<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
  <channel>
    <title>KLUEDO RSS Feed</title>
    <description>KLUEDO Dokumente/documents</description>
    <link>https://kluedo.ub.rptu.de/index/index/</link>
    <pubDate>Fri, 13 Nov 2026 12:36:44 +0100</pubDate>
    <lastBuildDate>Fri, 13 Nov 2026 12:36:44 +0100</lastBuildDate>
    <item>
      <title>Metrikenbasierte Informationspräsentation: Eine experimentelle Untersuchung zu Wahrnehmung, Bewertung und Konsequenzen aus Konsumentenperspektive</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13772</link>
      <description>Die vorliegende Dissertation untersucht, wie alternative Preismetriken als Form metrikenbasierter Informationspräsentation die Wahrnehmung, Verarbeitung und Bewertung von Preisinformationen sowie Entscheidungen und Verhaltensabsichten von Konsumenten beeinflussen. Im Mittelpunkt steht die Frage, welche Wirkungen die Wahl einer Bezugsgröße bei unveränderten Preisen entfaltet und unter welchen Bedingungen entsprechende Darstellungen die Informationsverarbeitung und Entscheidungsfindung unterstützen.&#13;
&#13;
Auf Grundlage verhaltenswissenschaftlicher Ansätze zur Preiswahrnehmung, kognitiven Passung, Bewertbarkeit und Informationsverarbeitung werden sieben Experimente in zwei Forschungsprojekten durchgeführt. Forschungsprojekt I untersucht alternative Preismetriken als Wahrnehmungs-, Bewertungs- und Entscheidungshilfe. Forschungsprojekt II betrachtet ihre Wirkung als Entscheidungsrahmung und vergleicht quantitative Preismetriken mit qualitativen Hinweisen.&#13;
&#13;
Die Ergebnisse zeigen, dass alternative Preismetriken die subjektive Preisverarbeitung erleichtern und die wahrgenommene Fairness der Preispräsentation erhöhen können. Über die Fairnessbewertung zeigen sich zudem indirekte Wirkungen auf Wiederkaufs- und Weiterempfehlungsabsichten. Die Arbeit erweitert die konsumentenbezogene Preisforschung um eine präsentationslogische Betrachtung alternativer Preismetriken.</description>
      <author>Marcel Mallach</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13772</guid>
      <pubDate>Fri, 13 Nov 2026 12:36:44 +0100</pubDate>
    </item>
    <item>
      <title>Closing the Divide: Bridging Knowledge and Data in Hybrid AI Systems</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13769</link>
      <description>Deep Neural Networks (DNNs) have gained prominence across numerous applications due to their ability to extract complex patterns from data. However, their performance heavily depends on the availability of large, high-quality training datasets, which are often impractical to obtain. In contrast, Knowledge-Driven Systems (KDS) leverage expert-defined rules and can operate with limited data but often suffer from lower accuracy and the complexity of knowledge formulation. While both approaches offer distinct advantages, existing hybrid methods primarily rely on ensemble techniques that combine independent predictions, potentially propagating inaccuracies from individual models. To bridge this gap, we propose Neural-Knowledge Networks (NKN), a novel knowledge fusion framework that synergistically integrates knowledge- and data-driven approaches. NKN explores three fusion strategies across time series and image modalities.&#13;
&#13;
The first strategy, residual fusion for time series forecasting, integrates KDS predictions at the input and output layers of DNNs, enabling the network to learn and correct inaccuracies in expert knowledge rather than directly learning input-output mappings. Evaluated on five time-series datasets, this approach improves upon State-of-the-Art (SotA) methods by 42%, maintaining a 40% improvement even with only half the training data.&#13;
&#13;
The second strategy, latent space fusion also addresses time series forecasting and integrates knowledge into intermediate DNN layers to learn joint representations from both knowledge and data. This approach enhances generalizability and explainability, achieving an average 35% improvement over SotA across nine time-series datasets.&#13;
&#13;
The third strategy extends NKN to the image modality through Guided Object Detection, where a specialized guidance network generates probabilistic spatial maps to direct the DNN toward regions likely to contain objects of interest, improving detection performance by 15% over the baseline.&#13;
&#13;
Overall, NKN demonstrates that complementary integration of expert knowledge and data-driven learning can improve accuracy, reduce data requirements, and enhance model generalizability. By enabling effective AI solutions with limited training data, NKN offers a promising step toward making advanced AI more accessible to resource-constrained industries.</description>
      <author>Muhammad Ali Chattha</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13769</guid>
      <pubDate>Wed, 11 Nov 2026 12:31:51 +0100</pubDate>
    </item>
    <item>
      <title>Baustein Stadtbibliothek - Öffentliche Bibliotheken als strategischer Baustein in der Stadtentwicklung und Impulsgeber für lebendige Stadtzentren</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13763</link>
      <description>Die Dissertation untersucht öffentliche Bibliotheken als Stadtbaustein und fragt aus stadtplanerischer Perspektive nach ihrer räumlichen und sozialräumlichen Wirkung sowie nach den Konsequenzen für planerische Entscheidungen in Städten und Gemeinden. Betrachtet wird die Bibliothek sowohl als physisches Gebäude und Ort als auch im Kontext ihres Standorts und des städtebaulichen Umfelds. Dabei werden Anforderungen an Lage und Gebäudetyp sowie die betrieblichen und institutionellen Rahmenbedingungen analysiert.&#13;
&#13;
Die Untersuchung stützt sich auf eine Auswertung aktueller Fachliteratur zu Kultur- und Bildungseinrichtungen in stadtplanerischen Strategien sowie auf Forschungen zu den strukturellen Rahmenbedingungen öffentlicher Bibliotheken. Auf dieser Basis wurden statistische Daten zu 100 Bibliotheksbauprojekten in Deutschland aus den Jahren 2010–2022 sowie zwölf qualitative Fallstudien ausgewertet. Die qualitativen Fallbeispiele aus Städten unterschiedlicher Größenordnungen sowie mit unterschiedlichen Gebäudetypen (Neubau, Altbau, Umnutzung von Gewerbebauten) dienen dazu, die quantitativen Daten zu vertiefen. In den Steckbriefen der Fallbeispiele werden ortsspezifische räumliche Wirkungen und Standortentscheidungen dargestellt. Gleichzeitig machen sie die Bandbreite an Themen, Akteurskonstellationen und Prozessen sichtbar, die in Städten unterschiedlicher Größenordnung auftreten. Die gewonnenen Erkenntnisse werden in einem weiteren Schritt auf Fragestellungen der Innenstadtentwicklung bezogen, um den potenziellen Beitrag öffentlicher Bibliotheken zu dieser Planungsaufgabe zu bestimmen.&#13;
&#13;
Die Ergebnisse verdeutlichen, dass neue Bibliothekskonzepte mit veränderten Anforderungen an Standortwahl und Gebäudetyp verbunden sind. Je nach Grad ihrer planerischen Einbettung lassen sich die untersuchten Projekte dabei in drei Typen einteilen: vom pragmatischen Lückenfüller über den aus bibliothekarischem Eigeninteresse entstandenen Neubau bis hin zum gezielt gesetzten Stadtbaustein. Zentrale Lagen, die Nähe zu Partnerinstitutionen und die gemeinsame Nutzung von Gebäuden gewinnen deutlich an Bedeutung. An Standorten, die diesen Kriterien entsprechen, steigen die Besuchszahlen messbar an. Bibliotheken können somit zu einem aktiven Baustein der Stadtentwicklung werden, der die Lebensqualität und sozialräumliche Teilhabe unterstützt und Stadträume belebt. Ihr Einfluss entfaltet sich dabei auf mehreren Ebenen: baulich-räumlich, sozialräumlich, funktional sowie im Zusammenspiel der beteiligten Akteur:innen. Die Fallbeispiele belegen zudem, dass Bibliotheken in vielfältigen Kontexten – vom ländlichen Raum über das Quartier bis zur Innenstadt – als wichtige soziale Knotenpunkte wirken und das Gemeinwesen spürbar stärken.&#13;
