@phdthesis{Wick2020,
  author    = {Wick, Christoph},
  title     = {Optical Medieval Music Recognition},
  doi       = {10.25972/OPUS-21434},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-214348},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2020},
  abstract  = {In recent years, great progress has been made in the area of Artificial Intelligence (AI) due to the possibilities of Deep Learning which steadily yielded new state-of-the-art results especially in many image recognition tasks. Currently, in some areas, human performance is achieved or already exceeded. This great development already had an impact on the area of Optical Music Recognition (OMR) as several novel methods relying on Deep Learning succeeded in specific tasks. Musicologists are interested in large-scale musical analysis and in publishing digital transcriptions in a collection enabling to develop tools for searching and data retrieving. The application of OMR promises to simplify and thus speed-up the transcription process by either providing fully-automatic or semi-automatic approaches. This thesis focuses on the automatic transcription of Medieval music with a focus on square notation which poses a challenging task due to complex layouts, highly varying handwritten notations, and degradation. However, since handwritten music notations are quite complex to read, even for an experienced musicologist, it is to be expected that even with new techniques of OMR manual corrections are required to obtain the transcriptions. This thesis presents several new approaches and open source software solutions for layout analysis and Automatic Text Recognition (ATR) for early documents and for OMR of Medieval manuscripts providing state-of-the-art technology. Fully Convolutional Networks (FCN) are applied for the segmentation of historical manuscripts and early printed books, to detect staff lines, and to recognize neume notations. The ATR engine Calamari is presented which allows for ATR of early prints and also the recognition of lyrics. Configurable CNN/LSTM-network architectures which are trained with the segmentation-free CTC-loss are applied to the sequential recognition of text but also monophonic music. Finally, a syllable-to-neume assignment algorithm is presented which represents the final step to obtain a complete transcription of the music. The evaluations show that the performances of any algorithm is highly depending on the material at hand and the number of training instances. The presented staff line detection correctly identifies staff lines and staves with an \$F_1\$-score of above \$99.5\\%\$. The symbol recognition yields a diplomatic Symbol Accuracy Rate (dSAR) of above \$90\\%\$ by counting the number of correct predictions in the symbols sequence normalized by its length. The ATR of lyrics achieved a Character Error Rate (CAR) (equivalently the number of correct predictions normalized by the sentence length) of above \$93\\%\$ trained on 771 lyric lines of Medieval manuscripts and of 99.89\\% when training on around 3.5 million lines of contemporary printed fonts. The assignment of syllables and their corresponding neumes reached \$F_1\$-scores of up to \$99.2\\%\$. A direct comparison to previously published performances is difficult due to different materials and metrics. However, estimations show that the reported values of this thesis exceed the state-of-the-art in the area of square notation. A further goal of this thesis is to enable musicologists without technical background to apply the developed algorithms in a complete workflow by providing a user-friendly and comfortable Graphical User Interface (GUI) encapsulating the technical details. For this purpose, this thesis presents the web-application OMMR4all. Its fully-functional workflow includes the proposed state-of-the-art machine-learning algorithms and optionally allows for a manual intervention at any stage to correct the output preventing error propagation. To simplify the manual (post-) correction, OMMR4all provides an overlay-editor that superimposes the annotations with a scan of the original manuscripts so that errors can easily be spotted. The workflow is designed to be iteratively improvable by training better models as soon as new Ground Truth (GT) is available.},
  subject      = {Neumenschrift},
  language  = {en}
}
@phdthesis{Schoeneberg2014,
  author    = {Sch{\"o}neberg, Hendrik},
  title     = {Semiautomatische Metadaten-Extraktion und Qualit{\"a}tsmanagement in Workflow-Systemen zur Digitalisierung historischer Dokumente},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-104878},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2014},
  abstract  = {Performing Named Entity Recognition on ancient documents is a time-consuming, complex and error-prone manual task. It is a prerequisite though to being able to identify related documents and correlate between named entities in distinct sources, helping to precisely recreate historic events. In order to reduce the manual effort, automated classification approaches could be leveraged. Classifying terms in ancient documents in an automated manner poses a difficult task due to the sources' challenging syntax and poor conservation states. This thesis introduces and evaluates approaches that can cope with complex syntactial environments by using statistical information derived from a term's context and combining it with domain-specific heuristic knowledge to perform a classification. Furthermore this thesis demonstrates how metadata generated by these approaches can be used as error heuristics to greatly improve the performance of workflow systems for digitizations of early documents.},
  subject      = {Klassifikation},
  language  = {de}
}
@phdthesis{Koeberle2021,
  author    = {K{\"o}berle, Philipp},
  title     = {High-resolution ultrasound for the identification of pathological patterns in patients with polyneuropathies and amyotrophic lateral sclerosis},
  doi       = {10.25972/OPUS-24580},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-245800},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2021},
  abstract  = {Neuropathies are a group of potentially treatable diseases with an often disabling and restricting course. Amyotrophic lateral sclerosis (ALS) is a lethal disease without causal treatment possibilities. The objective of this study was to examine the diagnostic utility of HRUS for the differentiation of subtypes of axonal and demyelinating neuropathies and to investigate its utility for the sonological differentiation of ALS. The hypothetical statement that neuropathy causes enlargement of peripheral nerves compared to healthy controls proved to be right, but the adjunctive assumption that ALS does not cause enlargement of peripheral nerves proved to be wrong - in patients with ALS slight enlargement of peripheral nerves was visible as well. The statement that nerve enlargement can be detected by measurement of the cross-sectional area (CSA) and the longitudinal diameter (LD) with comparable results proved to be right, but the enlargement was slightly less present by measurement of the LD. The statement that axonal and demyelinating neuropathies show distinct patterns of nerve enlargement must be answered differentiated: The comparison between axonal and demyelinating neuropathies showed a stronger nerve enlargement in patients with demyelinating neuropathies than in patients with axonal neuropathies at proximal nerve segments of upper extremities. In the comparison of diagnose-defined subgroups of inflammatory demyelinating neuropathies a respective specific pattern of nerve enlargement was visible. However, remarkable in this context was the strong nerve enlargement found in patients with NSVN, which is classified as an axonal neuropathy. Stratification for specific findings in nerve biopsy did not lead to constructive differences in comparison between the different groups. To sum up, HRUS showed to provide a useful contribution in the diagnostic process of neuropathies and ALS but needs to be integrated in a multimodal diagnostic approach.},
  subject      = {Polyneuropathie},
  language  = {en}
}
@phdthesis{Kluegl2015,
  author    = {Kl{\"u}gl, Peter},
  title     = {Context-specific Consistencies in Information Extraction: Rule-based and Probabilistic Approaches},
  publisher = {W{\"u}rzburg University Press},
  address   = {W{\"u}rzburg},
  isbn      = {978-3-95826-018-4 (print)},
  doi       = {10.25972/WUP-978-3-95826-019-1},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-108352},
  school      = {W{\"u}rzburg University Press},
  year      = {2015},
