Refine
Year of publication
- 2024 (216) (remove)
Document Type
- Doctoral Thesis (91)
- Article (36)
- Monograph/Edited Volume (26)
- Part of a Book (16)
- Other (11)
- Working Paper (9)
- Master's Thesis (8)
- Conference Proceeding (6)
- Part of Periodical (4)
- Postprint (2)
Is part of the Bibliography
- yes (216) (remove)
Keywords
- Judentum (5)
- Arctic (4)
- Arktis (4)
- experiment (4)
- machine learning (4)
- Christentum (3)
- Islam (3)
- Klimawandel (3)
- Kohlenstoff (3)
- Satzverarbeitung (3)
Institute
- Fachgruppe Politik- & Verwaltungswissenschaft (20)
- Extern (17)
- Historisches Institut (14)
- Institut für Biochemie und Biologie (14)
- Institut für Physik und Astronomie (13)
- Bürgerliches Recht (12)
- Fachgruppe Betriebswirtschaftslehre (11)
- Hasso-Plattner-Institut für Digital Engineering GmbH (11)
- Fachgruppe Soziologie (10)
- Fachgruppe Volkswirtschaftslehre (10)
Das in diesem Beitrag vorgestellte Projektseminarkonzept reagiert auf eine wahrgenommene Distanz und Unsicherheit Studierender im Fach Lebensgestaltung-Ethik-Religionskunde gegenüber religionsbezogenen Themen. Mittels verschiedener Strategien wurde, ausgehend von der Conceptual Change-Forschung, zur Wahrnehmung und Reflexion des eigenen kulturellen Standortes und der eigenen Konzepte in Bezug auf Religion(en) angeregt. Ihren Lernprozess haben die Studierenden in Arbeitsjournaleinträgen festgehalten. Diese Einträge wurden wiederum mittels einer qualitative Inhaltsanalyse untersucht. Nach der Darstellung der dabei erhobenen religions- und unterrichtsbezogenen Vorstellungen der Studierenden werden im Beitrag Anregungen gegeben, inwiefern die analysierten Befunde als Grundlage für die Verbesserung der Hochschullehre im Fachbereich dienen können.
Der Band präsentiert eine systematische Aufbereitung empirischer Befunde zum Lobbyismus in Deutschland und vermittelt, wie Lobbyist*innen, Entscheidungsträger*innen und institutionelle Rahmen miteinander interagieren. Untersucht werden politische Aktivitäten von sozialen Bewegungen, Verbänden, Unternehmen und Beratungsfirmen im Bundestag, der Bundesregierung und der Öffentlichkeit.
To manage tabular data files and leverage their content in a given downstream task, practitioners often design and execute complex transformation pipelines to prepare them. The complexity of such pipelines stems from different factors, including the nature of the preparation tasks, often exploratory or ad-hoc to specific datasets; the large repertory of tools, algorithms, and frameworks that practitioners need to master; and the volume, variety, and velocity of the files to be prepared. Metadata plays a fundamental role in reducing this complexity: characterizing a file assists end users in the design of data preprocessing pipelines, and furthermore paves the way for suggestion, automation, and optimization of data preparation tasks.
Previous research in the areas of data profiling, data integration, and data cleaning, has focused on extracting and characterizing metadata regarding the content of tabular data files, i.e., about the records and attributes of tables. Content metadata are useful for the latter stages of a preprocessing pipeline, e.g., error correction, duplicate detection, or value normalization, but they require a properly formed tabular input. Therefore, these metadata are not relevant for the early stages of a preparation pipeline, i.e., to correctly parse tables out of files. In this dissertation, we turn our focus to what we call the structure of a tabular data file, i.e., the set of characters within a file that do not represent data values but are required to parse and understand the content of the file. We provide three different approaches to represent file structure, an explicit representation based on context-free grammars; an implicit representation based on file-wise similarity; and a learned representation based on machine learning.
In our first contribution, we use the grammar-based representation to characterize a set of over 3000 real-world csv files and identify multiple structural issues that let files deviate from the csv standard, e.g., by having inconsistent delimiters or containing multiple tables. We leverage our learnings about real-world files and propose Pollock, a benchmark to test how well systems parse csv files that have a non-standard structure, without any previous preparation. We report on our experiments on using Pollock to evaluate the performance of 16 real-world data management systems.