&#13;
Die Untersuchung macht deutlich, dass sich das Potenzial öffentlicher Bibliotheken nicht von selbst entfaltet, sondern an die Bedingung einer frühen und verbindlichen Einbindung in kommunale Stadtentwicklungsprozesse sowie an politische und finanzielle Unterstützung geknüpft ist. Daraus lässt sich ableiten, dass Planungsprozesse stärker vernetzt und kooperativ gestaltet werden sollten. Die Arbeit schließt mit konkreten Handlungsempfehlungen für Stadtplaner:innen, Kommunen und Bibliotheksleitungen, die aufzeigen, wie öffentliche Bibliotheken strukturell, strategisch und räumlich als integraler Bestandteil gemeinwohlorientierter Stadtentwicklung berücksichtigt werden können.</description>
      <author>Ragna Körby</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13763</guid>
      <pubDate>Wed, 11 Nov 2026 09:00:38 +0100</pubDate>
    </item>
    <item>
      <title>Optimierung der Bioreaktorkultivierung von filamentösen Pilzen zur Produktion eines Proteaseinhibitors</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13766</link>
      <description>Filamentöse Pilze sind bekannte Produzenten bioaktiver Substanzen, von denen viele noch unerforscht sind. Die steigende Nachfrage nach neuen bioaktiven Verbindungen – insbesondere in der pharmazeutischen Industrie – hat das Interesse an diesen Mikroorganismen verstärkt. Natürliche bioaktive Verbindungen besitzen einzigartige Spezifität und Selektivität, was sie zu wertvollen Kandidaten für die Arzneimittelentwicklung macht. In dieser Arbeit wurde Penicillium sp. (IBWF 040-09) auf seine Herstellung eines Proteaseinhibitors untersucht, der auf eine der essenziellen Proteasen von Trypanosoma abzielt. Der Wirkstoff wurde als intrazellulärer Sekundärmetabolit identifiziert; daher wurden die Kultivierungsbedingungen optimiert, um die Pelletbildung statt des Wachstums von dispergiertem Myzel zu fördern. Der Proteaseinhibitor aus Penicillium sp. (IBWF 040-09) wurde in dieser Arbeit erstmals beschrieben und untersucht. Die Kultivierung von Penicillium sp. (IBWF 040-09) wurde in einem Einweg-Rührkesselbioreaktor mit mehreren Anpassungen durchgeführt, da dieser ursprünglich für die Kultivierung von Zellkulturen konzipiert war. Aufgrund der begrenzten Flexibilität des Einweg-Bioreaktors wurde die Kultivierung anschließend auf einen Glas-Bioreaktor übertragen. Dies erlaubte zudem die Entwicklung einer halbkontinuierlichen Kultivierungsstrategie, nämlich der wiederholten Satzkultivierung, um die Raum-Zeit-Ausbeute der Zielsubstanz zu verbessern. Zur Umsetzung dieser Strategie wurde während der Kultivierung eine gewisse Menge an Pilzbiomasse mithilfe eines sogenannten Sporenspeichers zurückgehalten. Die zurückgehaltene Biomasse durchlief zwischen den Satzkultivierungen eine Sporulationsphase, und die dabei entstehenden Sporen dienten als Inokulum für die nächste Satzkultivierung. Diese Strategie der wiederholten Satzkultivierung minimierte morphologische Veränderungen und bewahrte die Pelletform. Darüber hinaus wurde ein automatisierter Kultivierungsprozess etabliert, um den Ernteprozess zwischen den jeweiligen Satzkultivierungen zu rationalisieren, einen zuverlässigen Online-Biomasse-Sensor zu integrieren und eine Reinigung vor Ort unter sterilen Bedingungen nach jeder Satzkultivierung zu ermöglichen. Ein nicht-invasiver Biomasse-Sensor, der auf dem Prinzip der Rückstreulichtmessung basiert, wurde implementiert, und die kritischen Faktoren, die die Signalgenauigkeit beeinflussen, wurden systematisch bewertet. Ein steriles Cleaning-in-Place-Protokoll wurde entwickelt, um die effiziente Entfernung von Restbiomasse zwischen den Kultivierungszyklen sicherzustellen und damit die Reproduzierbarkeit der Bioprozesse zu verbessern. Zusammenfassend lässt sich sagen, dass diese Strategie einen vielversprechenden Ansatz für die langfristige, automatisierte Kultivierung filamentöser Mikroorganismen in Pelletform darstellt.</description>
      <author>Winda Astuti Soerjawinata</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13766</guid>
      <pubDate>Tue, 10 Nov 2026 13:08:51 +0100</pubDate>
    </item>
    <item>
      <title>Hybrid machine learning for enhanced prediction of diffusion coefficients in liquids</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13761</link>
      <description>Diffusion coefficients are key thermophysical properties for modeling mass transport in liquids, but experimental data are scarce, making reliable prediction methods indispensable. In the present work, we introduce a new method for predicting diffusion coefficients of molecular components at infinite dilution in pure liquid solvents by integrating the Stokes–Einstein (SE) equation with machine learning (ML). Unlike previous ML approaches, the resulting hybrid Enhanced Stokes–Einstein (ESE) model provides strictly physically consistent predictions for diffusion coefficients as a function of temperature across a broad range of binary mixtures. Trained and validated using an extensive compilation of literature data for infinite-dilution diffusion coefficients in binary liquid systems, ESE achieves significantly higher prediction accuracies than the previous state-of-the-art model, SEGWE, while requiring only the&#13;
SMILES strings encoding of the molecular formulae of the components of interest as additional inputs, which are always available. This simplicity makes ESE broadly applicable, e.g., for process design and optimization. The ESE model and its source code are fully disclosed and are directly accessible via an interactive web interface at https://ml-prop.mv.rptu.de/.</description>
      <author>Jens Wagner; Zeno Romero; Kerstin Münnemann; Sebastian Schmitt; Thomas Specht; Hans Hasse; Fabian Jirasek</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13761</guid>
      <pubDate>Tue, 10 Nov 2026 08:31:50 +0100</pubDate>
    </item>
    <item>
      <title>The Unpriced Sky - A Framework for a Global Carbon Currency to Advance Environmental and Social Development</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13760</link>
      <description>CORA is a global carbon currency extending the logic of cap-and-trade directly to individuals, allocating every person on the planet an equal annual carbon budget, spendable alongside conventional currency. The income generated by selling surplus allocations would lift nearly the entire lower-income half of the global population above every standard poverty threshold. If the parallels with unconditional cash transfer programs hold, that income transfer would catalyze improvements in health, education, child labor, mental wellbeing, and local economic activity. At the same time, by making the carbon cost of every good and service visible at the point of sale, CORA generates immediate and continuous competitive pressure on producers to reduce emissions. Many of the technical solutions required already exist; they have simply never been commercially attractive enough to justify implementation under conventional market conditions. The effects would extend across virtually every sector of the economy: construction, transport, energy, retail, and the second-hand and repair economy, among many others. The incentive structure would also nudge the informal economy toward formalization, with significant positive effects for tax bases and regulatory reach in lower-income countries. And where educational attainment rises, healthcare becomes accessible, and entrepreneurial resources materialize, the broader developmental trajectory, reduced poverty traps, more diversified economies, growing domestic markets, becomes highly likely.</description>