  abstract  = {Large amounts of communication, documentation as well as knowledge and information are stored in textual documents. Most often, these texts like webpages, books, tweets or reports are only available in an unstructured representation since they are created and interpreted by humans. In order to take advantage of this huge amount of concealed information and to include it in analytic processes, it needs to be transformed into a structured representation. Information extraction considers exactly this task. It tries to identify well-defined entities and relations in unstructured data and especially in textual documents. Interesting entities are often consistently structured within a certain context, especially in semi-structured texts. However, their actual composition varies and is possibly inconsistent among different contexts. Information extraction models stay behind their potential and return inferior results if they do not consider these consistencies during processing. This work presents a selection of practical and novel approaches for exploiting these context-specific consistencies in information extraction tasks. The approaches direct their attention not only to one technique, but are based on handcrafted rules as well as probabilistic models. A new rule-based system called UIMA Ruta has been developed in order to provide optimal conditions for rule engineers. This system consists of a compact rule language with a high expressiveness and strong development support. Both elements facilitate rapid development of information extraction applications and improve the general engineering experience, which reduces the necessary efforts and costs when specifying rules. The advantages and applicability of UIMA Ruta for exploiting context-specific consistencies are illustrated in three case studies. They utilize different engineering approaches for including the consistencies in the information extraction task. Either the recall is increased by finding additional entities with similar composition, or the precision is improved by filtering inconsistent entities. Furthermore, another case study highlights how transformation-based approaches are able to correct preliminary entities using the knowledge about the occurring consistencies. The approaches of this work based on machine learning rely on Conditional Random Fields, popular probabilistic graphical models for sequence labeling. They take advantage of a consistency model, which is automatically induced during processing the document. The approach based on stacked graphical models utilizes the learnt descriptions as feature functions that have a static meaning for the model, but change their actual function for each document. The other two models extend the graph structure with additional factors dependent on the learnt model of consistency. They include feature functions for consistent and inconsistent entities as well as for additional positions that fulfill the consistencies. The presented approaches are evaluated in three real-world domains: segmentation of scientific references, template extraction in curricula vitae, and identification and categorization of sections in clinical discharge letters. They are able to achieve remarkable results and provide an error reduction of up to 30\% compared to usually applied techniques.},
  subject      = {Information Extraction},
  language  = {en}
}
@phdthesis{Reutelshoefer2014,
  author    = {Reutelsh{\"o}fer, Jochen},
  title     = {A Meta-Engineering Approach for Document-Centered Knowledge Acquisition},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-107523},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2014},
  abstract  = {Today knowledge base authoring for the engineering of intelligent systems is performed mainly by using tools with graphical user interfaces. An alternative human-computer interaction para- digm is the maintenance and manipulation of electronic documents, which provides several ad- vantages with respect to the social aspects of knowledge acquisition. Until today it hardly has found any attention as a method for knowledge engineering. This thesis provides a comprehensive discussion of document-centered knowledge acquisition with knowledge markup languages. There, electronic documents are edited by the knowledge authors and the executable knowledge base entities are captured by markup language expressions within the documents. The analysis of this approach reveals significant advantages as well as new challenges when compared to the use of traditional GUI-based tools. Some advantages of the approach are the low barriers for domain expert participation, the simple integration of informal descriptions, and the possibility of incremental knowledge for- malization. It therefore provides good conditions for building up a knowledge acquisition pro- cess based on the mixed-initiative strategy, being a flexible combination of direct and indirect knowledge acquisition. Further it turns out that document-centered knowledge acquisition with knowledge markup languages provides high potential for creating customized knowledge au- thoring environments, tailored to the needs of the current knowledge engineering project and its participants. The thesis derives a process model to optimally exploit this customization po- tential, evolving a project specific authoring environment by an agile process on the meta level. This meta-engineering process continuously refines the three aspects of the document space: The employed markup languages, the scope of the informal knowledge, and the structuring and organization of the documents. The evolution of the first aspect, the markup languages, plays a key role, implying the design of project specific markup languages that are easily understood by the knowledge authors and that are suitable to capture the required formal knowledge precisely. The goal of the meta-engineering process is to create a knowledge authoring environment, where structure and presentation of the domain knowledge comply well to the users' mental model of the domain. In that way, the approach can help to ease major issues of knowledge-based system development, such as high initial development costs and long-term maintenance problems. In practice, the application of the meta-engineering approach for document-centered knowl- edge acquisition poses several technical challenges that need to be coped with by appropriate tool support. In this thesis KnowWE, an extensible document-centered knowledge acquisition environment is presented. The system is designed to support the technical tasks implied by the meta-engineering approach, as for instance design and implementation of new markup lan- guages, content refactoring, and authoring support. It is used to evaluate the approach in several real-world case-studies from different domains, such as medicine or engineering for instance. We end the thesis by a summary and point out further interesting research questions consid- ering the document-centered knowledge acquisition approach.},
  subject      = {Wissenstechnik},
  language  = {en}
}
@phdthesis{Ifland2014,
  author    = {Ifland, Marianus},
  title     = {Feedback-Generierung f{\"u}r offene, strukturierte Aufgaben in E-Learning-Systemen},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-106348},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2014},
  abstract  = {Bei Lernprozessen spielt das Anwenden der zu erlernenden T{\"a}tigkeit eine wichtige Rolle. Im Kontext der Ausbildung an Schulen und Hochschulen bedeutet dies, dass es wichtig ist, Sch{\"u}lern und Studierenden ausreichend viele {\"U}bungsm{\"o}glichkeiten anzubieten. Die von Lehrpersonal bei einer "Korrektur" erstellte R{\"u}ckmeldung, auch Feedback genannt, ist jedoch teuer, da der zeitliche Aufwand je nach Art der Aufgabe betr{\"a}chtlich ist. Eine L{\"o}sung dieser Problematik stellen E-Learning-Systeme dar. Geeignete Systeme k{\"o}nnen nicht nur Lernstoff pr{\"a}sentieren, sondern auch {\"U}bungsaufgaben anbieten und nach deren Bearbeitung quasi unmittelbar entsprechendes Feedback generieren. Es ist jedoch im Allgemeinen nicht einfach, maschinelle Verfahren zu implementieren, die Bearbeitungen von {\"U}bungsaufgaben korrigieren und entsprechendes Feedback erstellen. F{\"u}r einige Aufgabentypen, wie beispielsweise Multiple-Choice-Aufgaben, ist dies zwar trivial, doch sind diese vor allem dazu gut geeignet, sogenanntes Faktenwissen abzupr{\"u}fen. Das Ein{\"u}ben von Lernzielen im Bereich der Anwendung ist damit kaum m{\"o}glich. Die Behandlung dieser nach g{\"a}ngigen Taxonomien h{\"o}heren kognitiven Lernziele erlauben sogenannte offene Aufgabentypen, deren Bearbeitung meist durch die Erstellung eines Freitexts in nat{\"u}rlicher Sprache erfolgt. Die Information bzw. das Wissen, das Lernende eingeben, liegt hier also in sogenannter „unstrukturierter" Form vor. Dieses unstrukturierte Wissen ist maschinell nur schwer verwertbar, sodass sich Trainingssysteme, die Aufgaben dieser Art stellen und entsprechende R{\"u}ckmeldung geben, bisher nicht durchgesetzt haben. Es existieren jedoch auch offene Aufgabentypen, bei denen Lernende das Wissen in strukturierter Form eingeben, so dass es maschinell leichter zu verwerten ist. F{\"u}r Aufgaben dieser Art lassen sich somit Trainingssysteme erstellen, die eine gute M{\"o}glichkeit darstellen, Sch{\"u}lern und Studierenden auch f{\"u}r praxisnahe Anwendungen viele {\"U}bungsm{\"o}glichkeiten zur Verf{\"u}gung zu stellen, ohne das Lehrpersonal zus{\"a}tzlich zu belasten. In dieser Arbeit wird beschrieben, wie bestimmte Eigenschaften von Aufgaben ausgenutzt werden, um entsprechende Trainingssysteme konzipieren und implementieren zu k{\"o}nnen. Es handelt sich dabei um Aufgaben, deren L{\"o}sungen strukturiert und maschinell interpretierbar sind. Im Hauptteil der Arbeit werden vier Trainingssysteme bzw. deren Komponenten beschrieben und es wird von den Erfahrungen mit deren Einsatz in der Praxis berichtet: Eine Komponente des Trainingssystems „CaseTrain" kann Feedback zu UML Klassendiagrammen erzeugen. Das neuartige Trainingssystem „WARP" generiert zu UML Aktivit{\"a}tsdiagrammen Feedback in mehreren Ebenen, u.a. indem es das durch Aktivit{\"a}tsdiagramme definierte Verhalten von Robotern in virtuellen Umgebungen visualisiert. Mit „{\"U}PS" steht ein Trainingssystem zur Verf{\"u}gung, mit welchem die Eingabe von SQL-Anfragen einge{\"u}bt werden kann. Eine weitere in „CaseTrain" implementierte Komponente f{\"u}r Bildmarkierungsaufgaben erm{\"o}glicht eine unmittelbare, automatische Bewertung entsprechender Aufgaben. Die Systeme wurden im Zeitraum zwischen 2011 und 2014 an der Universit{\"a}t W{\"u}rzburg in Vorlesungen mit bis zu 300 Studierenden eingesetzt und evaluiert. Die Evaluierung ergab eine hohe Nutzung und eine gute Bewertung der Studierenden der eingesetzten Konzepte, womit belegt wurde, dass elektronische Trainingssysteme f{\"u}r offene Aufgaben in der Praxis eingesetzt werden k{\"o}nnen.},
