@inproceedings{SchauerSimbeckKuehn, author = {Schauer, Sophie and Simbeck, Katharina and K{\"u}hn, Patricia}, title = {Generative AI in Museums: Design Concepts for an Accessible Digital Mediation Station for a Renaissance Altar}, series = {Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering}, booktitle = {Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering}, publisher = {Springer}, address = {Heidelberg}, doi = {https://doi.org/10.1007/978-3-031-97254-6_20}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22703}, pages = {283 -- 295}, abstract = {Exploring the potential of artificial intelligence (AI) in museum settings, this study finds concepts for a digital and inclusive mediation station giving context to a Renaissance sandstone altar at the Staatliches Museum Schwerin. Today, many museum visitors have less theological knowledge and come from diverse, often non-Christian, backgrounds. Thus, understanding and appreciating historical artworks and artefacts becomes more difficult. Therefore, exploring ways to present cultural heritage in a manner that is understandable, playful, and inclusive, regardless of the audience's origin or beliefs, is necessary. Through a collaborative week-long workshop involving communication design students, conceptual frameworks and prototype designs were iteratively developed and evaluated, laying the groundwork for further refinement and subsequent integration into the museum's permanent exhibition.}, language = {en} } @inproceedings{SchauerKatharinaFernsel, author = {Schauer, Sophie and Katharina, Simbeck and Fernsel, Linda}, title = {The System Admin's Perspective: A Discussion on AI in Education with LMS Admins}, series = {International Conference on Information Technology Based Higher Education and Training (ITHET)}, booktitle = {International Conference on Information Technology Based Higher Education and Training (ITHET)}, publisher = {IEEE}, doi = {https://doi.org/10.1109/ITHET61869.2024.10837664}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22759}, pages = {1 -- 5}, abstract = {Many higher education institutions (HEIs) are implementing or reviewing the implementation of predictive Learning Analytics to improve learning processes, identify students at risk or provide personalized learning paths. The responsibility for selecting, reviewing, implementing and supporting those systems falls on system administrators, an under-researched stakeholder group of Artificial Intelligence in higher education. In this paper, we summarize qualitative insights from a workshop with system administrators in German HEIs. We find that the system administrators are highly aware not only of system requirements but also of the needs of various shareholder groups such as institutional leadership, learners and educators and that they put high emphasis on ethical, transparent and compliant system use. We conclude that system administrators should be involvedmore in research on the use of technology in education and that AI systems used in education need to provide possibilities to sufficiently test the system, including anonymous yet realistic test scenarios and data.}, language = {en} } @unpublished{Hafner, author = {Hafner, Ralf}, title = {Stada Revisited: A Back-of-the-Envelope Return Analysis}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22625}, pages = {1 -- 3}, abstract = {Dieser Beitrag kn{\"u}pft an meine 2017 erschienene Ver{\"o}ffentlichung zu Stada (https://opus4.kobv.de/opus4-htw/344) an. In dieser untersuchte ich, unter welchen Voraussetzungen die {\"u}blicherweise geforderten Eigenkapitalrenditen f{\"u}r Private-Equity-Investoren erreicht werden k{\"o}nnen. Nachdem Stada im Jahr 2025 ver{\"a}ußert wurde, ist es nun an der Zeit f{\"u}r eine erneute Analyse. Wie bereits 2017 basiert diese ausschließlich auf {\"o}ffentlich zug{\"a}nglichen Daten.}, language = {en} } @article{ZeitlerSchwarzkopf, author = {Zeitler, Marlene and Schwarzkopf, Julia}, title = {Assessing human rights due diligence in supply chains}, series = {Management Decision}, journal = {Management Decision}, publisher = {Emerald Publishing}, doi = {https://doi.org/10.1108/MD-03-2024-0583}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22639}, pages = {1 -- 19}, abstract = {Purpose - This study aims to broaden the understanding of auditing practices of sustainability initiatives that promote sustainability standards in supply chains. We examine the extent to which the UN Guiding Principles on Business and Human Rights are embedded in such auditing practices to support the legitimacy of firms' supply chain efforts. Design/methodology/approach - Building on legitimacy theory, this study explores the role and limitations of auditing practices to contribute to the legitimacy of corporate sustainable supply chain efforts. We analyze sustainability audit protocols of selected initiatives to determine their alignment with the UN Guiding Principles. Findings - We identify key auditing gaps, specifically concerning the provision of guidance, stakeholder engagement, impact assessment, and effectiveness measurement. Our findings extend discussions on legitimacy theory by emphasizing that formally adopting auditing practices calls into question the legitimacy of common sustainable supply chain practices. Research limitations/implications - The study emphasizes that legitimacy requires substantive alignment with human rights standards, which encourages greater scholarly attention to this normative dimension. We provide recommendations for practitioners to better integrate normative standards into their auditing practices. However, our research is built on a limited sample of published audit protocols, suggesting further research on additional initiatives and their practices. Originality - This article uses academic research on sustainable supply chain practices to highlight the limitations of initiatives' auditing practices in providing normative guidance for effective human rights due diligence.}, language = {en} } @article{KafedjiskaLeCorreKoebleretal., author = {Kafedjiska, Ivona and Le Corre, Vincent M. and K{\"o}bler, Hans and Levine, Igal and Schlatmann, Rutger and Lauermann, Iver}, title = {Investigation of interfacial charge-carrier dynamics, degradation, and recombination mechanisms in single-junction perovskite solar cells with NiOx and SAM hole-transporting layers via steady-state drift-diffusion model simulations}, series = {Sustainable Energy \& Fuels}, volume = {10}, journal = {Sustainable Energy \& Fuels}, number = {4}, publisher = {The Royal Society of Chemistry}, issn = {2398-4902}, doi = {10.1039/d5se00474h}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22345}, pages = {1174 -- 1191}, abstract = {We investigate the stability and the degradation pathways in single-junction perovskite solar cells with four varying hole-transporting layers (HTLs): pure nickel oxide (NiOx) and copper-doped (NiOx:Cu), with or without self-assembled monolayer (SAM) surface passivation. The cells are aged in a continuous MPP-tracking set-up in a nitrogen environment at 25 °C and the JV curves prior to and after the aging are fitted via drift-diffusion simulations. By using a set of experimentally-measured input parameters and correlating the results from the experiments with the simulations, we are able to test the reliability of the model and then extract important information about the interfacial charge-carrier dynamics, recombination, and degradation mechanisms in the solar cells. We