Following, we characterize the structure of files implicitly, by defining a measure of structural similarity for file pairs. We design a novel algorithm to compute this measure, which is based on a graph representation of the files' content. We leverage this algorithm and propose Mondrian, a graphical system to assist users in identifying layout templates in a dataset, classes of files that have the same structure, and therefore can be prepared by applying the same preparation pipeline.
Finally, we introduce MaGRiTTE, a novel architecture that uses self-supervised learning to automatically learn structural representations of files in the form of vectorial embeddings at three different levels: cell level, row level, and file level. We experiment with the application of structural embeddings for several tasks, namely dialect detection, row classification, and data preparation efforts estimation.
Our experimental results show that structural metadata, either identified explicitly on parsing grammars, derived implicitly as file-wise similarity, or learned with the help of machine learning architectures, is fundamental to automate several tasks, to scale up preparation to large quantities of files, and to provide repeatable preparation pipelines.
Social institutions
(2024)
Social institutions are a system of behavioral and relationship patterns that are densely interwoven and enduring and function across an entire society. They order and structure the behavior of individuals in core areas of society and thus have a strong impact on the quality of life of individuals. Institutions regulate the following: (a) family and relationship networks carry out social reproduction and socialization; (b) institutions in the realm of education and training ensure the transmission and cultivation of knowledge, abilities, and specialized skills; (c) institutions in the labor market and economy provide for the production and distribution of goods and services; (d) institutions in the realm of law, governance, and politics provide for the maintenance of the social order; (e) while cultural, media, and religious institutions further the development of contexts of meaning, value orientations, and symbolic codes.
Comparamos la labialización no asimiladora de nasales finales en español en tres corpus de español americano (mexicano, colombiano y paraguayo). Si bien es conocida la labialización no asimiladora en español yucateco, es en gran parte desconocida en otras regiones de habla hispana, por lo que a menudo se atribuye a la influencia maya. Ahora bien, se han señalado casualmente hábitos de pronunciación similares tanto en Paraguay como en Colombia. Comparando empíricamente la labialización en tres corpus constituidos sobre la misma base metodológica, concluimos que la evidencia a favor del contacto lingüístico es como mucho sumamente indirecta. Independientemente de esto, encontramos que la diferencia más marcada es que la tasa de labialización parece ser determinada por la duración de la pausa subsiguiente en los datos de la península yucateca, mas no en aquellos de Colombia y Paraguay. Argumentamos que es cierto que el contacto puede eventualmente haber desencadenado el desarrollo de este rasgo en el español yucateco, puesto que el español actual casi no conoce nasales labiales finales, pero el maya sí. Sin embargo, el perfil lingüístico (hablantes monolingües vs. bilingües) no tiene ningún efecto en nuestros datos yucatecos y paraguayos, y en el total de nuestros datos tampoco encontramos evidencia en favor de la hipótesis que el contacto lingüístico hubiera jugado un rol (importante) en el desarrollo de las labiales nasales en las tres variedades.
Sexualität in der Geschichte
(2024)
Jelena Tomović führt in diesem Band durch die Entwicklungen unserer sexuellen Sprache und Praktiken. Sie zeigt, dass die Art und Weise, wie über Sexualität gesprochen wird, nicht nur ein Spiegelbild, sondern auch ein treibender Faktor für soziale Veränderungen ist. Die Studie stellt die konventionelle Vorstellung von Sexualität in Frage und führt die Lesenden in eine Welt der subtilen Nuancen und kulturellen Veränderungen. Mit kommunikationstheoretischen Ansätzen, dem praxeologischen Ansatz, ihrer sozialkonstruktivistischen Grundannahme und einem klaren Fokus auf Akteur*innen bietet die Autorin eine frische Perspektive auf die Geschichte der Sexualität. Das Buch eröffnet neue Wege für die Erforschung und das Verständnis von Intimität und sozialer Kommunikation.
