and then a machine learning model is trained upon the artificial “weak“ data. In case of weak supervision for IR, the training dataset comes in the form of a tuple (query, passage). Dehghani et al. (2017) in their work used the AOL query logs (Pass et al., 2006), which is a set of millions of real web queries, and BM25 to retrieve the relevant passages for each of the user queries. A drawback with this approach is that it is hard to obtain query logs for every single different domain. [Objective]: This thesis proposes an intuitive approach for addressing the shortage of data in domains with limited or no data at all through transfer learning in the context of IR. We leverage Wikipedia’s structure for creating a Wikipedia-based generic IR training dataset for zero-shot neural models. [Method]: We create the “pseudo-queries“ by concatenating the titles of Wikipedia’s articles along with each of their title sections and we consider the associated section’s passage as the relevant passage of the pseudo-queries. All of our experiments are evaluated on a standard collection: MS Marco, which is a large scale web collection. For our zero-shot experiments, our proposed model, called “Wiki“, is a BERT model trained on the artificial Wikipedia-based dataset and the baseline is a default BERT model without any additional training. In our second line of experiments, we explore the benefits gained by pre-fine- tuning on the Wikipedia-based IR dataset and further fine-tuning on in-domain data. Our proposed model, "Wiki+Ma", is a BERT model pre-fine-tuned in the Wikipedia-based dataset and further fine-tuned in MS Marco, while the baseline is a BERT model fine-tuned only in MS Marco. [Results]: Results regarding our first experiments show that our BERT model trained on the Wikipedia-based IR dataset, called "Wiki", achieves a performance of 0.197 in MRR@10, which is about +10 points more in comparison to a BERT model with default weights; in addition, results in the development set indicate that the “Wiki“ model performs better than BERT model trained on in-domain data when the data is between 10k-50k instances. Results regarding our second line of experiments show that pre-fine-tuning on the Wikipedia-based IR dataset benefits later fine-tuning steps on in-domain data in terms of stability. [Conclusion]: Our findings suggest that transfer learning for IR tasks by leveraging the generic knowledge incorporated in Wikipedia is possible, though more experimentation is needed to understand its limitations in comparison with the traditional approaches such as the BM25.

The presented thesis aims to analyze the impact of social networking sites (SNS) in shaping our identity. Especially in the context of changes in everyday life caused by significant increase of SNS users. Unlike previous forms of online identity the identity of SNS users is anchored in the reality. Therefore our identity performance in this environment is closely tied to our real identity and real social network. Part of the thesis describe the influence of real social network on our lives, particularly the strong of weak ties is discussed. The analysis is focused on the processes and forms in which is our identity constructed and manifested within social networking sites. Detailed consideration is aimed on narrative and fragmentary characteristics of such identity. Also the consequences of total indexing of user-generated content and its persistent features are examined. This topic is followed by setting the framework of reputation management and protection of personal data. The findings of research present the forms of user identity on Facebook, the use of privacy settings and in particular the importance of the size of the networks of friends. The relationship between socialization in SNS and in the real world is described.

The current approach to assessing the educational quality applicable to assessing objects and processes formed and realized in producing spheres is widely spread. However, as education is a much more complicated anthropological, social and cultural object in comparison to that of production, the above mentioned approach is least effective. In education both \"strong\" and \"weak\" models are used. There do not exist measurement instruments for accurate assessing mild results. Self control, expert assessing method and portfolio are being put forward.

Im ersten Halbjahr 2016 wurde in Kooperation mit dem Martinsclub Bremen e.V. eine empirische Untersuchung zur funktionalen Varietät ' Leichte Sprache ' durchgeführt. Diese Ziele wurden verfolgt: 1) Für die ' Leichte Sprache ' wird u.a. vom ' Netzwerk Leichte Sprache ' reklamiert, für eine äußerst heterogene Zielgruppe von Nutzen zu sein. In der Untersuchung wird an Testitems, die z.T. erheblich gegen die ' Regeln ' für ' Leichte Sprache ' verstoßen, ermittelt werden, wie Gewährspersonen unterschiedlicher Zielgruppen (Menschen mit kognitiver Beeinträchtigung, Funktionale Analphabeten und Migranten) die ‚Schwierigkeit' und die ‚Akzeptabilität' der Items bewerten. 2) Die Regeln für ' Leichte Sprache ' sind (linguistisch) zu explizieren und auf die Bedürfnisse unterschiedlicher Zielgruppen zuzuschneiden. Beide Ziele wurden durch die Untersuchung für einzelne morphosyntaktische Phänomene erreicht. In diesem Artikel werden die Ergebnisse der Studie knapp vorgestellt und am Beispiel passivischer Strukturen des Deutschen im Detail diskutiert. Angrenzende Phänomenbereiche (sein-Perfekt, Kopulakonstruktionen mit sein und werden sowie werden mit futurischer und/oder epistemischer Lesart) werden nur angeschnitten.

We study the question of periodic homogenization of a variably scaled reaction-diffusion equation with non-linear drift of polynomial type. The non-linear drift was derived as hydrodynamic limit of a totally asymmetric simple exclusion process (TASEP) for a population of interacting particles crossing a domain with obstacle. We consider three different geometries: (i) Bounded domain crossed by a finitely thin flat composite layer; (ii) Bounded domain crossed by an infinitely thin flat composite layer; (iii) Unbounded composite domain.\end{itemize} For the thin layer cases, we consider our reaction-diffusion problem endowed with slow or moderate drift. Using energy-type estimates as well as concepts like thin-layer convergence and two-scale convergence, we derive homogenized evolution equations and the corresponding effective model parameters. Special attention is paid to the derivation of the effective transmission conditions across the separating limit interfaces. As a special scaling, the problem with large drift is treated separately for an unbounded composite domain. Because of the imposed large drift, this nonlinearity is expected to explode in the limit of a vanishing scaling parameter. To deal with this special case, we employ two-scale formal homogenization asymptotics with drift to derive the corresponding upscaled model equations as well as the structure of the effective transport tensors. Finally, we use Schauder's fixed point Theorem as well as monotonicity arguments to study the weak solvability of the upscaled model posed in the unbounded domain. This study wants to contribute with theoretical understanding needed when designing thin composite materials which are resistant to slow, moderate, and high velocity impacts. / We study the question of periodic homogenization of a variably scaled reaction-diffusion equation with non-linear drift of polynomial type. The non-linear drift was derived as hydrodynamic limit of a totally asymmetric simple exclusion process (TASEP) for a population of interacting particles crossing a domain with obstacle. We consider three different geometries: (i) Bounded domain crossed by a finitely thin composite layer; (ii) Bounded domain crossed by an infinitely thin composite  layer; (iii) Unbounded composite domain. For the thin layer cases, we consider our reaction-diffusion problem endowed with slow or moderate drift. Using energy-type  estimates, concepts like thin-layer convergence and two-scale convergence, we derive  homogenized  equations. Special attention is paid to the derivation of the effective transmission conditions across the separating limit interfaces. The problem with large drift is treated separately for an unbounded composite domain. Because of the imposed large drift, this nonlinearity is expected to explode in the limit of a vanishing scaling parameter.  This study wants to contribute with theoretical understanding needed when designing thin composite materials which are resistant to slow, moderate, and high velocity impacts.