find that NiOx induces severe electron trapping and poor band alignment at the NiOx-perovskite interface, thereby leading to the highest quasi-Fermi level splitting to open-circuit voltage (QFLS-Voc) offset among all the HTLs. As the cells age, the density of bulk traps when NiOx, NiOx:Cu, and NiOx:Cu + SAM are used increases by factors of 36, 3, and 8, respectively, while for NiOx + SAM it remains unchanged. For all of the HTLs, the non-radiative Shockley-Read-Hall (SRH) recombination via surface traps is the dominant recombination mechanism, as it is around 2-3 orders of magnitude higher than the direct or bulk-SRH recombination pathway. Additionally, NiOx exhibits an around 2 orders of magnitude higher rate of SRH interfacial recombination compared to the other three HTLs. However, as the cells age, the rate of the interface SRH recombination remains relatively stable, but the bulk SRH recombination increases by an order of magnitude in all cells, indicating that the degradation of the cells is directly proportional to the increase of the trap-assisted recombination in the perovskite bulk and its degradation. Finally, we investigate the correlation between the hysteresis factor (HF) and the ion concentration. We find that the devices with NiOx have the highest HF and the highest negative-ion concentration, in good agreement with the finding of electron trapping and the highest trap-assisted recombination rate for the NiOx samples. Combining all of this information, we can explain why NiOx is the least stable HTL among all the HTLs (15\% loss in the initial PCE) and how its stability can be improved with Cu doping (8\% loss in the initial PCE) and, to an extent, with SAM passivation (around 11\% loss in the initial PCE).}, language = {en} } @article{BergoBousounisToVuetal., author = {Bergo, Veronica and Bousounis, Pavlos and To Vu, Giang and Dout{\´e}, M{\´e}lodie and Polyzou, Aikaterini and Lalioti, Maria-Eleni and Grigorash, Bogdan B. and Tsurkan, Lyudmila and Morchel, Nicholas and Deboutte, Ward and Brau, Fr{\´e}d{\´e}ric and Manke, Thomas and Sagar, and Medyouf, Hind and Bulavin, Dmitry V. and Cabezas-Wallscheid, Nina and Derecka, Marta and Trompouki, Eirini}, title = {Lack of MDA5 delays hematopoietic aging by modulating inflammaging and proteostasis in mice}, series = {Nature Communications}, volume = {17}, journal = {Nature Communications}, number = {1}, publisher = {Nature Publishing Group UK}, address = {London}, issn = {2041-1723}, doi = {10.1038/s41467-026-69424-x}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22339}, abstract = {"Inflammaging", the chronic increase in inflammatory signaling with age, remains poorly understood in hematopoietic aging. Here, we identify the innate immune RNA sensor melanoma differentiation-associated protein 5 (MDA5) as an important factor of hematopoietic stem cell (HSC) aging. Aged Mda5 -/- mice exhibit reduced HSC accumulation and myeloid bias. Importantly, aged Mda5 -/- HSCs retain greater quiescence and superior repopulation capacity in noncompetitive transplants compared to wild-type counterparts. Multiomic analyses— including chromatin accessibility, transcriptomics, and metabolomics—reveal decreased inflammatory signaling, a youthful metabolic profile, and improved proteostasis in Mda5 -/- HSCs, through regulation of HSF1 and phospho-EIF2A, key proteostasis regulators. Activation of HSF1 in aged wild-type HSCs partially restores youthful features, supporting a causal role for proteostasis maintenance. Collectively, our findings demonstrate that attenuating MDA5-dependent inflammation preserves HSC function during aging by maintaining metabolic fitness and proteostasis and provide insight into potential therapeutic strategies for mitigating hematopoietic aging.}, language = {en} } @article{BergnerOrthMeissneretal., author = {Bergner, Joseph and Orth, Nico and Meissner, Lucas and Quaschning, Volker}, title = {Solar Charging—Lessons Learned from Field Observation}, series = {World Electric Vehicle Journal}, volume = {17}, journal = {World Electric Vehicle Journal}, number = {2}, editor = {Sierpiński, Grzegorz}, publisher = {MDPI}, issn = {2032-6653}, doi = {10.3390/wevj17020069}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22325}, abstract = {Although the combination of solar power and electric vehicles is widely considered beneficial, practical applications reveal substantial variance. To determine the proportion of solar energy used for charging and to identify the main drivers of a high solar share, a dataset containing measured 5 min energy time series of 725 households with PV and EVs was analyzed. In the existing literature, this represents a novelty, as most studies in this field are simulation-based, rely on synthetic profiles, use lower time resolutions, or are based on questionnaires. The share of solar energy used for EV charging is highly dispersed and varies by about ±40\% around a median of 60\%. The analysis shows that clustering by preferred charging times has strong explanatory potential: at the median, EVs charged predominantly during the daytime achieve a solar share that is more than 40\% higher than those charged in the evening. In the latter case, home battery storage increases the solar share by an average of 20 percentage points. A similar magnitude of a 25-percentage-point increase could be reached with solar surplus charging compared to uncontrolled charging. On average, households with PV, battery, and EVs cover more than 56\% of their total demand with self-generated solar energy; with solar-adapted charging, median values exceed 77\%. If a heat pump is used on site, the self-sufficiency decreases but can still reach median values above 45\% and up to 61\% for optimized households.}, subject = {Consumer behavior}, language = {en} } @article{WinartoSutterTockhornetal., author = {Winarto, Hanifah and Sutter, Johannes and Tockhorn, Philipp and Škorjanc, Viktor and Patil, Pramila and Berwig, Sebastian and Zimmermann, Lea and Mart{\´i}nez-Denegri, Guillermo and Albrecht, Steve and Becker, Christiane}, title = {Periodic Inverted Micropyramids for Optically Optimized Fully Textured Solution-Processed Perovskite Solar Cells}, series = {Solar RRL}, volume = {9}, journal = {Solar RRL}, number = {23}, publisher = {Wiley}, issn = {2367-198X}, doi = {10.1002/solr.202500613}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22175}, pages = {10}, abstract = {Optical performance of perovskite-based solar cells can be enhanced by utilizing fully textured interfaces. However, solution processing of perovskite films on textured surfaces is a nonstraightforward and challenging process, particularly if optically most efficient micrometer-sized textures are used. In this work, we present fully textured solution-processed perovskite solar cells on periodic inverted micropyramids. The textures have a period of 4 μm with varying pyramid depths and are fabricated by wet-chemical etching of silicon with subsequent replication on glass substrates using nanoimprint lithography. Inverted pyramids are shown to enable low reflectance similar to random micropyramids on silicon. Additionally, they are able to confine perovskite precursor solution within its structure during spin coating, resulting in a conformal, fully textured perovskite film. We demonstrate that the resulting fully textured single-junction perovskite solar cells feature a reduced reflection loss of up to 1.2 mA/cm 2 in short-circuit current density. Moreover, we observe that the amount of lead iodide in the perovskite precursor solution crucially impacts growth and nonradiative recombination losses of the fully textured perovskite solar cells on inverted micropyramids. Finally, we prove the versatility of our approach by also demonstrating conformal coating with slot-die coating, which is a scalable process considered for industrial application.