Filtern
Dokumenttyp
- Dissertation (13) (entfernen)
Sprache
- Englisch (7)
- Deutsch (5)
- Französisch (1)
Schlagworte
- Englisch (2)
- Japan (2)
- American (1)
- Artificial Intelligence (1)
- Ashikaga Yoshimitsu (1)
- Chinesen (1)
- Chinesisch-kanadische Literatur (1)
- Collexeme Analysis (1)
- Computation and Language (1)
- Computer Vision and Pattern Recognition (1)
Institut
- Fachbereich 2 (13) (entfernen)
Die Untersuchung verbindet Methoden der Korpuslinguistik und des close readings, um an einem repräsentativen Einzeltext mittlerer Länge das Verhältnis der syntaktischen und metrischen Ebene im mittelhochdeutschen Reimpaarvers zu untersuchen. Herausgearbeitet werden regelmäßig wiederkehrende Muster, die beide Ebenen stets gleich aufeinander abbilden. Diese Regelmäßigkeiten lassen sich aus den Lautstrukturen des mhd. Wortschatzes, den syntaktischen Bauplänen der Phrasen und Sätze, schließlich den Erfordernissen des metrischen Schemas erklären. Der häufig zur Erklärung herangezogene Reimzwang erweist sich bei näherer Betrachtung als eher sekundärer Einfluss auf die syntaktische Struktur. Neben typischen „Normalfällen“ bei denen sich statistisch häufige Betonungsmuster der Wörter, in üblichen, einfachen Satzstellungsmustern in immer gleicher Weise problemlos in den Reimpaarvers integrieren lassen, können auch wiederkehrende Abweichungsvarianten erklärt und beschrieben werden. Die festgestellten Regularitäten sind nur zu einem kleinen Teil und in wenigen Fällen deterministisch, es lässt sich jedoch, um die statistischen Auffälligkeiten zu begründen, zeigen, welche Vorteile sich aus bestimmten Varianten ergeben und welche Schwierigkeiten bei anderen entstehen, wie sich eine Variante durch eine andere ersetzen lässt. Beschrieben wird so der Gestaltungsraum des Dichters und die von ihm gewählten Lösungen. Indirekt ergibt sich zugleich ein Negativbild der Syntax, die den Zwängen des metrischen Schemas nicht unterworfen ist.
While humans find it easy to process visual information from the real world, machines struggle with this task due to the unstructured and complex nature of the information. Computer vision (CV) is the approach of artificial intelligence that attempts to automatically analyze, interpret, and extract such information. Recent CV approaches mainly use deep learning (DL) due to its very high accuracy. DL extracts useful features from unstructured images in a training dataset to use them for specific real-world tasks. However, DL requires a large number of parameters, computational power, and meaningful training data, which can be noisy, sparse, and incomplete for specific domains. Furthermore, DL tends to learn correlations from the training data that do not occur in reality, making DNNs poorly generalizable and error-prone.
Therefore, the field of visual transfer learning is seeking methods that are less dependent on training data and are thus more applicable in the constantly changing world. One idea is to enrich DL with prior knowledge. Knowledge graphs (KG) serve as a powerful tool for this purpose because they can formalize and organize prior knowledge based on an underlying ontological schema. They contain symbolic operations such as logic, rules, and reasoning, and can be created, adapted, and interpreted by domain experts. Due to the abstraction potential of symbols, KGs provide good prerequisites for generalizing their knowledge. To take advantage of the generalization properties of KG and the ability of DL to learn from large-scale unstructured data, attempts have long been made to combine explicit graph and implicit vector representations. However, with the recent development of knowledge graph embedding methods, where a graph is transferred into a vector space, new perspectives for a combination in vector space are opening up.
In this work, we attempt to combine prior knowledge from a KG with DL to improve visual transfer learning using the following steps: First, we explore the potential benefits of using prior knowledge encoded in a KG for DL-based visual transfer learning. Second, we investigate approaches that already combine KG and DL and create a categorization based on their general idea of knowledge integration. Third, we propose a novel method for the specific category of using the knowledge graph as a trainer, where a DNN is trained to adapt to a representation given by prior knowledge of a KG. Fourth, we extend the proposed method by extracting relevant context in the form of a subgraph of the KG to investigate the relationship between prior knowledge and performance on a specific CV task. In summary, this work provides deep insights into the combination of KG and DL, with the goal of making DL approaches more generalizable, more efficient, and more interpretable through prior knowledge.
In her poems, Tawada constructs liminal speaking subjects – voices from the in-between – which disrupt entrenched binary thought processes. Synthesising relevant concepts from theories of such diverse fields as lyricology, performance studies, border studies, cultural and postcolonial studies, I develop ‘voice’ and ‘in-between space’ as the frameworks to approach Tawada’s multifaceted poetic output, from which I have chosen 29 poems and two verse novels for analysis. Based on the body speaking/writing, sensuality is central to Tawada’s use of voice, whereas the in-between space of cultures and languages serves as the basis for the liminal ‘exophonic’ voices in her work. In the context of cultural alterity, Tawada focuses on the function of language, both its effect on the body and its role in subject construction, while her feminist poetry follows the general development of feminist academia from emancipation to embodiment to queer representation. Her response to and transformation of écriture féminine in her verse novels transcends the concept of the body as the basis of identity, moving to literary and linguistic, plural self-construction instead. While few poems are overtly political, the speaker’s personal and contextual involvement in issues of social conflict reveal the poems’ potential to speak of, and to, the multiply identified citizens of a globalised world, who constantly negotiate physical as well as psychological borders.