      <author>Kristina Schmidt</author>
      <category>masterthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13760</guid>
      <pubDate>Fri, 06 Nov 2026 08:31:00 +0100</pubDate>
    </item>
    <item>
      <title>Exploring the links between Self-Compassion and Resilience: Insights from measurement and intervention research in working populations</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13754</link>
      <description/>
      <author>Lisa von Boros</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13754</guid>
      <pubDate>Thu, 05 Nov 2026 15:54:34 +0100</pubDate>
    </item>
    <item>
      <title>Dependency Analysis of the Extra-Functional Properties of the Time-Triggered Architecture</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13758</link>
      <description>The time-triggered paradigm has proven highly suitable for closed safety-critical&#13;
distributed real-time systems with predetermined configurations and stringent timeliness and dependability requirements. The Time-Triggered Architecture (TTA) embodies this paradigm by combining a globally synchronized time base, offline-defined schedules, and architectural services that collectively provide key extra-functional properties, including determinism, temporal predictability, fault containment, and composability. While the TTA is successfully deployed across the avionics, space, rail, and automotive sectors, modern safety-critical systems are increasingly expected to support functionalities of different criticalities, higher levels of connectivity, long and untended operational lifetimes, and resilience against sophisticated cyber threats, motivating a transition toward open and highly interconnected platform architectures. Evolving the TTA to address these emerging requirements is fundamentally challenged by the rigidity of time-triggered operation and the tight coupling among its architectural services and mechanisms. Consequently, modifying or replacing individual services can unintentionally weaken its extra-functional properties. A systematic understanding is therefore required of how individual design decisions contribute to the emergence of these properties and which of these decisions are essential for preserving them. Prior work, however, has primarily focused on individual mechanisms rather than the&#13;
architecture as a whole. In this paper, we address this gap through a qualitative architectural analysis of TTA’s architectural services and the resulting extra-functional properties. The analysis enables system architects seeking to evolve the TTA to reason about the trade-offs involved in achieving these properties by&#13;
distinguishing inherent architectural constraints from those arising from implementation choices. For this purpose, we introduce a classification framework that categorizes the dependencies between architectural services and extra-functional properties as fundamental, implementation-based, or application-dependent. Our analysis shows that most dependencies are implementation-based or application-dependent rather than fundamental, revealing substantial design space for architectural adaptation while preserving TTA’s core safety guarantees.</description>
      <author>Mohammad Ibrahim Alkoudsi; Marcus Völp; Gerhard Fohler</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13758</guid>
      <pubDate>Thu, 05 Nov 2026 13:29:15 +0100</pubDate>
    </item>
    <item>
      <title>APEX1 as a novel component of the  ultrafine anaphase bridge resolution  machinery</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13750</link>
      <description>Replication stress is an important source of chromosomal instability and tumorigenesis because it interferes with faithful DNA synthesis. However, how replication stress affects mitotic chromosome segregation remains incompletely understood, particularly in relation to the formation and resolution of persistent sister chromatid connections such as ultrafine anaphase bridges (UFBs). UFBs are thin, histone-free DNA structures that connect sister chromatids during anaphase. While their exact origin and nature remain poorly characterized, they likely arise from under replicated DNA or from unresolved recombination intermediates. The UFBs can be visualized only through the recruitment of specific UFB-associated proteins, including Bloom syndrome helicase (BLM), the PLK1-interacting checkpoint translocase PICH, and Fanconi anemia proteins such as FANCD2 and FANCI. Although these proteins have been implicated in UFB resolution, the complete molecular mechanism and the identity of additional cofactors remain unclear. In this study, chromatin immunoprecipitation of know UFB-associated proteins coupled to mass spectrometry (ChIP-MS) was used to systematically identify novel proteins associated with UFB resolution factors that arise from mild replication stress. This screen identified several proteins that have not been associated with this type of damage, among them also APEX1, as a potential UFB associated protein. APEX1 is a multifunctional DNA repair and redox regulatory protein best known for its role in base excision repair and has also been identified as an R-loop-associated factor. However, its function in mitotic genome stability has not been well characterized. Functional analysis revealed increased UFB formation, chromatin bridges, and Binucleation in APEX1 knockout cells, suggesting that APEX1 contributes to accurate chromosome segregation during mitosis. Importantly, simultaneous depletion of BLM did not further increase these phenotypes, indicating a possible functional interaction between APEX1 and BLM in the prevention or resolution of persistent sister chromatid linkages. Together, these findings identify APEX1 as a novel factor involved in UFB resolution and suggest that APEX1 contributes to genome stability beyond its canonical role in DNA repair. Overall, this work supports a role for APEX1 in resolving replication stress–induced UFBs and limiting the persistence of fragile site–associated DNA intermediates into mitosis. By potentially cooperating with BLM-dependent processing pathways, APEX1 may help reduce mitotic and post mitotic defects, including, micronuclei, binucleation, and polyploidy.</description>
      <author>Farbod Mohseni</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13750</guid>
      <pubDate>Thu, 05 Nov 2026 09:05:04 +0100</pubDate>
    </item>
    <item>
      <title>From Promise to Practice: Advancing Bacterial Biomonitoring of Salmon Aquaculture</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13746</link>
      <description>The expansion of salmon aquaculture has increased the need for reliable and efficient approaches to assess its environmental impacts. Routine compliance monitoring of salmon farms has traditionally relied on morphology-based assessment of benthic macrofaunal communities, which requires considerable taxonomic expertise, time, and financial resources. Advances in environmental DNA (eDNA) metabarcoding have demonstrated that sediment bacterial communities respond rapidly and consistently to aquaculture-derived organic enrichment, and molecular approaches are beginning to be incorporated into environmental monitoring frameworks. However, their application in routine regulatory assessment remains comparatively limited. Broader implementation requires confidence that bacterial indicators are analytically robust, remain reliable despite natural ecological variability, are reproducible across analytical approaches and datasets, and can ultimately be translated into practical monitoring tools. The overarching aim of this thesis was therefore to evaluate whether sediment bacterial communities can provide reliable measures of benthic ecological condition under the methodological and environmental variability encountered in salmon aquaculture monitoring, and whether robust bacterial indicators can ultimately be translated into practical quantitative tools. To address this aim, bacterial community responses were examined across methodological treatments, temporal and environmental gradients, complementary indicator-selection approaches, and independent datasets. &#13;