}, subject = {Solar cells}, language = {en} } @article{Farias‐BasultoKafedjiskaBertrametal., author = {Farias-Basulto, Guillermo and Kafedjiska, Ivona and Bertram, Tobias and Riedel, Maximilian and Emery, Quiterie and Remec, Marko and Graniero, Paolo and Khenkin, Mark and Kaufmann, Christian A. and Lauermann, Iver and Klenk, Reiner and Schlatmann, Rutger and Ulbrich, Carolin}, title = {Perovskite-CIGSe Tandem Solar Cell: Over One Year of Outdoor Monitoring}, series = {Advanced Energy and Sustainability Research}, volume = {6}, journal = {Advanced Energy and Sustainability Research}, number = {12}, issn = {2699-9412}, doi = {10.1002/aesr.202500162}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22159}, pages = {11}, abstract = {Tandem solar cells can surpass the limitations of single-junction devices, promising increased performance due to lower thermalization losses. Even though many research and industrial upscaling efforts are based on perovskite-Si tandems, all-thin-film photovoltaic (PV) devices, for instance with chalcopyrite (CIGSe) and perovskite, can offer many advantages such as significant cost and material savings and access to niche markets like building integrated- and flexible PV. However, long-term stability and outdoor performance of perovskite-based tandem devices is to this day challenging. This work presents the first data analysis of year-round outdoor measurements (mpp-tracked) of a perovskite-chalcopyrite tandem device with a starting efficiency of about 23.14\% before encapsulation. The maximum outdoor performance of the tandem device changed during the period of observation, reaching the peak performance in April and then decreased due to the device degradation. At its maximum outdoor performance, the tandem could reach up to 68\% higher instantaneous power output, relative to its single-junction reference (CIGSe-SJ). In addition, a quantitative time series performance analysis, exemplary qualitative imaging characterization of the tandem before and after outdoor exposure, is shown. Finally, the possibility of predicting the immediate performance of an all-thin-film tandem is verified by using a multiple linear regression model with accuracies generally exceeding 90\%.}, subject = {Machine learning}, language = {en} } @article{KuettenbaumBramlHeinzeetal., author = {K{\"u}ttenbaum, Stefan and Braml, Thomas and Heinze, Marco and Kainz, Christian and Stettner, Christian and Taffe, Alexander}, title = {Reliability and partial factor-based assessment of a highway bridge supported by nondestructive testing}, series = {Structural Concrete}, volume = {26}, journal = {Structural Concrete}, number = {5}, publisher = {Wiley}, address = {Weinheim}, issn = {1464-4177}, doi = {10.1002/suco.202400717}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22149}, pages = {5535 -- 5554}, abstract = {The preservation of the degrading transport infrastructure is vital, as replacing it is impossible in view of limited budgets and environmental impacts. One component in overcoming this challenge is the evaluation of the reliability and the calculation of the remaining service life of existing structures. These structural reliability reassessments allow for the identification and utilization of load-bearing capacity reserves. The consideration of actual structural characteristics and environmental conditions, various of which can be gathered by means of numerous measurement and inspection techniques, has the potential for more realistic computation results and therefore more economic decisions about the operation as well as maintenance activities. This article attempts to shed light on the potential of nondestructive testing (NDT) methods for updating the input variables in the ultimate limit states during a recalculation. The approach of the NDT-supported reliability reassessment is demonstrated using a prestressed concrete bridge emphasizing the bending proof in transversal direction. As a result, the semi-probabilistic, probabilistic and NDT-based reassessment results are compared and the effects of NDT-supported structural analysis on the calculated reliability are highlighted.}, language = {en} } @misc{CortesHidalgo, type = {Master Thesis}, author = {Cortes Hidalgo, Carlos Bernardo}, title = {From Global to Local: Tailoring International Green Hydrogen Supply Models for Sustainable Mobility}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {31}, abstract = {Hydrogen is one of the most promising decarbonization technologies for the transportation sector since it only emits water vapor when produced with electrolysis. However, different aspects present a challenge for its implementation and there are currently limitations regarding hydrogen as an economy of scale. This paper analyzes the key factors for successful implementation of green hydrogen supply chains with road transportation as an end user application. In order to do so, a comprehensive literature review along with semi-structured stakeholders interviews were performed to gather real-life data of the hydrogen economy. Most of the bottlenecks for the implementation of green hydrogen are in CAPEX/OPEX costs or infrastructure, something that can be addressed by effective policy-making and a resilient supply chain design. Equally important, certain elements regarding innovation, partnerships, tailoring powertrain needs and diversification of business modes were found to be fundamental in success.}, language = {en} } @misc{Afridi, type = {Master Thesis}, author = {Afridi, Muhammad Abdullah}, title = {Navigating Funding Landscapes: Lessons from the EU for Advancing Sustainable Urban Utilities in Developing Nations}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {44}, abstract = {This thesis, titled "Navigating Funding Landscapes: Lessons from the EU for Advancing Sustainable Urban Utilities in Developing Nations," delves into the complex challenges and strategic solutions associated with municipal utilities accessing European Union (EU) funds. Through an in-depth examination of the Multiannual Financial Framework (MFF) and the European Green Deal (EGD), this study illuminates the bureaucratic, regulatory, financial, political, and operational obstacles that municipal utilities encounter. Utilizing qualitative methodologies, including expert interviews and thematic analysis, the research uncovers essential success factors and provides actionable recommendations for both the EU and municipal utilities to optimize the advantages of these funds. The findings underscore the importance of nuanced regulatory adjustments, streamlined processes, robust political support, and strategic partnerships. Additionally, the thesis investigates the transferability of EU successes to developing nations, emphasizing both the achievements and limitations of the EU model. This research highlights the urgent need for policy reforms and international cooperation to empower municipal utilities, thereby promoting resilient and sustainable urban development on a global scale.