  subject      = {E-Learning},
  language  = {de}
}
@phdthesis{Freiberg2015,
  author    = {Freiberg, Martina},
  title     = {UI-, User-, \& Usability-Oriented Engineering of Participative Knowledge-Based Systems},
  publisher = {W{\"u}rzburg University Press},
  isbn      = {978-3-95826-012-2 (print)},
  doi       = {10.25972/WUP-978-3-95826-013-9},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-106072},
  school      = {W{\"u}rzburg University Press},
  pages     = {232},
  year      = {2015},
  abstract  = {Knowledge-based systems (KBS) face an ever-increasing interest in various disciplines and contexts. Yet, the former aim to construct the 'perfect intelligent software' continuously shifts to user-centered, participative solutions. Such systems enable users to contribute their personal knowledge to the problem solving process for increased efficiency and an ameliorated user experience. More precisely, we define non-functional key requirements of participative KBS as: Transparency (encompassing KBS status mediation), configurability (user adaptability, degree of user control/exploration), quality of the KB and UI, and evolvability (enabling the KBS to grow mature with their users). Many of those requirements depend on the respective target users, thus calling for a more user-centered development. Often, also highly expertise domains are targeted — inducing highly complex KBs — which requires a more careful and considerate UI/interaction design. Still, current KBS engineering (KBSE) approaches mostly focus on knowledge acquisition (KA) This often leads to non-optimal, little reusable, and non/little evaluated KBS front-end solutions. In this thesis we propose a more encompassing KBSE approach. Due to the strong mutual influences between KB and UI, we suggest a novel form of intertwined UI and KB development. We base the approach on three core components for encompassing KBSE: (1) Extensible prototyping, a tailored form of evolutionary prototyping; this builds on mature UI prototypes and offers two extension steps for the anytime creation of core KBS prototypes (KB + core UI) and fully productive KBS (core KBS prototype + common framing functionality). (2) KBS UI patterns, that define reusable solutions for the core KBS UI/interaction; we provide a basic collection of such patterns in this work. (3) Suitable usability instruments for the assessment of the KBS artifacts. Therewith, we do not strive for 'yet another' self-contained KBS engineering methodology. Rather, we motivate to extend existing approaches by the proposed key components. We demonstrate this based on an agile KBSE model. For practical support, we introduce the tailored KBSE tool ProKEt. ProKEt offers a basic selection of KBS core UI patterns and corresponding configuration options out of the box; their further adaption/extension is possible on various levels of expertise. For practical usability support, ProKEt offers facilities for quantitative and qualitative data collection. ProKEt explicitly fosters the suggested, intertwined development of UI and KB. For seamlessly integrating KA activities, it provides extension points for two selected external KA tools: For KnowOF, a standard office based KA environment. And for KnowWE, a semantic wiki for collaborative KA. Therewith, ProKEt offers powerful support for encompassing, user-centered KBSE. Finally, based on the approach and the tool, we also developed a novel KBS type: Clarification KBS as a mashup of consultation and justification KBS modules. Those denote a specifically suitable realization for participative KBS in highly expertise contexts and consequently require a specific design. In this thesis, apart from more common UI solutions, we particularly also introduce KBS UI patterns especially tailored towards Clarification KBS.},
  subject      = {Wissensbasiertes System},
  language  = {en}
}
@phdthesis{Lemmerich2014,
  author    = {Lemmerich, Florian},
  title     = {Novel Techniques for Efficient and Effective Subgroup Discovery},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-97812},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2014},
  abstract  = {Large volumes of data are collected today in many domains. Often, there is so much data available, that it is difficult to identify the relevant pieces of information. Knowledge discovery seeks to obtain novel, interesting and useful information from large datasets. One key technique for that purpose is subgroup discovery. It aims at identifying descriptions for subsets of the data, which have an interesting distribution with respect to a predefined target concept. This work improves the efficiency and effectiveness of subgroup discovery in different directions. For efficient exhaustive subgroup discovery, algorithmic improvements are proposed for three important variations of the standard setting: First, novel optimistic estimate bounds are derived for subgroup discovery with numeric target concepts. These allow for skipping the evaluation of large parts of the search space without influencing the results. Additionally, necessary adaptations to data structures for this setting are discussed. Second, for exceptional model mining, that is, subgroup discovery with a model over multiple attributes as target concept, a generic extension of the well-known FP-tree data structure is introduced. The modified data structure stores intermediate condensed data representations, which depend on the chosen model class, in the nodes of the trees. This allows the application for many popular model classes. Third, subgroup discovery with generalization-aware measures is investigated. These interestingness measures compare the target share or mean value in the subgroup with the respective maximum value in all its generalizations. For this setting, a novel method for deriving optimistic estimates is proposed. In contrast to previous approaches, the novel measures are not exclusively based on the anti-monotonicity of instance coverage, but also takes the difference of coverage between the subgroup and its generalizations into account. In all three areas, the advances lead to runtime improvements of more than an order of magnitude. The second part of the contributions focuses on the \emph{effectiveness} of subgroup discovery. These improvements aim to identify more interesting subgroups in practical applications. For that purpose, the concept of expectation-driven subgroup discovery is introduced as a new family of interestingness measures. It computes the score of a subgroup based on the difference between the actual target share and the target share that could be expected given the statistics for the separate influence factors that are combined to describe the subgroup. In doing so, previously undetected interesting subgroups are discovered, while other, partially redundant findings are suppressed. Furthermore, this work also approaches practical issues of subgroup discovery: In that direction, the VIKAMINE II tool is presented, which extends its predecessor with a rebuild user interface, novel algorithms for automatic discovery, new interactive mining techniques, as well novel options for result presentation and introspection. Finally, some real-world applications are described that utilized the presented techniques. These include the identification of influence factors on the success and satisfaction of university students and the description of locations using tagging data of geo-referenced images.},
  subject      = {Data Mining},
  language  = {en}
}
@phdthesis{Agorastou2022,
  author    = {Agorastou, Vaia},
  title     = {Nycthemerale Augeninnendruckschwankungen und Glaukomprogression},
  doi       = {10.25972/OPUS-26417},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-264176},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2022},
  abstract  = {Die n{\"a}chtliche (24-st{\"u}ndige) {\"U}berwachung des intraokularen Drucks (IOD) bei station{\"a}ren Glaukompatienten wird in Europa seit mehr als 100 Jahren eingesetzt, um Spitzenwerte zu erkennen, die w{\"a}hrend der regul{\"a}ren Sprechstundenzeiten {\"u}bersehen werden. Daten, die diese Praxis unterst{\"u}tzen, fehlen, zum Teil weil es schwierig ist, manuell erstellte IOD-Kurven mit dem objektiven Verlauf des Glaukoms zu korrelieren. Um dieses Problem zu beheben, haben wir automatisierte IOD-Datenextraktionswerkzeuge eingesetzt und auf eine Korrelation mit einem fortschreitenden Verlust der retinalen Nervenfaserschicht auf der optischen Koh{\"a}renztomographie im Spektralbereich (SDOCT) getestet.},
  subject      = {Glaukom},
  language  = {de}
}
@misc{Kaempgen2009,
  type      = {Master Thesis},
  author    = {Kaempgen, Benedikt},
  title     = {Deskriptives Data-Mining f{\"u}r Entscheidungstr{\"a}ger: Eine Mehrfachfallstudie},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-46343},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2009},
