University of Hildesheim
Not a member yet
1144 research outputs found
Sort by
Few-Shot Meta-Learning in Heterogeneous Contexts
No full textFew-shot learning has become an effective technique for helping parametric models improve performance on various tasks by leveraging knowledge from similar datasets. However, existing few-shot learning approaches are limited in their application to specific data types and tasks. This thesis addresses the challenges of few-shot learning in different domains: unstructured data, time series forecasting, and human motion prediction.
The first part of the thesis introduces Chameleon, a novel model that enables meta-learning weight initialization across tasks with different schemas in unstructured data. Chameleon learns to align predictor schemas to a common representation, allowing for successful parameter initializations across diverse tasks. Experimental results on various datasets demonstrate the effectiveness of Chameleon in achieving few-shot classification on unstructured data.
In the second part, the thesis presents HIDRA, a meta-learning approach designed to tackle the problem of few-shot learning in time series forecasting with varying numbers of target variables. HIDRA employs a master neuron to initialize output neurons for new tasks, providing a flexible and robust solution for training and evaluating across datasets with different target variable counts. Extensive experiments validate the superiority of HIDRA over standard approaches, especially in low-capacity models and complex tasks with a high number of classes.
In the third part of the thesis, we present TimeHetNet. We formalize the problem of few-shot time series forecasting with heterogeneous channels for the first time. Extending recent work on heterogeneous attributes in vector data, we develop a model composed of permutation-invariant deep set blocks incorporating a temporal embedding. We assemble the first meta-dataset of 40 multivariate time-series datasets and show through experiments that our model provides a good generalization, outperforming baselines from more straightforward scenarios that either fail to learn across tasks or miss temporal information.
The final part of the thesis focuses on few-shot motion prediction, a challenging task of predicting future sensor readings of motion body sensors. To address this, the thesis proposes a novel extension of the TimeHetNet model incorporating graph neural networks and deep set blocks to explicitly include spatial graph information while generalizing motion tasks with heterogeneous sensors. Experimental evaluations on various motion tasks showcase significant performance improvements compared to state-of-the-art models, demonstrating the model's effectiveness in capturing spatial and temporal information.
Overall, this thesis contributes to the advancement of few-shot learning techniques in diverse domains, providing valuable insights and innovative solutions to enhance the performance of parametric models in scenarios with limited labeled data
Explainable Case-Based Reasoning: Rapid Prototyping of Domain Independent Explainable Case-Based Reasoning Applications
No full textThis dissertation investigates the explainability competence of case-based reasoning systems and enriches developers of CBR applications with a flexible extension of their already existing CBR applications by an additional component, which enables existing CBR applications to explain, how they reached their diagnosis result. For doing so, an approach presented by Jean-Baptiste Lamy et al. [83] has been implemented in Java and has been successfully tested in the domain of “Intrusion Detection Systems” as well as the domain of “Blood Alcohol Level Prediction”. The extension has been developed in the context of the myCBR framework for the development of Java-based CBR applications and, thus, been published publicly using an open-source license. A user study with 38 participants showed a good acceptance rate – independent of the participants’ experiences on artificial intelligence or the general stance towards artificial intelligence
Tales of Math Island The Tri-Island Area of Number Theory, Graph Theory, and Games
No full textThis Habilitation thesis covers twelve articles in the areas number theory, graph theory, and games. The main part, namely the content of eight of the twelve articles, is about peg solitaire on graphs. The classical game of peg solitaire has been known for many centuries, its true origins are unknown. In this game, the player is given a board of cells, each of which a peg can be in. A jump is done by taking a peg, jumping over another peg into an empty cell, removing the peg over which it jumped. This can be generalized on graphs as follows: For an (undirected) graph G, we put pegs in some of the vertices of G. Given three vertices u,v,w with pegs in u and v and a hole in w such that uv as well vw are adjacent, we can jump with the peg from u over v into w, removing the peg in v. The goal of the game is to remove as much pegs as possible via such jumps. Several questions are discussed, such as the minimal number of remaining pegs for special graphs or graph operations. We also present a new method for peg solitaire on graphs, pagoda funtions, that can be used to give a bound on the minimal number of remaining pegs
Chang Dobin. History of Korean Thought (1925)
No full textIn 1925, during the Japanese rule, the Korean historian and resistance activist Chang Dobin (1888-1963) wrote the first history of Korean philosophy in the Korean language, History of Korean Thought. For Chang, the history of philosophy is a powerful tool for nation-building and anti-colonial resistance: the Korean people should be freed from the sinocentric burden of the past, in order to develop a form of historical subjectivity suitable for a modern nation-state. Chang constructs a linear narrative from prehistory to the 19th century, creating a prototype of Korean intellectual history as a dialectical competition between the Korean and the cultural Other
Informationskompetenz und Künstliche Intelligenz: Konzepte, Herausforderungen und Chancen Ein Beitrag der Fachgruppe Informationskompetenz der KIBA
No full textAm 20. September 2024 organisierte die Fachgruppe Informationswissenschaft der KIBA an der Universität Hildesheim einen Workshop zum Thema „Informationskompetenz und Künstliche Intelligenz“.