}, language = {en} } @misc{Arndt, type = {Master Thesis}, author = {Arndt, Philipp}, title = {Bayesian Word Embeddings}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {47}, abstract = {Word embeddings are an area of natural language processing that has received a lot of attention in recent years as models have become more accurate. An embedding is a representation of a word, usually in a multi-dimensional vector space, that attempts to capture the semantic and syntactic properties and thus its meaning. Machine learning models that learn this type of representation typically assume that similar words occur in similar contexts and try to place the words in the vector space so that similar words are close together and dissimilar words are far apart. However, traditional embedding methods only assign a single vector to each word and therefore fail to correctly represent words with multiple meanings. Several approaches have been proposed to overcome this weakness, with Bayesian embeddings being a particularly promising one. In this thesis, a Bayesian word embedding model based on the variational autoencoder framework is developed, trained, evaluated and compared with state-of-the-art solutions. The results suggest that the chosen approach is suitable for learning word embeddings that capture the semantic and syntactic properties of words based on the contexts in which they occur, as the model is able to correctly classify similar words. While the performance on some standard evaluation benchmarks is rather poor, it can be speculated that this is most likely due to the small dataset on which the model was trained. It's also not clear whether the embeddings encode a sense of uncertainty or not, since the implemented evaluation methods don't allow any statements in this regard.}, language = {en} } @article{BehrendtEppinger, author = {Behrendt, Tanita and Eppinger, Elisabeth}, title = {Circularity for Sustainable Textiles: Aligning Fiber Compositions of T-Shirts with Ecodesign and Recyclability}, series = {Sustainability}, volume = {17}, journal = {Sustainability}, number = {22}, editor = {Gai, Chao}, publisher = {MDPI}, issn = {2071-1050}, doi = {10.3390/su172210057}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21828}, abstract = {The sustainability transition of the textile industry requires amongst other strategies circular approaches. Ecodesign guidelines and design for recycling are approaches that reduce resource consumption and textile waste. Garments are made of a large variety of different materials, from blended fibers to haberdashery items, colorants, and finishings, making it challenging to predict the composition of post-consumer textile waste. This mix of materials complicates recycling efforts, contributing to globally less than 1\% of fiber-to-fiber recycling. This study investigates material compositions of one of the most popular and widespread garments: T-shirts. While about half of our sample contains cotton only, the other items contain two or more fibers, revealing huge variations in fiber blends, including varying degrees of elastane contents, which are not linked to functional requirements. These blends, especially the varying levels of elastane, increase costs and efforts for recycling, making fiber-to-fiber recycling less attractive and more expensive than new fiber production. They also contribute to avoidable microfiber pollution. Accordingly, this study underlines the requirements for providing detailed ecodesign guidelines and applying the extended producer responsibility to incorporate environmental lifecycle costs, to help shift the industry towards a circular economy.}, subject = {Fibers}, language = {en} } @article{UschmajewZeiser, author = {Uschmajew, Andr{\´e} and Zeiser, Andreas}, title = {Discontinuous Galerkin discretization of conservative dynamical low-rank approximation schemes for the Vlasov-Poisson equation}, series = {BIT Numerical Mathematics}, volume = {65}, journal = {BIT Numerical Mathematics}, number = {4}, publisher = {Springer}, address = {Dordrecht}, issn = {0006-3835}, doi = {10.1007/s10543-025-01085-6}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21983}, abstract = {A numerical dynamical low-rank approximation (DLRA) scheme for the solution of the Vlasov-Poisson equation is presented. Based on the formulation of the DLRA equations as Friedrichs' systems in a continuous setting, it combines recently proposed conservative DLRA methods with a discontinuous Galerkin discretization. The resulting scheme is shown to ensure mass and momentum conservation at the discrete level. In addition, a new formulation of the conservative integrator is proposed based on its interpretation as a tangent space projector splitting scheme. Numerical experiments validate our approach in one- and two-dimensional simulations of Landau damping. As a demonstration of feasibility, it is also shown that the rank-adaptive unconventional integrator can be combined with mesh adaptivity.}, language = {en} } @article{ParkPaientko, author = {Park, Bomee and Paientko, Tetiana}, title = {Assessing the Impact of Capital Expenditure on Corporate Profitability in South Korea's Electronics Industry: A Regression Analysis Approach}, series = {Analytics}, volume = {4}, journal = {Analytics}, number = {4}, editor = {Leung, Carson K.}, publisher = {MDPI}, issn = {2813-2203}, doi = {10.3390/analytics4040036}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22062}, abstract = {This study investigates the relationship between capital expenditure (CAPEX) and long-term corporate profitability in South Korea's electronics industry. Using panel data from 126 listed electronics firms covering 2005-2019, the research applies fixed-effects regression analysis to examine how CAPEX influences profitability, measured by EBITDA/total assets. The results confirm that CAPEX exerts a positive and statistically significant long-term effect on profitability, with stronger but not significantly different impacts for large firms compared to SMEs. The findings contribute to empirical evidence on capital investment efficiency and the implications of economies and diseconomies of scale in capital-intensive industries.}, language = {en} } @article{HiebertHofstetterThomasetal., author = {Hiebert, Tina and Hofstetter, Florian and Thomas, Carsten and Mushtaq, Savera and Kaan, Eero and Parameswaran, Biranavan}, title = {Labels4Rails: A Railway Image Annotation Tool and Associated Reference Dataset}, series = {Data}, volume = {10}, journal = {Data}, number = {12}, editor = {Torres-Sospedra, Joaqu{\´i}n}, publisher = {MDPI}, issn = {2306-5729}, doi = {10.3390/data10120210}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22070}, abstract = {The development of autonomous train systems relies heavily on machine learning (ML) models, which in turn depend on large, high-quality annotated datasets for training and evaluation. The railway domain lacks adequate public datasets and efficient annotation tools. To address this gap, we present Labels4Rails, a tool designed specifically for the annotation of railway scenes. It captures track topology, switch states including switch directions, and informational tags regarding the images' content and leverages consistent camera perspectives and the fixed track geometries inherent to railways for annotation efficiency. We used Labels4Rails to create the L4R_NLB reference dataset from Norwegian railway footage. The dataset contains 10,253 annotated images across four seasons, including 1415 switch annotations. Both the tool and dataset are publicly available.