Methodological robustness was evaluated by comparing PCR amplification protocols and quantifying sampling-related variability relative to conventional macrofaunal monitoring. Temporal dynamics were investigated across two salmon production cycles to determine the relative influence of aquaculture impact and natural seasonal variability on bacterial community structure. Reproducible bacterial indicators were subsequently identified using three complementary analytical approaches and compared across two independent Scottish aquaculture datasets. Finally, selected indicators were developed into targeted digital PCR (dPCR) assays and quantified across samples representing different levels of benthic ecological quality.&#13;
The findings demonstrated that methodological standardization is a prerequisite for robust bacterial biomonitoring. Although alternative PCR amplification protocols produced substantial differences in bacterial diversity estimates, they yielded highly consistent ecological patterns across the organic enrichment gradient. Likewise, sampling-related variability in bacterial eDNA was comparable to that observed for conventional macrofaunal monitoring, indicating that the molecular approach did not introduce greater background variability than the established monitoring method. This consistency extended beyond methodological variation to temporal ecological dynamics. Across two salmon production cycles, organic enrichment remained the dominant driver of bacterial community structure (59%), whereas temperature variations explained a considerably smaller proportion of community change (1%). Bacterial communities also reflected changes associated with farm management, particularly at intermediate impact levels. &#13;
Importantly, the ecological signal was reproducible at the level of individual indicators. The three indicator-selection approaches converged on highly similar sets of enrichment-responsive taxa within each dataset, and several indicators were recovered across both independent datasets. Selected indicators were successfully quantified using dPCR, and their absolute abundances followed the expected ecological quality gradients across independent salmon farm samples. These results demonstrate that sequencing-derived bacterial indicators can be converted into quantitative molecular targets.&#13;
Overall, this thesis advances bacterial biomonitoring beyond the demonstration of ecological potential towards the requirements of practical application. The findings show that bacterial responses to aquaculture-derived organic enrichment remain interpretable despite important sources of methodological and natural variability, that reproducible indicators can be identified across complementary analytical approaches and independent datasets, and that these indicators can be translated from sequencing data into quantitative molecular targets. Importantly, the contribution of this work lies not in proposing bacterial metabarcoding as a replacement for established benthic assessment, but in strengthening the scientific basis for its integration into routine environmental monitoring. By connecting methodological evaluation, ecological validation, reproducible indicator discovery, and targeted quantification within a single framework, this thesis helps bridge a gap that has persisted between microbial ecology and its practical application in environmental assessment. It therefore contributes both to the development of bacterial bioindicators as reliable indicators of benthic ecological condition and to building the scientific foundation and confidence required to advance molecular biomonitoring “from promise to practice” in salmon aquaculture.</description>
      <author>Mahshid Oladi Kalarijani</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13746</guid>
      <pubDate>Wed, 04 Nov 2026 12:59:25 +0100</pubDate>
    </item>
    <item>
      <title>Material Utilization in Additively Manufactured Layered Systems with a Porous Substrate: A Numerical Case Study of a Thrust Ball Bearing</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13748</link>
      <description>In layered systems with porous substrates and a dense solid surface, stiffness and strength are inherently coupled through porosity-dependent relations, influencing their load-bearing behaviour. This work presents a systematic methodology for the assessment and design of such layer-substrate systems based on a criterion of balanced material utilization. A dimensionless parameter is defined to characterize the stress state in both components relative to their admissible limits, from which the optimal layer thickness is determined at equal stress levels in both constituents. Stress distributions are calculated using a numerical half-space model for layered contacts and evaluated through material-dependent equiva-&#13;
lent stress criteria. The relationship between material utilization and load-carrying capacity is reduced to a scaling factor that combines the influence of porosity-dependent parameters. The approach establishes a direct link between the governing material parameters and structural design variables. Across the investigated parameter range, the utilization rate scales linearly with optimal layer thickness, whereas the load-carrying capacity follows a cubic relation. For a representative Ashby strength scaling coefficient of Cσ = 0.3, for example, a substrate porosity of 90% leads to a scaling factor of 1.6, corresponding to a&#13;
possible load amplification of 60% relative to the homogeneous reference.</description>
      <author>Olaf Grutza; Simon Graf; Stefan Paulus; Stefan Thielen; Oliver Koch</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13748</guid>
      <pubDate>Wed, 04 Nov 2026 11:50:53 +0100</pubDate>
    </item>
    <item>
      <title>Design of Energy‑Efficient Ultra‑High‑Throughput LDPC Decoders for Future Communication Systems</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13744</link>
      <description>Forward error correction (FEC) is essential for reliable communication, but decoders are major contributors to computational complexity and power consumption. Among modern FEC codes, many communication standards adopt low-density parity-check (LDPC) codes because they can provide excellent error-correction performance and support high-throughput decoder architectures. At the required throughputs, only application-specific integrated circuit implementations are feasible. However, as semiconductor technologies continue to evolve, implementation complexity increasingly arises from interconnects and memory rather than arithmetic operations. Moreover, transistor density is increasing faster than power is improving, resulting in higher power density. This limits further throughput increase for a fixed thermal design power. Therefore, the central question is how to leverage the strengths of LDPC decoders while addressing the limitations imposed by silicon implementation.&#13;