  abstract  = {Das Potenzial der Wissensentdeckung in Daten wird h{\"a}ufig nicht ausgenutzt, was haupts{\"a}chlich auf Barrieren zwischen dem Entwicklerteam und dem Endnutzer des Data-Mining zur{\"u}ckzuf{\"u}hren ist. In dieser Arbeit wird ein transparenter Ansatz zum Beschreiben und Erkl{\"a}ren von Daten f{\"u}r Entscheidungstr{\"a}ger vorgestellt. In Entscheidungstr{\"a}ger-zentrierten Aufgaben werden die Projektanforderungen definiert und die Ergebnisse zu einer Geschichte zusammengestellt. Eine Anforderung besteht dabei aus einem tabellarischen Bericht und ggf. Mustern in seinem Inhalt, jeweils verst{\"a}ndlich f{\"u}r einen Entscheidungstr{\"a}ger. Die technischen Aufgaben bestehen aus einer Datenpr{\"u}fung, der Integration der Daten in einem Data-Warehouse sowie dem Generieren von Berichten und dem Entdecken von Mustern wie in den Anforderungen beschrieben. Mehrere Data-Mining-Projekte k{\"o}nnen durch Wissensmanagement sowie eine geeignete Infrastruktur voneinander profitieren. Der Ansatz wurde in zwei Projekten unter Verwendung von ausschließlich Open-Source-Software angewendet.},
  subject      = {Data Mining},
  language  = {de}
}
@phdthesis{Bisenius2023,
  author    = {Bisenius, Fabian},
  title     = {Zum Stand der Versorgung chronisch herzinsuffizienter Patienten durch niedergelassene Kardiologen in Bayern - Ein Qualit{\"a}tssicherungsprojekt},
  doi       = {10.25972/OPUS-30303},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-303032},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2023},
  abstract  = {In dieser Arbeit wurde ein Kollektiv chronisch herzinsuffizienter Patienten aus der niedergelassenen kardiologischen Betreuung in Bayern analysiert und auf die Umsetzung der zum Zeitpunkt der HF-Bavaria Studie g{\"u}ltigen Leitlinien untersucht. Dabei wurde das Patientenkollektiv nach dem Geschlecht und zus{\"a}tzlich auch nach den neu definierten Herzinsuffizienz-Klassen der aktuell g{\"u}ltigen Leitlinien eingeteilt, um Unterschiede und Gemeinsamkeiten innerhalb dieser Differenzierungen darstellen zu k{\"o}nnen und einen Vergleich zu den Studien der j{\"u}ngeren Vergangenheit zu erm{\"o}glichen. Die Patienten der HF-Bavaria Studie waren zu 65,9 \% m{\"a}nnlich (n = 3569) und zu 34,1 \% weiblich (n = 1848). Die Frauen litten h{\"a}ufiger unter HFpEF, waren seit k{\"u}rzerer Zeit herzinsuffizient und waren in der Vergangenheit seltener zur Therapieintensivierung oder Intervention hospitalisiert. Die Patientinnen berichteten dabei weniger h{\"a}ufig Komorbidit{\"a}ten. So fanden sich bei den Frauen seltener KHK, Niereninsuffizienz oder Diabetes mellitus, hingegen h{\"a}ufiger Herzklappenerkrankungen und Vorhofflimmern. Weiterhin wurden die Patientinnen weniger h{\"a}ufig mit ACE-Hemmer, Betablocker und MRA, dagegen h{\"a}ufiger mit ARB und Digitalis behandelt. Im Patientenkollektiv der HF-Bavaria Studie hatten 29,0 \% eine HFrEF (n = 1581), 28,9 \% eine HFmrEF (n = 1577) und 42,0 \% eine HFpEF (n = 2291). Patienten mit HFrEF waren {\"u}berwiegend m{\"a}nnlich, zum gr{\"o}ßten Teil seit mehr als 5 Jahren herzinsuffizient und im Vergleich zu den anderen Herzinsuffizienz-Klassen h{\"a}ufiger in den NYHA-Stadien III und IV eingestuft. HFrEF Patienten hatten den gr{\"o}ßten Anteil an bereits erfolgten Interventionen und Device-Therapien und die durchschnittlich h{\"o}chste Anzahl an Komorbidit{\"a}ten. Das Komorbidit{\"a}tenspektrum bei Patienten mit HFmrEF lag prozentual in den meisten Kategorien zwischen den beiden anderen Herzinsuffizienz-Klassen. Patienten mit HFpEF waren {\"u}berThe ewiegend weiblich, wiesen vergleichsweise am h{\"a}ufigsten eine komorbide Hypertonie oder ein Vorhofflimmern auf, w{\"a}hrend eine KHK deutlich seltener vorlag, als es in den anderen Herzinsuffizienz-Klassen der Fall war. Die Pr{\"u}fung der leitliniengerechten Pharmakotherapie bei HFrEF-Patienten ergab eine insgesamt gleichwertige Verschreibungsh{\"a}ufigkeit im geschlechtsspezifischen Vergleich der nach NYHA-Stadium indizierten Medikamentenklassen und Kombinationstherapien. Lediglich im NYHA-Stadium III konnte gezeigt werden, dass M{\"a}nner signifikant h{\"a}ufiger mit einem Betablocker therapiert wurden. Weiterhin zeigte sich, bis auf wenige Ausnahmen, eine auch im nationalen und internationalen Vergleich hohe prozentuale Einnahme der stadienabh{\"a}ngig indizierten Medikamente. Die Einnahmerate von MRAs war vergleichsweise noch geringer als zu erwarten w{\"a}re, jedoch konnte das begleitende Vorliegen relevanter Kontraindikationen nicht zuverl{\"a}ssig genug erfasst werden, um die tats{\"a}chliche Versorgungsl{\"u}cke zu quantifizieren. Die Analyse der Pharmakotherapie von HFmrEF- und HFpEF-Patienten zeigte, trotz bisher fehlender wissenschaftlicher Erkenntnisse zur optimalen medikament{\"o}sen Therapie dieser Patientengruppen, sehr {\"a}hnliche Einnahmeh{\"a}ufigkeiten der verschiedenen Substanzklassen im Vergleich zu den HFrEF-Patienten. Die Therapie mit Devices war im Patientenkollektiv der HF-Bavaria Studie vergleichsweise selten und dabei h{\"a}ufiger bei m{\"a}nnlichen Patienten vorzufinden. Die Analyse der leitliniengetreuen Indikationen von ICDs, CRTs und CRT-ICDs zu den tats{\"a}chlich implantierten Devices ergab Hinweise auf eine Unterversorgung vermittels apparativer Therapiem{\"o}glichkeiten. Die Auswertung der HF-Bavaria Studie best{\"a}tigte die von uns erwartete Heterogenit{\"a}t und Komplexit{\"a}t der herzinsuffizienten Patienten in der niedergelassenen kardiologischen Betreuung. In dieser Arbeit konnte gezeigt werden, dass bedeutsame Unterschiede im Hinblick auf das Profil, den Verlauf und die Therapie von m{\"a}nnlichen und weiblichen herzinsuffizienten Patienten bestehen. Die Therapieempfehlungen der Leitlinien richten sich trotz dieser Unterschiede vorrangig nach der Herzinsuffizienz-Klasse der Patienten. Bisher existierten in den Leitlinien vorrangig Therapieempfehlungen f{\"u}r Patienten mit einer HFrEF (und LVEF <40 \%). Im Patientenkollektiv fanden sich jedoch zu 71 \% Patienten mit einer LVEF ≥40 \%. Dies bedeutet, dass f{\"u}r den Großteil der Patienten in unserer Studie bisher keine evidenzbasierten Behandlungsalgorithmen existieren, insbesondere zur Pharmakotherapie. K{\"u}nftig sollte die Forschung vermehrt auf diese Evidenzl{\"u}cken eingehen und idealerweise eine personalisierte Therapie erm{\"o}glichen. Abschließend l{\"a}sst sich feststellen, dass die leitliniengerechte Therapie der herzinsuffizienten Patienten in der niedergelassenen kardiologischen Versorgung in Bayern eine im nationalen und internationalen Kontext fortgeschrittene Qualit{\"a}t besitzt. Dennoch wurden erwartungsgem{\"a}ß M{\"o}glichkeiten zur Qualit{\"a}tsverbesserung im vorliegenden Projekt identifiziert.},
  subject      = {Chronische Herzinsuffizienz},
  language  = {de}
}
@phdthesis{Schloer2022,
  author    = {Schl{\"o}r, Daniel},
  title     = {Detecting Anomalies in Transaction Data},
  doi       = {10.25972/OPUS-29856},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-298569},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2022},
  abstract  = {Detecting anomalies in transaction data is an important task with a high potential to avoid financial loss due to irregularities deliberately or inadvertently carried out, such as credit card fraud, occupational fraud in companies or ordering and accounting errors. With ongoing digitization of our world, data-driven approaches, including machine learning, can draw benefit from data with less manual effort and feature engineering. A large variety of machine learning-based anomaly detection methods approach this by learning a precise model of normality from which anomalies can be distinguished. Modeling normality in transactional data, however, requires to capture distributions and dependencies within the data precisely with special attention to numerical dependencies such as quantities, prices or amounts. To implicitly model numerical dependencies, Neural Arithmetic Logic Units have been proposed as neural architecture. In practice, however, these have stability and precision issues. Therefore, we first develop an improved neural network architecture, iNALU, which is designed to better model numerical dependencies as found in transaction data. We compare this architecture to the previous approach and show in several experiments of varying complexity that our novel architecture provides better precision and stability. We integrate this architecture into two generative neural network models adapted for transaction data and investigate how well normal behavior is modeled. We show that both architectures can successfully model normal transaction data, with our neural architecture improving generative performance for one model. Since categorical and numerical variables are common in transaction data, but many machine learning methods only process numerical representations, we explore different representation learning techniques to transform categorical transaction data into dense numerical vectors. We extend this approach by proposing an outlier-aware discretization, thus incorporating numerical attributes into the computation of categorical embeddings, and investigate latent spaces, as well as quantitative performance for anomaly detection. Next, we evaluate different scenarios for anomaly detection on transaction data. We extend our iNALU architecture to a neural layer that can model both numerical and non-numerical dependencies and evaluate it in a supervised and one-class setting. We investigate the stability and generalizability of our approach and show that it outperforms a variety of models in the balanced supervised setting and performs comparably in the one-class setting. Finally, we evaluate three approaches to using a generative model as an anomaly detector and compare the anomaly detection performance.},