}, subject = {Artificial intelligence}, language = {en} } @masterthesis{Adhikari, type = {Bachelor Thesis}, author = {Adhikari, Kapil}, title = {Implementation_of_Multichannel_Systems_on_a_Zynq-Based_System-on-Module_for_Terahertz_Spectroscopy}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {27}, abstract = {The Terahertz Sensors and Systems group at Fraunhofer HHI has developed a single channel continuous wave terahertz spectrometer. This spectrometer utilizes photoconductive antennas to convert an optical beat signal, generated by two lasers operating at different wavelengths, into the THz frequency range. Its potential for real-world applications is demonstrated through thickness measurements of multilayer dielectric samples. Scaling the system to a multichannel configuration would enable simultaneous measurements at multiple spatial points, enhancing imaging resolution and expanding it to industrial applications, such as radar-based object tracking and detection in autonomous vehicles and drone navigation. This thesis investigates two innovative approaches for scaling the single channel terahertz spectrometer to a multichannel configuration using the RedPitaya STEMLab 125-14. The RedPitaya board consists of a Zynq7020 System on Chip (SoC), a dual channel Analog-Digital-Converter (ADC) and a dual channel Digital-Analog-Convert (DAC). The first approach explores channel expansion through an external 8 channel ADC (ADC20Click), utilizing the Linux kernel's SPI driver for communication. The second approach implements a synchronized cluster of multiple RedPitaya devices in a master-slave architecture, leveraging their integrated ADCs and distributed clock synchronization. The SPI communication between the RedPitaya and the ADC20Click, operating on a Linux-based system, faced challenges in achieving synchronized data acquisition due to the non-deterministic nature of OS scheduling. Consequently, the sampling rate varied between 5 ksps to 8 ksps. Given the current system architecture, this variability makes the approach unsuitable for integration. However, for other THz applications that can tolerate lower sample rates and non-synchronized operation, this method could still be a viable option. In contrast, the approach utilizing multiple RedPitaya boards in a cluster successfully achieved synchronized data acquisition across multiple ADC channels. This was accomplished by generating and distributing clock and synchronization signals from the master device to the connected slave devices. A constant delay between sample acquisitions across devices was observed, which could be compensated using a delay unit with configurable delay lengths. The system maintained a unified and scalable hardware architecture. Integrating the new approach into the existing THz spectrometer, it demonstrated synchronized measurements across three channels, utilizing one ADC channel from each RedPitaya board in the cluster. Compared to the current system, the RedPitaya cluster managed to achieve a similar dynamic range of 60 dB for a single shot measurement, enabling the development of a multi-channel CW spectrometer using the proposed approach.}, language = {en} } @misc{Busch, type = {Master Thesis}, author = {Busch, Josefine}, title = {Alternative Byte-Pair Encoding for DNA-Specific Classification Tasks with Neural Networks}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {54}, abstract = {Traditional tokenization algorithms used for DNA classification tasks are optimized for Natural Language Processing (NLP) tasks and ignore key-characteristics of DNA sequence data. A specialized tokenizer for DNA data could utilize those characteristics to perform better on DNA classification tasks. This master's thesis describes the implementation and evaluation of HelixSwap a novel, tokenization algorithm based on Byte Pair Encoding (BPE). In two evaluationscenarios, machine learning (ML) models trained using HelixSwap are compared to models using standard BPE for tokenization. HelixSwap leads to an improved pretraining performance in all investigated use cases. The models using the novel algorithm achieve comparable results during fine-tuning to the standard BPE tokenizer models.}, language = {en} } @masterthesis{Szalontay, type = {Bachelor Thesis}, author = {Szalontay, No{\´e}mi}, title = {Museum of the Invisible - How museums can meaningfully engage with and support people experiencing homelessness}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {47}, abstract = {This thesis investigates the potential for museums to engage meaningfully with people experiencing homelessness through a spectrum of practical and strategic measures. Drawing on case studies, theoretical frameworks, and participatory methodologies, it examines how institutions can move beyond representational approaches to foster co-creation, empowerment, and long-term social impact. The research identifies actions ranging from simple, immediately implementable measures to complex, resource intensive commitments, applicable to museums of any size or type. Particular attention is given to trans-disciplinary collaboration, the recognition of lived experience as expertise, and the integration of social justice into institutional structures. The paper argues that museums have both the capacity and the responsibility to act as civic actors in addressing homelessness.}, language = {en} } @masterthesis{Chammam, type = {Bachelor Thesis}, author = {Chammam, Youssef}, title = {A Practical Exploration of discv5 in Distributed Applications}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {70}, abstract = {This thesis explores the discv5 discovery protocol, analyzing its role and implementation in decentralized network systems. Beginning with an overview of distributed networks and their architectural overlays, the study delves into the operational principles of the Kademlia Distributed Hash Table (DHT) as a foundational technology for discv5. The evolution from discv4 to discv5 is scrutinized, focusing on the enhancements and their implications for decentralized networking. The practical component of the thesis presents the integration of the discv5 Rust library into a prototype decentralized network. This network utilizes discv5 for node discovery and dynamic routing table management, coupled with a DHT for decentralized data storage, presenting a model for key-value storage in distributed environments. The implementation details are discussed, showcasing how the discv5 library can be adapted for real-world applications. Extensive testing of the library under various scenarios—including IP changes, ENR updates, and concurrent requests—provides insights into its resilience and e↵ectiveness against security threats, such as eclipse attacks. These tests highlight the library's capabilities and limitations within a simulated decentralized social media platform, demonstrating its potential to enhance privacy and reduce reliance on central authorities. The findings underscore the viability of discv5 as a critical component for decentralized applications, emphasizing its security features, scalability, and adaptability. This thesis not only extends the theoretical understanding of distributed discovery protocols but also contributes practical insights into their application in enhancing decentralized digital ecosystems.