&#13;
This thesis presents new algorithmic and architectural approaches for designing energy-efficient, high-throughput LDPC decoders for future communication systems. My contributions to the established research on LDPC decoding are organized into three areas. First, I investigate the conflicting design principles of entropy and locality in LDPC decoders and propose a methodology to assess their impact on the implementation complexity of high-throughput decoders. Second, I analyze two principal strategies for representing exchange messages in LDPC decoders at the architectural level, comparing node and decoder architectures. The results show that finite-alphabet (FA) decoders outperform conventional decoders. On the algorithmic level, I propose novel techniques for designing lookup tables for FA decoders tailored to the underlying architecture, balancing error-correction performance with implementation complexity. Finally, I propose adaptive sliding window decoding for spatially coupled LDPC codes, translating channel variability into power savings. All contributions are validated against state-of-the-art using convergence analysis, extensive Monte Carlo simulations, and post-place-and-route “virtual silicon” implementations.</description>
      <author>Oliver Griebel</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13744</guid>
      <pubDate>Wed, 04 Nov 2026 10:15:56 +0100</pubDate>
    </item>
    <item>
      <title>Analysis of associative olfactory learning and memory in Drosophila</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13743</link>
      <description>Throughout their lives organisms encounter a wide range of complex sensory stimuli and must be able to flexibly adjust their behavior. Innate behavior, which is genetically programmed, differs from learned behavior, which relies on experience. The output of both behavioral pathways is integrated to generate appropriate responses to environmental cues. The detection and evaluation of these stimuli are essential for survival and show remarkable similarities between vertebrates and insects, especially in the olfactory system. Memory enables organisms to use past experiences by storing relevant information that can be accessed later to guide behavior and adapt to future changes. In Drosophila melanogaster the mushroom body (MB) acts as the key brain region for learning and memory formation. Through associative learning, where sensory cues are linked with either positive or negative reinforcement, behavioral changes can become lasting. The establishment of long-term memory in the MB involves local molecular adaptation and synaptic remodeling.&#13;
In this thesis I investigate the formation and function of memory traces in the olfactory system of the fruit fly. Central to this study is the CRE-activity-dependent memory engram label (CAMEL)-tool, a genetic reporter that labels and manipulates memory-relevant Kenyon cells (KCs). I demonstrated that the number of fluorescently labeled CAMEL cells was modulated dynamically by sensory input. Environmental enrichment led to a significant increase in CAMEL-labeled KCs within the MB, indicating that CAMEL activity reflects experience-dependent memory encoding. Optogenetic reactivation of these CAMEL-cells after exposure to specific stimuli resulted in behavioral changes, confirming their functional relevance.&#13;
To investigate molecular changes within individual engram cells, I combined the fluorescent non-canonical amino-acid tagging (FUNCAT) technique with the CAMEL-tool to visualize newly synthesized proteins. FUNCAT relies on the incorporation of a methionine analog into nascent proteins, which can subsequently be fluorescently labeled using bioorthogonal click chemistry. This approach enables sensitive and specific detection of protein synthesis at the single-cell level, allowing precise analysis of translational activity within defined neuronal populations without relying on conventional antibody staining.&#13;
I established an optimized immunohistochemical labeling protocol to improve the detection of sparsely labeled neuronal populations by increasing signal-to-noise ratio and overall staining consistency. This enabled more reliable detection of sparsely labeled neuronal populations and helped to overcome limitations observed with previous protocols.&#13;
In an artificial aversive conditioning paradigm the odor cue was replaced by optogenetic activation of CAMEL cells during training. In the subsequent light preference test using the circular arena, we were not able to recall artificial aversive memory using this approach.&#13;
Finally, I optogenetically manipulated dopaminergic signaling pathways by expressing a light-activatable variant of the dopamine receptor D1R1 in MB neurons. I confirmed proper subcellular localization of the receptor, and behavioral experiments demonstrated that D1R1 function in KCs is required for learning. However, when the electric shock was replaced by light-induced activation of D1R1 during odor presentation, this artificial activation alone was insufficient to induce associative memory formation.</description>
      <author>Vanessa Burre</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13743</guid>
      <pubDate>Fri, 30 Oct 2026 11:20:50 +0100</pubDate>
    </item>
    <item>
      <title>Fuzzy Cognitive Maps for Wastewater Treatment Selection: Constructed Wetlands vs. Conventional Plants</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13740</link>
      <description>The Fuzzy Cognitive Map (FCM) framework provides a useful tool for representing the complex interdependencies involved in wastewater treatment selection, particularly when social, ecological, climatic, and economic criteria are considered simultaneously. In this study, the FCM approach was applied to compare two wastewater treatment approaches in Ahvaz, Iran: constructed wetlands (CWs) as a nature-based solution and energy-based wastewater treatment plants. The developed model included 30 components and 127 causal links, and was used to examine four scenarios representing CWs, energy-based treatment, a hybrid approach, and direct wastewater discharge. The results showed that both CWs and energy-based solutions had similar effects on public health, while the hybrid scenario produced the greatest improvement. CWs had a positive effect on ecosystem restoration and showed better performance in heavy metal removal, whereas energy-based solutions had a greater negative influence on environmental conditions and climate-related components. In addition, the economic results indicated that CWs were more favorable in terms of capital cost, energy consumption, and operational cost. Sensitivity analysis using ±10% variations in causal weights showed that the main scenario-response patterns remained generally unchanged. Overall, the findings suggest that CWs and energy-based systems each have specific advantages and limitations, while the hybrid approach offers the most balanced performance across the evaluated criteria. This study demonstrates the usefulness of the FCM approach for supporting wastewater management decisions and for identifying trade-offs among treatment alternatives in sustainable water resource planning.</description>
      <author>Mohamad Azizipour; Narges Baahmadi; Amin E. Bakhshipour; Ulrich Ditmer</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13740</guid>
      <pubDate>Fri, 30 Oct 2026 11:07:28 +0100</pubDate>
    </item>
    <item>