  subject      = {Anomalieerkennung},
  language  = {en}
}
@misc{Igl2002,
  type      = {Master Thesis},
  author    = {Igl, Wilmar},
  title     = {Komplexes Probleml{\"o}sen in Multiagentensimulationsszenarien : Untersuchungen zur Formalisierung von Strategien f{\"u}r die Bek{\"a}mpfung von Waldbr{\"a}nden},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-10771},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2002},
  abstract  = {Die vorliegende Arbeit ist in zwei Teile gegliedert, von denen der erste Teil den theoretischen Hintergrund und empirische Befunde zum Thema „Komplexes Probleml{\"o}sen" behandelt. Der zweite Teil beinhaltet Methodik und Ergebnisse der durchgef{\"u}hrten Untersuchung. Nach der Einleitung in Kapitel 1 werden in Kapitel 2 die „Grundkonzepte des Komplexen Probleml{\"o}sens" vorgestellt, wobei mit der Abgrenzung des Bereichs „Komplexes Probleml{\"o}sen" begonnen wird. Anschließend werden die Eigenschaften von komplexen Systemen und deren Anforderungen an Probleml{\"o}ser beschrieben, wobei die Taxonomie1 von D{\"o}rner et al. (1994) zugrunde gelegt wird. In Kapitel 3 werden Modelle der Wissensrepr{\"a}sentation und des Probleml{\"o}sens vorgestellt. Dabei wird der Begriff der „Strategie" diskutiert und im Zusammenhang mit verschiedenen allgemeinen Modellen des Probleml{\"o}sens erl{\"a}utert. Kapitel 4 behandelt das Konzept „Delegation". Delegation wird in dieser Arbeit als Methode verwendet, um Versuchspersonen zur Formalisierung ihrer Strategien zu bewegen, wobei sie die Ausf{\"u}hrung der Strategien gleichzeitig beobachten k{\"o}nnen. Es werden vor allem Befunde aus der Organisationspsychologie und Unternehmensf{\"u}hrung berichtet und die Anwendung von Delegation in der Interaktion zwischen Mensch und k{\"u}nstlichem Agent er{\"o}rtert. In Kapitel 5 werden Waldbrandsimulationen behandelt. Diese z{\"a}hlen zu den klassischen Simulationen, die zur Untersuchung von Komplexem Probleml{\"o}sen verwendet werden. Zuerst wird auf computergest{\"u}tzte Simulation im Allgemeinen eingegangen, wobei Unterschiede zu traditionellen Untersuchungsmethoden angesprochen werden. Dabei wird auch die Bedeutung der Multiagentensimulation f{\"u}r die Komplexe Probleml{\"o}seforschung hervorgehoben. Anschließend wird Feuerverhalten und Feuerbek{\"a}mpfung als Vorbild f{\"u}r Waldbrandsimulationen erl{\"a}utert. Dadurch k{\"o}nnen sowohl Anhaltspunkte zur Beurteilung der Plausibilit{\"a}t als auch f{\"u}r die Implementierung einer Waldbrandsimulation gewonnen werden. Im Anschluss daran werden drei bekannte Beispiele f{\"u}r Waldbrandsimulationen vorgestellt, wobei auch auf dom{\"a}nen- bzw. simulationsspezifische Strategien eingegangen wird. In Kapitel 6 wird ein {\"U}berblick {\"u}ber verschiedene empirische Befunde aus dem Bereich des Komplexen Probleml{\"o}sens gegeben. Diese betreffen sowohl Eigenschaften von komplexen Systemen als auch Merkmale des Probleml{\"o}sers. In Kapitel 7 werden die wichtigsten Kritikpunkte und Probleme, mit denen die Komplexe Probleml{\"o}seforschung zu k{\"a}mpfen hat, zusammengefasst. Die konkreten Fragestellungen der Untersuchung werden in Kapitel 8 vorgestellt, wobei Kapitel 9 und 10 erl{\"a}utern, mit welcher Methodik diese Fragen untersucht werden. In diesem Zusammenhang wird auch die Simulationsumgebung SeSAm vorgestellt. Im folgenden Kapitel 11 wird auf die Eigenschaften der implementierten Waldbrandsimulation eingegangen. Kapitel 12 beschreibt den Aufbau und Ablauf der Untersuchung, mit der die Daten gewonnen werden, die in Kapitel 13 berichtet werden. Eine Diskussion der Befunde im Hinblick auf die Fragestellungen und ihre Bedeutung f{\"u}r die zuk{\"u}nftige Forschung erfolgt in Kapitel 14.},
  language  = {de}
}
@phdthesis{Furth2018,
  author    = {Furth, Sebastian},
  title     = {Linkable Technical Documentation},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-174185},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2018},
  abstract  = {The success of semantic systems has been proven over the last years. Nowadays, Linked Data is the driver for the rapid development of ever new intelligent systems. Especially in enterprise environments semantic systems successfully support more and more business processes. This is especially true for after sales service in the mechanical engineering domain. Here, service technicians need effective access to relevant technical documentation in order to diagnose and solve problems and defects. Therefore, the usage of semantic information retrieval systems has become the new system metaphor. Unlike classical retrieval software Linked Enterprise Data graphs are exploited to grant targeted and problem-oriented access to relevant documents. However, huge parts of legacy technical documents have not yet been integrated into Linked Enterprise Data graphs. Additionally, a plethora of information models for the semantic representation of technical information exists. The semantic maturity of these information models can hardly be measured. This thesis motivates that there is an inherent need for a self-contained semantification approach for technical documents. This work introduces a maturity model that allows to quickly assess existing documentation. Additionally, the approach comprises an abstracting semantic representation for technical documents that is aligned to all major standard information models. The semantic representation combines structural and rhetorical aspects to provide access to so called Core Documentation Entities. A novel and holistic semantification process describes how technical documents in different legacy formats can be transformed to a semantic and linked representation. The practical significance of the semantification approach depends on tools supporting its application. This work presents an accompanying tool chain of semantification applications, especially the semantification framework CAPLAN that is a highly integrated development and runtime environment for semantification processes. The complete semantification approach is evaluated in four real-life projects: in a spare part augmentation project, semantification projects for earth moving technology and harvesting technology, as well as an ontology population project for special purpose vehicles. Three additional case studies underline the broad applicability of the presented ideas.},
  subject      = {Linked Data},
  language  = {en}
}
@phdthesis{Dietrich2019,
  author    = {Dietrich, Georg},
  title     = {Ad Hoc Information Extraction in a Clinical Data Warehouse with Case Studies for Data Exploration and Consistency Checks},
  doi       = {10.25972/OPUS-18464},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-184642},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2019},
  abstract  = {The importance of Clinical Data Warehouses (CDW) has increased significantly in recent years as they support or enable many applications such as clinical trials, data mining, and decision making. CDWs integrate Electronic Health Records which still contain a large amount of text data, such as discharge letters or reports on diagnostic findings in addition to structured and coded data like ICD-codes of diagnoses. Existing CDWs hardly support features to gain information covered in texts. Information extraction methods offer a solution for this problem but they have a high and long development effort, which can only be carried out by computer scientists. Moreover, such systems only exist for a few medical domains. This paper presents a method empowering clinicians to extract information from texts on their own. Medical concepts can be extracted ad hoc from e.g. discharge letters, thus physicians can work promptly and autonomously. The proposed system achieves these improvements by efficient data storage, preprocessing, and with powerful query features. Negations in texts are recognized and automatically excluded, as well as the context of information is determined and undesired facts are filtered, such as historical events or references to other persons (family history). Context-sensitive queries ensure the semantic integrity of the concepts to be extracted. A new feature not available in other CDWs is to query numerical concepts in texts and even filter them (e.g. BMI > 25). The retrieved values can be extracted and exported for further analysis. This technique is implemented within the efficient architecture of the PaDaWaN CDW and evaluated with comprehensive and complex tests. The results outperform similar approaches reported in the literature. Ad hoc IE determines the results in a few (milli-) seconds and a user friendly GUI enables interactive working, allowing flexible adaptation of the extraction. In addition, the applicability of this system is demonstrated in three real-world applications at the W{\"u}rzburg University Hospital (UKW). Several drug trend studies are replicated: Findings of five studies on high blood pressure, atrial fibrillation and chronic renal failure can be partially or completely confirmed in the UKW. Another case study evaluates the prevalence of heart failure in inpatient hospitals using an algorithm that extracts information with ad hoc IE from discharge letters and echocardiogram report (e.g. LVEF < 45 ) and other sources of the hospital information system. This study reveals that the use of ICD codes leads to a significant underestimation (31\%) of the true prevalence of heart failure. The third case study evaluates the consistency of diagnoses by comparing structured ICD-10-coded diagnoses with the diagnoses described in the diagnostic section of the discharge letter. These diagnoses are extracted from texts with ad hoc IE, using synonyms generated with a novel method. The developed approach can extract diagnoses from the discharge letter with a high accuracy and furthermore it can prove the degree of consistency between the coded and reported diagnoses.},