}, language = {en} } @masterthesis{Zuendorf, type = {Bachelor Thesis}, author = {Z{\"u}ndorf, Lennard Peter}, title = {Building an Interpretable Natural Language AI Tool based on Transformer Models and Approaches of Explainable AI}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {48}, abstract = {The rapid advancement of Artificial Intelligence, particularly in the context of transformer-based Large Language Models, has ushered in a new era in consumer and business software. However, a significant aspect often overlooked is the integration of Explainable AI, specifically interpretable machine learning approaches. This thesis explores the relevance and applicability of existing XAI methods to contemporary transformer based models, widely used in applications such as ChatGPT. Through the development of an explainable application prototype, this study assesses the feasibility of integrating XAI techniques into practical, industry level applications. To this end, open-source technologies were employed to create a usable chatbot application prototype that provides decision-making explanations. Despite the successful development of a functional prototype, challenges remain in achieving userfriendly interpretability, particularly with complex models. Thus, this study demonstrates the feasibility of integrating XAI into chat applications but also highlights the vital areas for future research in enhancing post hoc interpretability methods. The prototype reveals limitations in current post hoc interpretability methods and underscores the need for improved approaches, especially considering the increasing size and complexity of LLMs.}, language = {en} } @article{MpofuMwasiagiMechaetal., author = {Mpofu, Nonsikelelo Sheron and Mwasiagi, Josphat Igadwa and Mecha, Cleophas Achisa and Nganyi, Eric Oyondi}, title = {Characterization and Performance Evaluation of Cotton Fabrics Functionalized via In Situ Green Synthesis of Silver Nanoparticles Using Solanum tuberosum Peel Extract}, series = {Polymers}, volume = {17}, journal = {Polymers}, number = {19}, editor = {Antunes, Marcelo}, publisher = {MDPI}, issn = {2073-4360}, doi = {10.3390/polym17192598}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21704}, abstract = {The functionalization of textiles with nanomaterials through green synthesis offers a promising pathway for sustainable material innovation. This study explores the in situ green synthesis of silver nanoparticles (AgNPs) onto cotton fabrics using Solanum tuberosum (potato) peel extract as a natural reducing and stabilizing agent. The synthesis conditions were optimized by varying silver nitrate concentration, extract volume, temperature, pH, and reaction time, after which the optimized protocol was applied for fabric treatment. The presence and distribution of AgNPs were confirmed through UV-Visible spectroscopy, Fourier-transform infrared spectroscopy, scanning electron microscopy and dynamic light scattering. The treated fabrics demonstrated strong and durable antibacterial performance, with inhibition zones of 23 ± 0.02 against Escherichia coli and 16 ± 0.01 against Staphylococcus aureus. Notably, antibacterial activity was retained even after 20 washing cycles, demonstrating the durability of the treatment. Mechanical testing revealed a 32.25\% increase in tensile strength and a corresponding 10.47\% reduction in elongation at break compared to untreated fabrics, suggesting improved durability with moderate stiffness. Air permeability decreased by 8.8\%, correlating with the rougher surface morphology observed in Scanning Electron Microscopy images. Thermal analysis showed a decrease in thermal stability relative to untreated cotton, highlighting the influence of AgNPs on degradation behavior. Overall, this work demonstrates that potato peel waste, an abundant and underutilized biomass, can be used as a sustainable source for the green synthesis of AgNP-functionalized textiles. The approach provides a cost-effective and environmentally friendly strategy for developing multifunctional fabrics, while supporting circular economy goals in textile engineering.}, language = {en} } @article{FariasBasultoMehlhopOttoetal., author = {Farias-Basulto, Guillermo and Mehlhop, Thede and Otto, Nicolas J. and Bertram, Tobias and Jäger, Klaus and Gall, Stefan and Weinberger, Nikolaus and Schlatmann, Rutger and Lauermann, Iver and Klenk, Reiner and List-Kratochvil, Emil and Kaufmann, Christian A.}, title = {Improving Perovskite/CIGS Tandem Solar Cells for Higher Power Conversion Efficiency through Light Management and Bandgap Engineering}, series = {ACS Applied Materials \& Interfaces}, volume = {17}, journal = {ACS Applied Materials \& Interfaces}, number = {40}, publisher = {American Chemical Society}, issn = {1944-8244}, doi = {10.1021/acsami.5c15458}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21725}, pages = {56250 -- 56255}, abstract = {Perovskite and chalcopyrite materials are excellent absorbers for highly efficient, all-thin-film tandem solar cells. This work presents a certified world record for such a device, achieving a power conversion efficiency of 24.6 ± 1.1\% under steady-state conditions. The best IV parameters extracted from certified current-voltage measurements presented a short-circuit current density of around 19.3 mA/cm 2 , an open-circuit voltage of 1.765 V, and a fill factor of 71.8\%. In comparison to our previous record, the current density improved considerably, mainly due to the lowering of the bandgap of the bottom subcell and the improved optics of the top perovskite cell.}, subject = {Photovoltaic}, language = {en} } @article{HohlBoenischMankeetal., author = {Hohl, Tobias and B{\"o}nisch, Ulrike and Manke, Thomas and Arrigoni, Laura}, title = {Enhancing single-cell ATAC sequencing with formaldehyde fixation, cryopreservation, and multiplexing for flexible analysis}, series = {BMC Research Notes}, volume = {18}, journal = {BMC Research Notes}, number = {1}, publisher = {Springer Nature}, address = {London}, issn = {1756-0500}, doi = {10.1186/s13104-025-07547-y}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21739}, abstract = {Objective: The need for freshly isolated cells in bulk or single cell ATAC-seq experiments creates considerable logistical barriers and increases susceptibility to batch effects. This makes it difficult to coordinate complex or longitudinal studies. Our goal was to develop a sample preservation strategy that overcomes these limitations, enabling consistent and high-quality chromatin accessibility profiling from archived samples. Results: We established a workflow that incorporates mild formaldehyde fixation prior to cryopreservation, preserving both bulk and single-cell ATAC-seq data quality at levels comparable to fresh samples in HepG2 cells. This protocol reliably maintains key data quality metrics, including signal-to-noise ratio and fragment distributions. Furthermore, the method is fully compatible with transposase-based sample multiplexing using custom Tn5 barcodes. To address barcode hopping inherent to multiplexing, we introduced a computational demultiplexing strategy based on fragment ratios, which accurately assigns single cells to their sample of origin. Our approach streamlines experimental logistics and ensures reproducibility across diverse and temporally dispersed samples, broadening the scope for ATAC-seq-based studies, including those in clinical research settings where coordinated sample collection is challenging.