      <title>On algorithmic and structural aspects of Roman domination in graphs</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13739</link>
      <description>Nowadays, the problem of distributing limited resources is more relevant than ever.&#13;
One needs to find answers to the question how to allocate resources such that all relevant regions are covered.&#13;
Additionally, there may be different kinds of demands.&#13;
This problem can mathematically be modelled by variants of the dominating set problem.&#13;
&#13;
The dominating set problem is a well-known problem in combinatorial optimisation.&#13;
The goal is to find a minimum size subset of the vertices of a graph such that every vertex not in the set has a neighbour in the set.&#13;
There is a vast amount of papers on this topic and its variants.&#13;
The Roman domination problem is a generalisation of the dominating set problem that has been investigated since the early 2000's.&#13;
For a graph \(G\) one seeks to find a map that assigns each vertex a number in \(\{0,1,2\}\) such that every vertex with value \(0\) has a neighbour with value \(2\); then this map is called a Roman dominating function.&#13;
The goal is to find such a function with the minimum sum of its function values (the Roman domination number \(\gamma_R\)).&#13;
In the resource distribution setting, that is, each region needs to be covered by itself (values \(1\) and \(2\)) or it is acceptable if it is not but then there has to be a neighbouring region that has "level \(2\)" kind of resources.&#13;
This thesis adds more results to variants and related problems regarding the Roman domination problem.&#13;
&#13;
In the first variants we study the behaviour of \(\gamma_R\) for different subtrees of a graph.&#13;
We show that among the spanning trees of a graph the one with minimum possible \(\gamma_R\) has exactly the Roman domination number of the graph itself.&#13;
Finding a Steiner tree with minimum possible \(\gamma_R\) turns out to be, in general, NP-hard.&#13;
Both problems become polynomial-time solvable on graphs of bounded tree-width, as we show by giving an algorithm.&#13;
When looking at the maximum possible \(\gamma_R\) of a spanning tree, the situation is different.&#13;
We show that the ratio of the maximum \(\gamma_R\) of a spanning tree and the Roman domination number of the graph grows linear in the number of vertices of the graph.&#13;
&#13;
As a related problem, we consider the dual problem of Roman domination in a linear optimisation sense.&#13;
We arrive at the \(2\)-limited packing problem in which one wants to find a maximum size subset of vertices of a graph such that in each closed neighbourhood at most two vertices are contained in said subset.&#13;
This problem has been introduced in 2010 as a special case of the \(k\)-limited packing problem.&#13;
We compare this problem to Roman domination and as a main result show strong duality on trees.&#13;
&#13;
Another variant is simultaneous Roman domination where given different graphs we want to Roman dominate each single graph with one shared function.&#13;
We show that this problem is NP-hard even when there are only two graphs given which both are paths.&#13;
For simultaneously Roman dominating all spanning trees of a graph we characterise all graphs where the optimal solution value equals the number of vertices.&#13;
&#13;
Finally, we investigate the Roman domination polynomial which was first introduced in 2021.&#13;
We show some first results directing to the unimodality of said polynomial and investigate the behaviour of the polynomial evaluated at \(-1\).</description>
      <author>Helena Weiß</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13739</guid>
      <pubDate>Thu, 29 Oct 2026 09:55:09 +0100</pubDate>
    </item>
    <item>
      <title>Realbrandversuchs zur brandschutztechnischen Untersuchung von Aufstockungen in Holzbauweise</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13519</link>
      <description>Um schnell Wohnraum zu schaffen, der überwiegend &#13;
im innerstädtischen Bereich benötigt wird, bietet sich &#13;
eine Aufstockung in einer Holzrahmen-/ &#13;
Holztafelbauweise in besonderer Weise an. Ein &#13;
wesentliches Augenmerk ist dabei jedoch auf den &#13;
Brandschutz zu legen. Durch die brennbare &#13;
Eigenschaft des Holzes sind vor allem &#13;
Überschreitungen der Geschosshöhen in die &#13;
Gebäudeklassen 4 und 5 zu beachten. Mithilfe eines &#13;
Realbrandversuches wurden konstruktionsbedingte &#13;
Schwachstellen, wie die Anschlussfuge der &#13;
Holzbaugeschosse auf die Bestandsbauteile &#13;
untersucht um eine mögliche Rauchgasverschleppung &#13;
ausschließen zu können. Auch die Einbindung von &#13;
Baustoffen aus natürlichen Rohstoffen war Teil des &#13;
Realbrandversuchs, um dessen Auswirkungen auf &#13;
einen Brandverlauf mit beurteilen zu können.</description>
      <author>Sabine Scheidel; Tobias Götz; Sebastian Dienst; Oliver Kornadt</author>
      <category>conferenceobject</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13519</guid>
      <pubDate>Wed, 28 Oct 2026 08:33:35 +0100</pubDate>
    </item>
    <item>
      <title>Pool-scAN: Experimental Setup for Student-Oriented Wet-Chemical Analysis to Detect Pool Chemicals in Pool Water within a Game-Based Inquiry Learning Approach</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13733</link>
      <description>Inquiry-based learning (IBL) has already proven to be an effective approach to promote scientific inquiry skills. However, its implementation in school practice faces several persistent challenges such as limited time, necessary support measures for students, difficulties in assessment or rigid curricula. To address these barriers, a novel experimental setting─Pool-scAN─in a game-based inquiry learning (GBIL) approach was developed for upper secondary chemistry education. In this setting, students investigate a fictional murder case with integrated supportive game elements (e.g. story, feedback, competition, guidance) as forensic scientists through qualitative and quantitative wet-chemical analyses of highly authentic pool water samples. The setting was evaluated with n = 51 high school students in Germany using a pre-post design based on an adapted R-Comp questionnaire assessing research competence according to the RMRC-K model. Results revealed a highly significant improvement in students’ perceived overall research competence (d = 1.163***) across all subscales, with the largest effect observed for domain knowledge. Considering the limitations of self-assessments, findings indicate that game-based learning combined with inquiry-based learning is beneficial for research competence development. Furthermore, the results suggest that the adapted R-Comp questionnaire could enable a simple and quick assessment of research competence in schools. On the other hand, an effective experimental learning environment on the topic of wet chemical analysis was developed using the novel context of pool chemistry in the game-based inquiry learning approach. For this reason, the specific influence of game-based learning on research competence and other variables such as motivation or cognitive load of students within the learning environment will be examined in more detail in further studies.</description>
      <author>Nicolas Wunn; Daniela Becker; Nils Fitting; Johann-Nikolaus Seibert</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13733</guid>
      <pubDate>Mon, 26 Oct 2026 14:03:35 +0100</pubDate>
    </item>
    <item>
      <title>Untersuchung des raumseitigen Wärmeübergangs an Bauteiloberflächen und Entwicklung eines Rechenmodells sowie einer effizienten Simulationsmethode</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13736</link>