  subject      = {Information Extraction},
  language  = {en}
}
@phdthesis{Krug2020,
  author    = {Krug, Markus},
  title     = {Techniques for the Automatic Extraction of Character Networks in German Historic Novels},
  doi       = {10.25972/OPUS-20918},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-209186},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2020},
  abstract  = {Recent advances in Natural Language Preprocessing (NLP) allow for a fully automatic extraction of character networks for an incoming text. These networks serve as a compact and easy to grasp representation of literary fiction. They offer an aggregated view of the text, which can be used during distant reading approaches for the analysis of literary hypotheses. In their core, the networks consist of nodes, which represent literary characters, and edges, which represent relations between characters. For an automatic extraction of such a network, the first step is the detection of the references of all fictional entities that are of importance for a text. References to the fictional entities appear in the form of names, noun phrases and pronouns and prior to this work, no components capable of automatic detection of character references were available. Existing tools are only capable of detecting proper nouns, a subset of all character references. When evaluated on the task of detecting proper nouns in the domain of literary fiction, they still underperform at an F1-score of just about 50\%. This thesis uses techniques from the field of semi-supervised learning, such as Distant supervision and Generalized Expectations, and improves the results of an existing tool to about 82\%, when evaluated on all three categories in literary fiction, but without the need for annotated data in the target domain. However, since this quality is still not sufficient, the decision to annotate DROC, a corpus comprising 90 fragments of German novels was made. This resulted in a new general purpose annotation environment titled as ATHEN, as well as annotated data that spans about 500.000 tokens in total. Using this data, the combination of supervised algorithms and a tailored rule based algorithm, which in combination are able to exploit both - local consistencies as well as global consistencies - yield an algorithm with an F1-score of about 93\%. This component is referred to as the Kallimachos tagger. A character network can not directly display references however, instead they need to be clustered so that all references that belong to a real world or fictional entity are grouped together. This process widely known as coreference resolution is a hard problem in the focus of research for more than half a century. This work experimented with adaptations of classical feature based machine learning, with a dedicated rule based algorithm and with modern techniques of Deep Learning, but no approach can surpass 55\% B-Cubed F1, when evaluated on DROC. Due to this barrier, many researchers do not use a fully-fledged coreference resolution when they extract character networks, but only focus on a more forgiving subset- the names. For novels such as Alice's Adventures in Wonderland by Lewis Caroll, this would however only result in a network in which many important characters are missing. In order to integrate important characters into the network that are not named by the author, this work makes use of automatic detection of speaker and addressees for direct speech utterances (all entities involved in a dialog are considered to be of importance). This problem is by itself not an easy task, however the most successful system analysed in this thesis is able to correctly determine the speaker to about 85\% of the utterances as well as about 65\% of the addressees. This speaker information can not only help to identify the most dominant characters, but also serves as a way to model the relations between entities. During the span of this work, components have been developed to model relations between characters using speaker attribution, using co-occurrences as well as by the usage of true interactions, for which yet again a dataset was annotated using ATHEN. Furthermore, since relations between characters are usually typed, a component for the extraction of a typed relation was developed. Similar to the experiments for the character reference detection, a combination of a rule based and a Maximum Entropy classifier yielded the best overall results, with the extraction of family relations showing a score of about 80\% and the quality of love relations with a score of about 50\%. For family relations, a kernel for a Support Vector Machine was developed that even exceeded the scores of the combined approach but is behind on the other labels. In addition, this work presents new ways to evaluate automatically extracted networks without the need of domain experts, instead it relies on the usage of expert summaries. It also refrains from the uses of social network analysis for the evaluation, but instead presents ranked evaluations using Precision@k and the Spearman Rank correlation coefficient for the evaluation of the nodes and edges of the network. An analysis using these metrics showed, that the central characters of a novel are contained with high probability but the quality drops rather fast if more than five entities are analyzed. The quality of the edges is mainly dominated by the quality of the coreference resolution and the correlation coefficient between gold edges and system edges therefore varies between 30 and 60\%. All developed components are aggregated alongside a large set of other preprocessing modules in the Kallimachos pipeline and can be reused without any restrictions.},
  subject      = {Textanalyse},
  language  = {en}
}
@phdthesis{Reul2020,
  author    = {Reul, Christian},
  title     = {An Intelligent Semi-Automatic Workflow for Optical Character Recognition of Historical Printings},
  doi       = {10.25972/OPUS-20923},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-209239},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2020},
  abstract  = {Optical Character Recognition (OCR) on historical printings is a challenging task mainly due to the complexity of the layout and the highly variant typography. Nevertheless, in the last few years great progress has been made in the area of historical OCR resulting in several powerful open-source tools for preprocessing, layout analysis and segmentation, Automatic Text Recognition (ATR) and postcorrection. Their major drawback is that they only offer limited applicability by non-technical users like humanist scholars, in particular when it comes to the combined use of several tools in a workflow. Furthermore, depending on the material, these tools are usually not able to fully automatically achieve sufficiently low error rates, let alone perfect results, creating a demand for an interactive postcorrection functionality which, however, is generally not incorporated. This thesis addresses these issues by presenting an open-source OCR software called OCR4all which combines state-of-the-art OCR components and continuous model training into a comprehensive workflow. While a variety of materials can already be processed fully automatically, books with more complex layouts require manual intervention by the users. This is mostly due to the fact that the required Ground Truth (GT) for training stronger mixed models (for segmentation as well as text recognition) is not available, yet, neither in the desired quantity nor quality. To deal with this issue in the short run, OCR4all offers better recognition capabilities in combination with a very comfortable Graphical User Interface (GUI) that allows error corrections not only in the final output, but already in early stages to minimize error propagation. In the long run this constant manual correction produces large quantities of valuable, high quality training material which can be used to improve fully automatic approaches. Further on, extensive configuration capabilities are provided to set the degree of automation of the workflow and to make adaptations to the carefully selected default parameters for specific printings, if necessary. The architecture of OCR4all allows for an easy integration (or substitution) of newly developed tools for its main components by supporting standardized interfaces like PageXML, thus aiming at continual higher automation for historical printings. In addition to OCR4all, several methodical extensions in the form of accuracy improving techniques for training and recognition are presented. Most notably an effective, sophisticated, and adaptable voting methodology using a single ATR engine, a pretraining procedure, and