}, subject = {Single Cell Analysis}, language = {en} } @article{UschmajewZeiser, author = {Uschmajew, Andr{\´e} and Zeiser, Andreas}, title = {Discontinuous Galerkin discretization of conservative dynamical low-rank approximation schemes for the Vlasov-Poisson equation}, series = {BIT Numerical Mathematics}, volume = {65}, journal = {BIT Numerical Mathematics}, number = {4}, publisher = {Springer Nature}, address = {Dordrecht}, issn = {0006-3835}, doi = {10.1007/s10543-025-01085-6}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21753}, abstract = {A numerical dynamical low-rank approximation (DLRA) scheme for the solution of the Vlasov-Poisson equation is presented. Based on the formulation of the DLRA equations as Friedrichs' systems in a continuous setting, it combines recently proposed conservative DLRA methods with a discontinuous Galerkin discretization. The resulting scheme is shown to ensure mass and momentum conservation at the discrete level. In addition, a new formulation of the conservative integrator is proposed based on its interpretation as a tangent space projector splitting scheme. Numerical experiments validate our approach in one- and two-dimensional simulations of Landau damping. As a demonstration of feasibility, it is also shown that the rank-adaptive unconventional integrator can be combined with mesh adaptivity.}, subject = {-}, language = {en} } @inproceedings{HenschelGetzinRoffiaetal., author = {Henschel, Thomas and Getzin, Fares and Roffia, Paolo and Fonseca, Mark and Otter, Nadine and Kuttner, Michael and Bution, Jefferson Luiz and Oliva, F{\´a}bio Lotti and Fenner, Germano and Snippe, Linda and Jansen, Pieter and Saveliev, Mykyta and Florio, Cristina and Nastari, Raffaela}, title = {11th ERRN Conference - Risk Management Challenges in Times of Change}, series = {11th ERRN Conference}, booktitle = {11th ERRN Conference}, editor = {Henschel, Thomas and Getzin, Fares}, doi = {10.34702/opus4-2177}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21770}, pages = {1 -- 53}, abstract = {It was with great pleasure that the 11th European Risk Research Network Conference was hosted at HTW Berlin under the theme "Risk Management Challenges in Times of Change." The conference brought together leading academics, practitioners, and emerging scholars from across Europe, as well as North and South America, to engage in a shared exploration of the evolving landscape of risk management. The international character of the event reflected the global nature of today's risk environment—an interconnected world where new threats and opportunities continue to emerge at an accelerating pace. The conference aimed to provide a comprehensive examination of contemporary risk management challenges and to highlight innovative approaches to addressing them. In doing so, it sought to bridge the gap between academic research and professional practice. By fostering meaningful dialogue between scholars and industry experts, the event served as a platform where theoretical insights and practical experiences were exchanged, contributing to a deeper understanding of complex risk phenomena. The contributions showcased during the conference—and now documented in these proceedings—reflect this interdisciplinary and practice-oriented approach. They examine risk management challenges from various angles, providing both analytical insight and practical applicability. Together, they highlight key research gaps and suggest potential directions for future development in the field. As a result, this volume serves as a valuable resource for academics, professionals, and policymakers aiming to navigate the increasing complexity of modern risk management. Particular recognition is due to the PhD students who presented their projects. Their fresh perspectives and innovative ideas provided valuable impulses to the discussions and underscored the importance of supporting the next generation of risk management scholars. Their contributions have further enriched the diversity and intellectual vibrancy of the conference. By bringing together a distinguished group of experts from academia and practice, the conference and its proceedings have provided an in-depth examination of the challenges associated with managing risk in a rapidly changing world. The methods, insights, and case studies presented here are intended to inspire further research, reflection, and collaboration among all those committed to advancing the discipline and practice of risk management.}, subject = {Risk Management}, language = {en} } @misc{Heise, type = {Master Thesis}, author = {Heise, Roxana Rebecca Elin}, title = {Molecular mechanisms of biocide-induced antibiotic tolerance in bacteria}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {39}, abstract = {Antimicrobial resistance represents a significant and imminent threat to global health, according to the World Health Organization (WHO). The emergence of resistance in bacteria and other microorganisms is a consequence of the sustained and excessive utilisation of antimicrobial agents. Considering these developments, concerns are rising regarding the potential for cross-resistance between biocides and antibiotics. There are mechanisms, that affect the susceptibility of microorganisms to both classes of antimicrobial agents. Three modes of reduced susceptibility are distinguished between resistance, tolerance and persistence. Resistance describes a decrease of susceptibility and the ability to grow in presence of concentrations, that inhibit closely related strains. Tolerance is described as a loss of susceptibility and an increase in survival of the entire population. Persistence refers to a subpopulation that is able survive due to changes in gene expression, often without genetic alterations. In previous experiments conducted at BAM, exposure to a subinhibitory concentration of glutaraldehyde was observed to result in enhanced survival against antibiotics and an induction of prolonged lag. A prolongation of the lag time is associated with antibiotic tolerance and increased survival. Using transcriptomics two genes of interest were discovered: yadV, a putative fimbrial chaperone and yjfL, an inner membrane protein. The genes and their role in lag and antibiotic survival are not well studied. The aim of this thesis is to investigate, the role of the genes in lag and antibiotic tolerance. To conduct the experiments, a few mutants were obtained.: I) a knockout mutant, II) an overexpression mutant and III) a fluorescent reporter mutant of yadV and yjfL, together with respective controls. The overexpression and reporter mutant were transformed into the wild type E. coli MG1655 used in the previous experiments. The minimal inhibitory concentration (MIC) and the minimal bactericidal concentration (MBC) were determined for the knockout mutants and the wildtypes. The knockout of the genes of interest does not show a significant difference compared to the wild type. Next, time-kill-curves for the knockout and overexpression mutants were gathered in the presence of lethal doses of antibiotics. The knockout mutants were induced with subinhibitory concentrations of glutaraldehyde to assess the effect of the mutations on antibiotic tolerance. The knockout mutants do not indicate a difference to the wildtypes again, whereas the overexpression of yadV was observed to induce tolerance. Scanlag experiments were performed to examine the distribution of single-cell lag-times and detect changes, as an extended lag-time promotes survival. The prolonged lag was detected for knockout mutants and the wildtypes. This suggests the tolerance induced by glutaraldehyde is independent of yadV and yjfL. Overexpression mutants exhibit the same lag as the wildtype. While the overexpression of yadV facilitates tolerance, it is not tolerance by lag. A different approach was employed for the reporter mutants. The cells were compared with cells exposed to a sublethal concentrations of glutaraldehyde, to detect heterogeneity. A heterogenous expression may indicate a tolerant subpopulation, that functions as persisters. Such a subpopulation was identified for yadV as well as the positive control. The knockout of the genes does not impact the survival against antibiotics. However, the overexpression of yadV impacts and promotes the survival.