      <description>Die Bestimmung des raumseitigen Wärmeübergangs ist eine zentrale Voraussetzung für die Vorhersage von Bauteiloberflächentemperaturen und Wärmeströmen. Sie spielt eine wesentli-che Rolle bei energetischen und feuchtetechnischen Berechnungen sowie bei der Bewertung der thermischen Behaglichkeit. Bisherige Erkenntnisse und Berechnungsmodelle sowie norma-tive Vorgaben zum raumseitigen Wärmeübergang berücksichtigen realistische Raumszenarien jedoch nur unzureichend. In der Realität hängt der raumseitige Wärmeübergang von verschie-denen Einflussgrößen des jeweiligen Raumszenarios ab, darunter z.B. die Raumgeometrie, die thermischen Eigenschaften der Gebäudehülle und die Strategien zur Raumtemperierung.&#13;
Die vorliegende Arbeit gliedert sich in zwei Hauptteile. Der erste Teil umfasst die Analyse des raumseitigen Wärmeübergangs an der Außenwand sowie die Entwicklung eines einfachen Be-rechnungsmodells für stationäre und quasistationäre Raumszenarien auf Basis stationärer nu-merischer Strömungssimulationen (CFD) mit dem Reynolds-gemittelten Navier-Stokes (RANS) Modell. Da instationäre CFD-Simulationen im bauphysikalischen Planungskontext für Räume mit realen Abmessungen und für Langzeitsimulationen aufgrund des hohen Rechenaufwands nicht praktikabel sind, befasst sich der zweite Teil mit der Entwicklung einer effizienten Simula-tionsmethode zur Bestimmung des transienten lokalen Wärmeübergangs.&#13;
Zunächst wurden die RANS-Simulationen anhand von Referenzdaten aus messtechnischen Untersuchungen, die sowohl Strömungsmessungen mittels Particle Tracking Velocimetry als auch thermische Messungen des Temperaturfeldes der Raumluft und der Wärmeströme an der Wandoberfläche umfassten, validiert. Anschließend wurden umfangreiche Raumszenarien si-mulativ untersucht. Dabei wurden Differenzen von bis zu 76 % gegenüber den normativen Vor-gaben festgestellt. Auf Basis der gewonnenen Simulationsdaten wurden die Eigenschaften des raumseitigen Wärmeübergangs detailliert analysiert. Zudem wurden raumszenarienabhängige Korrelationsmodelle zur realitätsnahen Beschreibung des flächengemittelten Wärmeübergangs abgeleitet und praktische Handlungsempfehlungen formuliert.&#13;
Für stark instationäre Situationen, beispielsweise bei kurzfristigen Änderungen der Raumtempe-ratur oder bei stark transienten Außenbedingungen, wurde ein effizientes Simulationsverfahren für den raumseitigen Wärmeübergang auf Basis von Fast Fluid Dynamics (FFD) entwickelt. Die FFD wurde in drei Aspekten optimiert. Erstens wurde die Genauigkeit verbessert, sodass die-selbe Genauigkeitsgrößenordnung wie bei konventionellen transienten CFD-Verfahren erreicht werden konnte. Zweitens wurde der Simulationsaufwand weiter reduziert. Im Testfall war die optimierte FFD dadurch bis zu 76-mal schneller als die konventionelle transiente CFD-Simulation. Drittens wurde die Anwendbarkeit verbessert, indem die Simulation des konjugierten Wärmeübergangs durch Kopplung mit einer thermischen Festkörpersimulation ermöglicht wur-de.&#13;
Diese Arbeit liefert detaillierte Erkenntnisse zum raumseitigen Wärmeübergang an Bauteilober-flächen in realistischen Raumszenarien. Die entwickelten Rechenmodelle und Simulationsan-sätze können in Nachweisverfahren und Simulationsrechnungen zum raumseitigen Wärme-übergang eingesetzt werden und ermöglichen dabei unter Berücksichtigung realistischer Raumszenarien eine höhere Genauigkeit als bisherige Rechenmodelle sowie eine deutlich ver-besserte rechnerische Effizienz gegenüber bisherigen Simulationsansätzen.</description>
      <author>Zhenming Peng</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13736</guid>
      <pubDate>Mon, 26 Oct 2026 12:36:08 +0100</pubDate>
    </item>
    <item>
      <title>Direct Laser Writing of Ferromagnetic Nickel Utilizing the Principle of Sensitized Triplet–Triplet Annihilation Upconversion</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13726</link>
      <description>Direct laser writing of ferromagnetic microstructures is of great interest for sensing and data storage in compact three-dimensional architectures. However, reliable direct laser writing of metallic, and even more so ferromagnetic, materials remains a major challenge. Here, we discuss a developed photoresist suitable to direct laser write ferromagnetic nickel based on sensitized triplet–triplet annihilation upconversion. By combining an in situ photochemical deoxygenation process with a sensitized triplet–triplet annihilation upconversion process and photoreduction of Ni\(^{2+}\) ions, the deposition of nickel-containing structures is enabled under ambient conditions. Using this approach, nickel-based structures are fabricated as a proof of concept. Scanning electron microscopy and EDX analysis indicate spatially confined deposition of nickel, while magnetic characterization by vibrating sample magnetometry and scanning NV magnetometry confirm ferromagnetic behavior of the printed structures. This work provides a possible route toward extending direct laser writing to metallic and especially ferromagnetic materials.</description>
      <author>Kristin E. J. Kühl; Katharina Rediger; Nikhita Khera; Ephraim Spindler; Gereon Niedner-Schatteburg; Elke Neu; Mathias Weiler; Georg von Freymann</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13726</guid>
      <pubDate>Thu, 22 Oct 2026 15:05:04 +0200</pubDate>
    </item>
    <item>
      <title>Immediate effects of the retrocapital pad and toe bar in sensorimotor foot orthoses on the electromyographic activity of the gastrocnemius muscle in healthy adults</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13732</link>
      <description/>
      <author>Stephan Becker; Jonas Dully; Alessia Arcieri; Lukas Maurer; Carlo Dindorf; Michael Fröhlich</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13732</guid>
      <pubDate>Thu, 22 Oct 2026 14:55:49 +0200</pubDate>
    </item>
    <item>
      <title>Infobrief FBK 79/26</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13708</link>
      <description>Informationen aus dem Lehrstuhl für Fertigungstechnik und Betriebsorganisation</description>
      <author/>
      <category>periodicalpart</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13708</guid>
      <pubDate>Wed, 21 Oct 2026 08:22:16 +0200</pubDate>
    </item>
    <item>
      <title>Grievance forums in autocracies: Surrogates for protest or instruments of control?</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13725</link>
      <description>This article examines Russia’s Regional Governance Centres, state-run regional offices that process citizen complaints via social media, and the conditions under which citizens use these channels to articulate grievances. From an affordance perspective, I argue that citizens view opposition-led protest networks and state-run grievance portals as alternative channels for expressing grievances, choosing between them based on perceived efficacy and risk. Building on theories of collective action and selective incentives, I test whether the full-scale invasion of Ukraine and the resulting rise in the costs of public protest shifted available affordances towards state-managed digital complaint systems offering individualised, lower-risk solutions. Using original dataset of VKontakte messages, I show that regional protest intensity positively correlates with citizen reliance on this digital grievance system. Using an interactive fixed-effects design, I find no statistically significant differences in complaint posting after the 2022 invasion between regions with stronger and weaker pre-war protest records. Overall, the results show that digital grievance platforms and street protest form complementary repertoires of claim-making under ordinary conditions, although the evidence is strongest in cross-regional comparisons rather than in more demanding fixed-effects specifications. The analysis does not provide evidence that digital grievance channels substitute for street protest under harsher repression.</description>