an Active Learning (AL) component are proposed. Experiments showed that combining pretraining and voting significantly improves the effectiveness of book-specific training, reducing the obtained Character Error Rates (CERs) by more than 50\%. The proposed extensions were further evaluated during two real world case studies: First, the voting and pretraining techniques are transferred to the task of constructing so-called mixed models which are trained on a variety of different fonts. This was done by using 19th century Fraktur script as an example, resulting in a considerable improvement over a variety of existing open-source and commercial engines and models. Second, the extension from ATR on raw text to the adjacent topic of typography recognition was successfully addressed by thoroughly indexing a historical lexicon that heavily relies on different font types in order to encode its complex semantic structure. During the main experiments on very complex early printed books even users with minimal or no experience were able to not only comfortably deal with the challenges presented by the complex layout, but also to recognize the text with manageable effort and great quality, achieving excellent CERs below 0.5\%. Furthermore, the fully automated application on 19th century novels showed that OCR4all (average CER of 0.85\%) can considerably outperform the commercial state-of-the-art tool ABBYY Finereader (5.3\%) on moderate layouts if suitably pretrained mixed ATR models are available.},
  subject      = {Optische Zeichenerkennung},
  language  = {en}
}
@phdthesis{Djebko2020,
  author    = {Djebko, Kirill},
  title     = {Quantitative modellbasierte Diagnose am Beispiel der Energieversorgung des SONATE-Nanosatelliten mit automatisch kalibrierten Modellkomponenten},
  doi       = {10.25972/OPUS-20662},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-206628},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2020},
  abstract  = {Von technischen Systemen wird in der heutigen Zeit erwartet, dass diese stets fehlerfrei funktionieren, um einen reibungslosen Ablauf des Alltags zu gew{\"a}hrleisten. Technische Systeme jedoch k{\"o}nnen Defekte aufweisen, die deren Funktionsweise einschr{\"a}nken oder zu deren Totalausfall f{\"u}hren k{\"o}nnen. Grunds{\"a}tzlich zeigen sich Defekte durch eine Ver{\"a}nderung im Verhalten von einzelnen Komponenten. Diese Abweichungen vom Nominalverhalten nehmen dabei an Intensit{\"a}t zu, je n{\"a}her die entsprechende Komponente an einem Totalausfall ist. Aus diesem Grund sollte das Fehlverhalten von Komponenten rechtzeitig erkannt werden, um permanenten Schaden zu verhindern. Von besonderer Bedeutung ist dies f{\"u}r die Luft- und Raumfahrt. Bei einem Satelliten kann keine Wartung seiner Komponenten durchgef{\"u}hrt werden, wenn er sich bereits im Orbit befindet. Der Defekt einer einzelnen Komponente, wie der Batterie der Energieversorgung, kann hierbei den Verlust der gesamten Mission bedeuten. Grunds{\"a}tzlich l{\"a}sst sich Fehlererkennung manuell durchf{\"u}hren, wie es im Satellitenbetrieb oft {\"u}blich ist. Hierf{\"u}r muss ein menschlicher Experte, ein sogenannter Operator, das System {\"u}berwachen. Diese Form der {\"U}berwachung ist allerdings stark von der Zeit, Verf{\"u}gbarkeit und Expertise des Operators, der die {\"U}berwachung durchf{\"u}hrt, abh{\"a}ngig. Ein anderer Ansatz ist die Verwendung eines dedizierten Diagnosesystems. Dieses kann das technische System permanent {\"u}berwachen und selbstst{\"a}ndig Diagnosen berechnen. Die Diagnosen k{\"o}nnen dann durch einen Experten eingesehen werden, der auf ihrer Basis Aktionen durchf{\"u}hren kann. Das in dieser Arbeit vorgestellte modellbasierte Diagnosesystem verwendet ein quantitatives Modell eines technischen Systems, das dessen Nominalverhalten beschreibt. Das beobachtete Verhalten des technischen Systems, gegeben durch Messwerte, wird mit seinem erwarteten Verhalten, gegeben durch simulierte Werte des Modells, verglichen und Diskrepanzen bestimmt. Jede Diskrepanz ist dabei ein Symptom. Diagnosen werden dadurch berechnet, dass zun{\"a}chst zu jedem Symptom eine sogenannte Konfliktmenge berechnet wird. Dies ist eine Menge von Komponenten, sodass der Defekt einer dieser Komponenten das entsprechende Symptom erkl{\"a}ren k{\"o}nnte. Mithilfe dieser Konfliktmengen werden sogenannte Treffermengen berechnet. Eine Treffermenge ist eine Menge von Komponenten, sodass der gleichzeitige Defekt aller Komponenten dieser Menge alle beobachteten Symptome erkl{\"a}ren k{\"o}nnte. Jede minimale Treffermenge entspricht dabei einer Diagnose. Zur Berechnung dieser Mengen nutzt das Diagnosesystem ein Verfahren, bei dem zun{\"a}chst abh{\"a}ngige Komponenten bestimmt werden und diese von symptombehafteten Komponenten belastet und von korrekt funktionierenden Komponenten entlastet werden. F{\"u}r die einzelnen Komponenten werden Bewertungen auf Basis dieser Be- und Entlastungen berechnet und mit ihnen Diagnosen gestellt. Da das Diagnosesystem auf ausreichend genaue Modelle angewiesen ist und die manuelle Kalibrierung dieser Modelle mit erheblichem Aufwand verbunden ist, wurde ein Verfahren zur automatischen Kalibrierung entwickelt. Dieses verwendet einen Zyklischen Genetischen Algorithmus, um mithilfe von aufgezeichneten Werten der realen Komponenten Modellparameter zu bestimmen, sodass die Modelle die aufgezeichneten Daten m{\"o}glichst gut reproduzieren k{\"o}nnen. Zur Evaluation der automatischen Kalibrierung wurden ein Testaufbau und verschiedene dynamische und manuell schwierig zu kalibrierende Komponenten des Qualifikationsmodells eines realen Nanosatelliten, dem SONATE-Nanosatelliten modelliert und kalibriert. Der Testaufbau bestand dabei aus einem Batteriepack, einem Laderegler, einem Tiefentladeschutz, einem Entladeregler, einem Stepper Motor HAT und einem Motor. Er wurde zus{\"a}tzlich zur automatischen Kalibrierung unabh{\"a}ngig manuell kalibriert. Die automatisch kalibrierten Satellitenkomponenten waren ein Reaktionsrad, ein Entladeregler, Magnetspulen, bestehend aus einer Ferritkernspule und zwei Luftspulen, eine Abschlussleiterplatine und eine Batterie. Zur Evaluation des Diagnosesystems wurde die Energieversorgung des Qualifikationsmodells des SONATE-Nanosatelliten modelliert. F{\"u}r die Batterien, die Entladeregler, die Magnetspulen und die Reaktionsr{\"a}der wurden die vorher automatisch kalibrierten Modelle genutzt. F{\"u}r das Modell der Energieversorgung wurden Fehler simuliert und diese diagnostiziert. Die Ergebnisse der Evaluation der automatischen Kalibrierung waren, dass die automatische Kalibrierung eine mit der manuellen Kalibrierung vergleichbare Genauigkeit f{\"u}r den Testaufbau lieferte und diese sogar leicht {\"u}bertraf und dass die automatisch kalibrierten Satellitenkomponenten eine durchweg hohe Genauigkeit aufwiesen und damit f{\"u}r den Einsatz im Diagnosesystem geeignet waren. Die Ergebnisse der Evaluation des Diagnosesystems waren, dass die simulierten Fehler zuverl{\"a}ssig gefunden wurden und dass das Diagnosesystem in der Lage war die plausiblen Ursachen dieser Fehler zu diagnostizieren.},
  subject      = {Satellit},
  language  = {de}
}
@phdthesis{Steininger2023,
  author    = {Steininger, Michael},
  title     = {Deep Learning for Geospatial Environmental Regression},
  doi       = {10.25972/OPUS-31312},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-313121},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2023},
  abstract  = {Environmental issues have emerged especially since humans burned fossil fuels, which led to air pollution and climate change that harm the environment. These issues' substantial consequences evoked strong efforts towards assessing the state of our environment. Various environmental machine learning (ML) tasks aid these efforts. These tasks concern environmental data but are common ML tasks otherwise, i.e., datasets are split (training, validatition, test), hyperparameters are optimized on validation data, and test set metrics measure a model's generalizability. This work focuses on the following environmental ML tasks: Regarding air pollution, land use regression (LUR) estimates air pollutant concentrations at locations where no measurements are available based on measured locations and each location's land use (e.g., industry, streets). For LUR, this work uses data from London (modeled) and Zurich (measured). Concerning climate change, a common ML task is model output statistics (MOS), where a climate model's output for a study area is altered to better fit Earth observations and provide more accurate climate data. This work uses the regional climate model (RCM) REMO and Earth observations from the E-OBS dataset for MOS. Another task regarding climate is grain size distribution interpolation where soil properties at locations without measurements are estimated based on the few measured locations. This can provide climate models with soil information, that is important for hydrology. For this task, data from Lower Franconia is used. Such environmental ML tasks commonly have a number of properties: (i) geospatiality, i.e., their data refers to locations relative to the Earth's surface. (ii) The environmental variables to estimate or predict are usually continuous. (iii) Data can be imbalanced due to relatively rare extreme events (e.g., extreme precipitation). (iv) Multiple related potential target variables can be available per location, since measurement devices often contain different sensors. (v) Labels