Ziel des Workshops war es, das Konzept der Informationskompetenz im Kontext der neuen Herausforderungen und Chancen, die mit der rasanten Verbreitung von Künstlicher Intelligenz (insbesondere generativer Künstlicher Intelligenz seit Ende 2022) einhergehen, aus einer forschungsbezogenen Perspektive zu beleuchten und zu diskutieren.
13 Expert:innen, die zuvor ein Positionspapier zu diesem Thema verfasst hatten, nahmen am Workshop teil. Die Positionspapiere und die Workshop-Diskussion konzentrierten sich auf drei zentrale Fragestellungen:
1. Welche Auswirkungen hat Künstliche Intelligenz auf bestehende Konzepte der Informationskompetenz?
2. Welche neuen Fragestellungen ergeben sich aus Künstlicher Intelligenz für informationswissenschaftliche Forschung im Bereich Informationskompetenz?
3. Welche Herausforderungen und Chancen ergeben sich in der Förderung von Informationskompetenz durch Künstliche Intelligenz?
Dieser Band präsentiert die Positionspapiere sowie die zentralen Ergebnisse des Workshops
"Für Gott, die Kirche und das Vaterland" Spiritualität und Pädagogik der Jesuiten im Spiegel des Gymnasiums Mariano-Josephinum in Hildesheim (1595-1773)
No full textIn der Bischofsstadt Hildesheim erneuerten Jesuiten angesichts der Reformation seit 1595 die mittelalterliche Domschule im Zeichen eines christlichen Humanismus katholischer Prägung. Bis heute sind die Bibliotheks- und Archivbestände nahezu vollständig erhalten. Auf dieser Quellenbasis stellt die vorliegende Studie exemplarisch den Zusammenhang von Spiritualität und Pädagogik dar, wie er für das Bildungsapostolat der Jesuiten bis zur Ordensaufhebung 1773 typisch war
Investigation of The Psychological Wellbeing after a One-Week-Skiing Intervention with Childhood Cancer Survivors
No full textAfter a child’s cancer diagnosis, the life of the whole family changes immediately. Besides physical exposure due to both the disease and the treatment, psychological stress, anxiety and uncertainties arise for all family members. In a family-oriented winter sport rehabilitation camp in the alps the increased physical fitness should be focused besides the adventure and the psychological wellbeing, thus basically the path back to normality.
In order to investigate the latter, the parents of eight patients (10.87 yrs. ± 5.33 yrs.) were surveyed before and after the journey via the Strength and Difficulties Questionnaire (SDQ) consisting of 25 items divided into five subscales (social behavior, hyperactivity, emotional problems, problems with peer groups, behavioral problems).
It was pointed out that for the given setting open answer options were more meaningful than closed answers. There was only one significance for the subscale behavioral problems. However, the parents stated distinctly positive effects regarding the social togetherness of the children, the atmosphere within the families, the independence and self- confidence of the children. According to the participants the rehabilitation journey could be recommended not only for the improvement of motor parameters and physical effects, but also on a psychological level and for an increased perceived health-related life qualit
Fachkräfte der Einrichtungsaufsicht und -beratung im Kontakt mit jungen Menschen Ein Forschungsprojekt zur Qualitätsentwicklung von Beteiligungs- und Beschwerderechten in Einrichtungen der Kinder- und Jugendhilfe
No full textDieser Reflexionsbogen wurde im Rahmen eines Forschungsprojektes gemeinsam mit Fachkräften der Einrichtungsaufsicht und -beratung und jungen Menschen erarbeitet. Er dient Fachkräften der betriebs-erlaubniserteilenden Behörden in ihrer Beratungstätigkeit, soll Einrichtungen bei der Erstellung und Überarbeitung von Beteiligungs- und Beschwerdekonzepten gemäß § 45 SGB VIII unterstützen und Einrichtungen einladen, ihre Beteiligungs- und Beschwerdekonzepte regelmäßig und reflexiv mit allen Beteiligten - insbesondere junge Menschen - weiterzuentwickeln.
Die Reflexionsfragen sollen zum Denken anregen und können je nach Bedarf, Zielgruppen junger Menschen und Rahmenbedingungen der Einrichtung angepasst werden. Fachkräfte der betriebserlaubniserteilenden Behörden können den Bogen in Gesprächen mit Trägern nutzen oder ihn zur internen Nutzung - zum Beispiel für Teamsitzungen, Fachkonferenzen ö. ä. - zur Verfügung stellen. Es lohnt sich auch, einzelne Fragen mit jungen Menschen zu besprechen, zum Beispiel in Gruppensitzungen oder Sitzungen von Kinder- und Jugendvertretungen