      <author>Aidar Zinnatullin</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13725</guid>
      <pubDate>Mon, 19 Oct 2026 17:28:16 +0200</pubDate>
    </item>
    <item>
      <title>What drives affective polarization? A systematic literature review</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13724</link>
      <description>This article synthesizes causal evidence on affective polarization through a systematic literature review spanning 2012–2025. Using an emergence perspective as an interpretive lens, we catalog 135 peer-reviewed studies located via Google Scholar and Scopus and retained using SCImago Journal Rank indexing (PRISMA-consistent). We code each article for outcome (vertical attitudes toward parties/elites vs. horizontal attitudes toward citizens), identification strength, measures, setting, and five driver families: psychological factors, media and information environment, identity and group processes, institutional factors, and socioeconomic factors. The corpus is highly U.S.-centric, but it also includes cross-national and diverse country cases. Horizontal and vertical outcomes are studied in comparable numbers, yet their joint dynamics are rarely examined. Cross-sectional associative studies account for the largest share of the corpus (44.4%), followed by experimental designs (28.9%) and panel designs (20.0%). Quasi-experimental studies remain rare (3.0%), and other approaches account for 3.7%. Explanations centered on psychology, media exposure/content, and identity processes dominate; institutional and socioeconomic mechanisms remain comparatively thin. Scholarly output rises sharply after 2020, particularly in areas related to psychological and media drivers, with recent growth in institutional accounts. We highlight research needs, including comparative designs that leverage variation in media and party systems, integrative analyses of horizontal-vertical interactions, more rigorous and comparable measurement, and longitudinal/panel studies to link short-run shocks with slow-moving trends.</description>
      <author>Manuel Josef Kogel; Aidar Zinnatullin</author>
      <category>article</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13724</guid>
      <pubDate>Mon, 19 Oct 2026 17:20:38 +0200</pubDate>
    </item>
    <item>
      <title>Interaction-Aware Maneuver and Lane Change Prediction Using Deep Learning Architectures for Autonomous Driving</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13713</link>
      <description>Maneuver prediction is a critical component of decision-making in autonomous driving, particularly in complex urban environments where traffic participants continuously interact under uncertainty. Accurate and interpretable prediction of future driving maneuvers, such as lane changes, is essential for safe and anticipatory planning. This thesis investigates interaction-aware maneuver prediction from the ego-vehicle perspective using deep learning-based approaches that integrate spatial reasoning, temporal modeling, and multi-modal perception.&#13;
The research presents a progression of architectures that address maneuver and lane change prediction at increasing levels of complexity. First, interaction-aware models based on graph neural networks (GNNs) and recurrent neural networks (RNNs) jointly capture spatial interactions between traffic agents and the temporal evolution of driving behavior for scene understanding. Building on this, a cognitively inspired two-layer hierarchical decision-making framework separates planned maneuvers from stimuli and cause reasoning to improve interpretability. Subsequently, a deep multi-modal architecture based on Edge-Featured Graph Attention Networks (EGATs) is developed for lane change prediction, integrating visual perception, spatial interaction graphs, ego-vehicle state signals, and lane-level information. To support this research, a new in-house multi-modal dataset, WylonSet, is introduced and later extended to WylonSet++, providing high-quality annotations for lane change maneuvers and, in its extended version, for maneuver phases in diverse urban traffic scenarios. Finally, a transformer-based multi-modal architecture is proposed for multi-horizon lane change and phase prediction. The model employs attention-based fusion across modalities and incorporates uncertainty estimation at the modality, fusion, and prediction levels, enabling interpretable and future-aware behavior prediction. Experiments on four real-world driving datasets show that interaction-aware, multi-modal, and attention-based architectures consistently improve prediction accuracy and temporal consistency compared with uni-modal and non-interaction-aware variants. The transformer-based architecture predicts lane change maneuvers and their phases across horizons of up to 3 s, achieving an overall macro F1-score above 97% on WylonSet++ and detecting lane change events on average about 1 s before their onset.&#13;
Overall, this thesis contributes new models, datasets, and evaluation methodology for maneuver and lane change prediction, providing a foundation for interpretable and deployment-oriented decision-making systems in autonomous driving.</description>
      <author>Petrit Rama</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13713</guid>
      <pubDate>Mon, 19 Oct 2026 16:05:51 +0200</pubDate>
    </item>
    <item>
      <title>Deep Persistent Homology</title>
      <link>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13728</link>
      <description>The integration of deep learning and algebraic topology opens new perspectives for the analysis of complex data. Autoencoders, a central method of unsupervised learning, can map high-dimensional data spaces to lower dimensions in a nonlinear way, yet it remains unclear which topological invariants are preserved in the process.&#13;
This dissertation investigates autoencoders through the lens of algebraic topology. Persistent homology is employed to systematically analyze the representations learned by autoencoders. The central question is whether autoencoders are capable of inducing isomorphisms between the persistent homology groups of the relevant spaces. Recent advances from fiber theory, together with newly developed bounds on the interleaving distance, enable a precise mathematical characterization of the behavior of such models. Both the computation of persistent homology over field coefficients and the application of global persistent homology in the deep learning context encounter fundamental limitations. One of these limitations is examined through a combined mathematical and computational–experimental study of torsion phenomena in the autoencoder setting. As a further solution perspective, it is proposed to embed local persistent homology into neural architectures via the persis&#13;
tent Mayer–Vietoris sequence. In this way, the dissertation contributes to the&#13;
mathematical foundations of topological deep learning.</description>
      <author>Maria Walch</author>
      <category>doctoralthesis</category>
      <guid>https://kluedo.ub.rptu.de/frontdoor/index/index/docId/13728</guid>
      <pubDate>Fri, 16 Oct 2026 14:14:43 +0200</pubDate>
    </item>
  </channel>
</rss>