are spatially often only sparsely available since conducting measurements at all locations of interest is usually infeasible. These properties present challenges but also opportunities when designing ML methods for such tasks. In the past, environmental ML tasks have been tackled with conventional ML methods, such as linear regression or random forests (RFs). However, the field of ML has made tremendous leaps beyond these classic models through deep learning (DL). In DL, models use multiple layers of neurons, producing increasingly higher-level feature representations with growing layer depth. DL has made previously infeasible ML tasks feasible, improved the performance for many tasks in comparison to existing ML models significantly, and eliminated the need for manual feature engineering in some domains due to its ability to learn features from raw data. To harness these advantages for environmental domains it is promising to develop novel DL methods for environmental ML tasks. This thesis presents methods for dealing with special challenges and exploiting opportunities inherent to environmental ML tasks in conjunction with DL. To this end, the proposed methods explore the following techniques: (i) Convolutions as in convolutional neural networks (CNNs) to exploit reoccurring spatial patterns in geospatial data. (ii) Posing the problems as regression tasks to estimate the continuous variables. (iii) Density-based weighting to improve estimation performance for rare and extreme events. (iv) Multi-task learning to make use of multiple related target variables. (v) Semi-supervised learning to cope with label sparsity. Using these techniques, this thesis considers four research questions: (i) Can air pollution be estimated without manual feature engineering? This is answered positively by the introduction of the CNN-based LUR model MapLUR as well as the off-the-shelf LUR solution OpenLUR. (ii) Can colocated pollution data improve spatial air pollution models? Multi-task learning for LUR is developed for this, showing potential for improvements with colocated data. (iii) Can DL models improve the quality of climate model outputs? The proposed DL climate MOS architecture ConvMOS demonstrates this. Additionally, semi-supervised training of multilayer perceptrons (MLPs) for grain size distribution interpolation is presented, which can provide improved input data. (iv) Can DL models be taught to better estimate climate extremes? To this end, density-based weighting for imbalanced regression (DenseLoss) is proposed and applied to the DL architecture ConvMOS, improving climate extremes estimation. These methods show how especially DL techniques can be developed for environmental ML tasks with their special characteristics in mind. This allows for better models than previously possible with conventional ML, leading to more accurate assessment and better understanding of the state of our environment.},
  subject      = {Deep learning},
  language  = {en}
}
@phdthesis{Eismann2023,
  author    = {Eismann, Simon},
  title     = {Performance Engineering of Serverless Applications and Platforms},
  doi       = {10.25972/OPUS-30313},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-303134},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2023},
  abstract  = {Serverless computing is an emerging cloud computing paradigm that offers a highlevel application programming model with utilization-based billing. It enables the deployment of cloud applications without managing the underlying resources or worrying about other operational aspects. Function-as-a-Service (FaaS) platforms implement serverless computing by allowing developers to execute code on-demand in response to events with continuous scaling while having to pay only for the time used with sub-second metering. Cloud providers have further introduced many fully managed services for databases, messaging buses, and storage that also implement a serverless computing model. Applications composed of these fully managed services and FaaS functions are quickly gaining popularity in both industry and in academia. However, due to this rapid adoption, much information surrounding serverless computing is inconsistent and often outdated as the serverless paradigm evolves. This makes the performance engineering of serverless applications and platforms challenging, as there are many open questions, such as: What types of applications is serverless computing well suited for, and what are its limitations? How should serverless applications be designed, configured, and implemented? Which design decisions impact the performance properties of serverless platforms and how can they be optimized? These and many other open questions can be traced back to an inconsistent understanding of serverless applications and platforms, which could present a major roadblock in the adoption of serverless computing. In this thesis, we address the lack of performance knowledge surrounding serverless applications and platforms from multiple angles: we conduct empirical studies to further the understanding of serverless applications and platforms, we introduce automated optimization methods that simplify the operation of serverless applications, and we enable the analysis of design tradeoffs of serverless platforms by extending white-box performance modeling.},
  subject      = {Leistungsbewertung},
  language  = {en}
}
@phdthesis{Krenzer2023,
  author    = {Krenzer, Adrian},
  title     = {Machine learning to support physicians in endoscopic examinations with a focus on automatic polyp detection in images and videos},
  doi       = {10.25972/OPUS-31911},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-319119},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2023},
  abstract  = {Deep learning enables enormous progress in many computer vision-related tasks. Artificial Intel- ligence (AI) steadily yields new state-of-the-art results in the field of detection and classification. Thereby AI performance equals or exceeds human performance. Those achievements impacted many domains, including medical applications. One particular field of medical applications is gastroenterology. In gastroenterology, machine learning algorithms are used to assist examiners during interventions. One of the most critical concerns for gastroenterologists is the development of Colorectal Cancer (CRC), which is one of the leading causes of cancer-related deaths worldwide. Detecting polyps in screening colonoscopies is the essential procedure to prevent CRC. Thereby, the gastroenterologist uses an endoscope to screen the whole colon to find polyps during a colonoscopy. Polyps are mucosal growths that can vary in severity. This thesis supports gastroenterologists in their examinations with automated detection and clas- sification systems for polyps. The main contribution is a real-time polyp detection system. This system is ready to be installed in any gastroenterology practice worldwide using open-source soft- ware. The system achieves state-of-the-art detection results and is currently evaluated in a clinical trial in four different centers in Germany. The thesis presents two additional key contributions: One is a polyp detection system with ex- tended vision tested in an animal trial. Polyps often hide behind folds or in uninvestigated areas. Therefore, the polyp detection system with extended vision uses an endoscope assisted by two additional cameras to see behind those folds. If a polyp is detected, the endoscopist receives a vi- sual signal. While the detection system handles the additional two camera inputs, the endoscopist focuses on the main camera as usual. The second one are two polyp classification models, one for the classification based on shape (Paris) and the other on surface and texture (NBI International Colorectal Endoscopic (NICE) classification). Both classifications help the endoscopist with the treatment of and the decisions about the detected polyp. The key algorithms of the thesis achieve state-of-the-art performance. Outstandingly, the polyp detection system tested on a highly demanding video data set shows an F1 score of 90.25 \% while working in real-time. The results exceed all real-time systems in the literature. Furthermore, the first preliminary results of the clinical trial of the polyp detection system suggest a high Adenoma Detection Rate (ADR). In the preliminary study, all polyps were detected by the polyp detection system, and the system achieved a high usability score of 96.3 (max 100). The Paris classification model achieved an F1 score of 89.35 \% which is state-of-the-art. The NICE classification model achieved an F1 score of 81.13 \%. Furthermore, a large data set for polyp detection and classification was created during this thesis. Therefore a fast and robust annotation system called Fast Colonoscopy Annotation Tool (FastCAT) was developed. The system simplifies the annotation process for gastroenterologists. Thereby the i gastroenterologists only annotate key parts of the endoscopic video. Afterward, those video parts are pre-labeled by a polyp detection AI to speed up the process. After the AI has pre-labeled the frames, non-experts correct and finish the annotation. This annotation process is fast and ensures high quality. FastCAT reduces the overall workload of the gastroenterologist on average by a factor of 20 compared to an open-source state-of-art annotation tool.},
  subject      = {Deep Learning},
  language  = {en}
}