Enhancing higher entrepreneurship education: insights from practitioners for curriculum improvement
(2024)
Curricula for higher entrepreneurship education should meet the requirements of both a solid theoretical foundation and a practical orientation. When these curricula are designed by education specialists, entrepreneurs are usually not consulted. To explore practitioners’ curricular recommendations, we conducted 73 semi-structured interviews with entrepreneurs with at least five years of professional experience. We collected 49 items for teaching and learning objectives, 37 for contents, 28 for teaching methods, and 17 for assessment methods. The respondents are convinced that students should acquire solid knowledge in business and management, legal issues, and entrepreneurship. For the latter, only some core aspects are provided. The entrepreneurs put greater emphasis on entrepreneurial skills and attitudes and consider experiential learning designs as most suitable, both in the secure setting of the classroom and in real life. The findings can help reflect on current entrepreneurship curriculum designs.
Deep learning has seen widespread application in many domains, mainly for its ability to learn data representations from raw input data. Nevertheless, its success has so far been coupled with the availability of large annotated (labelled) datasets. This is a requirement that is difficult to fulfil in several domains, such as in medical imaging. Annotation costs form a barrier in extending deep learning to clinically-relevant use cases. The labels associated with medical images are scarce, since the generation of expert annotations of multimodal patient data at scale is non-trivial, expensive, and time-consuming. This substantiates the need for algorithms that learn from the increasing amounts of unlabeled data. Self-supervised representation learning algorithms offer a pertinent solution, as they allow solving real-world (downstream) deep learning tasks with fewer annotations. Self-supervised approaches leverage unlabeled samples to acquire generic features about different concepts, enabling annotation-efficient downstream task solving subsequently.
Nevertheless, medical images present multiple unique and inherent challenges for existing self-supervised learning approaches, which we seek to address in this thesis: (i) medical images are multimodal, and their multiple modalities are heterogeneous in nature and imbalanced in quantities, e.g. MRI and CT; (ii) medical scans are multi-dimensional, often in 3D instead of 2D; (iii) disease patterns in medical scans are numerous and their incidence exhibits a long-tail distribution, so it is oftentimes essential to fuse knowledge from different data modalities, e.g. genomics or clinical data, to capture disease traits more comprehensively; (iv) Medical scans usually exhibit more uniform color density distributions, e.g. in dental X-Rays, than natural images. Our proposed self-supervised methods meet these challenges, besides significantly reducing the amounts of required annotations.
We evaluate our self-supervised methods on a wide array of medical imaging applications and tasks. Our experimental results demonstrate the obtained gains in both annotation-efficiency and performance; our proposed methods outperform many approaches from related literature. Additionally, in case of fusion with genetic modalities, our methods also allow for cross-modal interpretability. In this thesis, not only we show that self-supervised learning is capable of mitigating manual annotation costs, but also our proposed solutions demonstrate how to better utilize it in the medical imaging domain. Progress in self-supervised learning has the potential to extend deep learning algorithms application to clinical scenarios.
Organic solar cells (OSCs) represent a new generation of solar cells with a range of captivating attributes including low-cost, light-weight, aesthetically pleasing appearance, and flexibility. Different from traditional silicon solar cells, the photon-electron conversion in OSCs is usually accomplished in an active layer formed by blending two kinds of organic molecules (donor and acceptor) with different energy levels together.
The first part of this thesis focuses on a better understanding of the role of the energetic offset and each recombination channel on the performance of these low-offset OSCs. By combining advanced experimental techniques with optical and electrical simulation, the energetic offsets between CT and excitons, several important insights were achieved: 1. The short circuit current density and fill-factor of low-offset systems are largely determined by field-dependent charge generation in such low-offset OSCs. Interestingly, it is strongly evident that such field-dependent charge generation originates from a field-dependent exciton dissociation yield. 2. The reduced energetic offset was found to be accompanied by strongly enhanced bimolecular recombination coefficient, which cannot be explained solely by exciton repopulation from CT states. This implies the existence of another dark decay channel apart from CT.
The second focus of the thesis was on the technical perspective. In this thesis, the influence of optical artifacts in differential absorption spectroscopy upon the change of sample configuration and active layer thickness was studied. It is exemplified and discussed thoroughly and systematically in terms of optical simulations and experiments, how optical artifacts originated from non-uniform carrier profile and interference can manipulate not only the measured spectra, but also the decay dynamics in various measurement conditions. In the end of this study, a generalized methodology based on an inverse optical transfer matrix formalism was provided to correct the spectra and decay dynamics manipulated by optical artifacts.
Overall, this thesis paves the way for a deeper understanding of the keys toward higher PCEs in low-offset OSC devices, from the perspectives of both device physics and characterization techniques.