}, language = {en} } @inproceedings{KalffSimbeck, author = {Kalff, Yannick and Simbeck, Katharina}, title = {Explained, yet misunderstood: How AI Literacy shapes HR Managers' interpretation of User Interfaces in Recruiting Recommender Systems}, series = {Proceedings of the 5th Workshop on Recommender Systems for Human Resources (RecSys in HR 2025)}, booktitle = {Proceedings of the 5th Workshop on Recommender Systems for Human Resources (RecSys in HR 2025)}, editor = {Kaya, Mesut and Bogers, Toine and Bied, Guillaume and Johnson, Chris and Decorte, Jens-Joris}, publisher = {CEUR}, issn = {1613-0073}, url = {http://nbn-resolving.de/urn:nbn:de:0074-4046-X}, pages = {1 -- 10}, abstract = {AI-based recommender systems increasingly influence recruitment decisions. Thus, transparency and responsible adoption in Human Resource Management (HRM) are critical. This study examines how HR managers' AI literacy influences their subjective perception and objective understanding of explainable AI (XAI) elements in recruiting recommender dashboards. In an online experiment, 410 German-based HR managers compared baseline dashboards to versions enriched with three XAI styles: important features, counterfactuals, and model criteria. Our results show that the dashboards used in practice do not explain AI results and even keep AI elements opaque. However, while adding XAI features improves subjective perceptions of helpfulness and trust among users with moderate or high AI literacy, it does not increase their objective understanding. It may even reduce accurate understanding, especially with complex explanations. Only overlays of important features significantly aided the interpretations of high-literacy users. Our findings highlight that the benefits of XAI in recruitment depend on users' AI literacy, emphasizing the need for tailored explanation strategies and targeted literacy training in HRM to ensure fair, transparent, and effective adoption of AI.}, subject = {Explainable AI}, language = {en} } @article{SeilerLoidoltKrügervonKlitzingetal., author = {Seiler, Maximilian and Loidolt-Kr{\"u}ger, Maria and von Klitzing, Regine and Drews, Anja}, title = {Effect of pH and Particle Charge on the Interfacial Properties of Biocatalytic Pickering EmulsionsWhere Are the Enzymes Located?}, series = {Langmuir}, volume = {41}, journal = {Langmuir}, number = {37}, publisher = {American Chemical Society}, issn = {0743-7463}, doi = {10.1021/acs.langmuir.5c02459}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21539}, pages = {25264 -- 25276}, abstract = {Pickering emulsions (PEs), where water-in-oil (w/o) droplets are stabilized by nanoparticles (NPs), offer a promising platform for biocatalysis by providing a large interfacial area crucial for efficient substrate conversion. While several lipase catalyzed reactions in PEs have been demonstrated, the exact interfacial structure is unknown. This study focuses on the interfacial network formed by NPs and Candida rugosa lipase (CRL) at the octanol/water-interface by varying pH and NP charge. By applying different methods, the location of lipases within a PE was identified and the enzyme concentration profile quantified for the first time. Positively charged nanoparticles (NP+) adsorbed at the o/w-interface together with CRL to form a network-structure. The relation between individual and simultaneous adsorption showed a constant value of 0.75 for the investigated pH range. Negatively charged particles (NP-) did not adsorb spontaneously at the negatively charged octanol/water-interface and therefore showed no influence on the enzyme adsorption behavior. Interfacial shear rheology measurements further revealed distinct elastic behavior of the enzyme-particle network due to attractive interactions between positively charged nanoparticles and CRL. This was shown by a 4.4-fold increase in the interfacial storage modulus. In contrast, repulsive interactionseither between CRL and positively charged particles at low pH or with negatively charged particlesdid not enhance the elastic response of CRL at the interface. Confocal laser scanning microscopy of prepared PE droplets showed an interfacial CRL layer thickness of 0.75 μm for NP+ and 0.51 μm for NP-. Using NP+ results in a 30\% higher interfacial enzyme concentration, indicating a more compact layer structure. These insights contribute to optimizing biocatalytic systems using PEs for industrial applications and provide a basis for the quantitative analysis of the interfacial layer in a Pickering emulsion.}, language = {en} } @article{RufoloPaientkoDziergwa, author = {Rufolo, Alessia and Paientko, Tetiana and Dziergwa, Katrin}, title = {M\&As and Corporate Financial Performance: An Empirical Study of DAX 40 Firms}, series = {FinTech}, volume = {4}, journal = {FinTech}, number = {3}, editor = {Liu, David (Yulong) and Shahzad, Jawad Syed}, publisher = {MDPI}, issn = {2674-1032}, doi = {10.3390/fintech4030043}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21404}, abstract = {This study examines the impact of mergers and acquisitions (M\&As) on the financial performance of firms listed in Germany's DAX 40 index. Although M\&As are a widely used strategic tool intended to create value through synergies and market expansion, existing research provides conflicting evidence about their effectiveness. Using an empirical approach, we analyze the financial data of acquiring companies before and post-M\&A transactions to evaluate changes in profitability, liquidity and solvency. Our findings suggest that financial performance does not universally improve following acquisitions. Instead, results vary significantly based on deal characteristics and internal management factors. These results suggest that, while M\&A can be a pathway to growth, success depends heavily on the quality of execution and organizational integration. This paper contributes to the ongoing debate about the effectiveness of M\&As and provides insights for corporate decision-makers, investors, and policy stakeholders.}, subject = {Mergers and Acquisitions}, language = {en} } @article{Sandberg, author = {Sandberg, Berit}, title = {Startup Culture as a Masculinity Contest: An Exploratory Study on Prevalence and Gender Dynamics}, series = {Social Sciences}, volume = {14}, journal = {Social Sciences}, number = {7}, editor = {Morando, Martina and Platania, Silvia Maria}, publisher = {MDPI}, issn = {2076-0760}, doi = {10.3390/socsci14070438}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21075}, abstract = {Startups are often praised for their innovative power and dynamic work environments, but are also criticized for workplace cultures that perpetuate traditional masculine norms of competitiveness, workaholism, emotional resilience, and strength. This exploratory study examines the prevalence of Masculinity Contest Culture (MCC) in German startups and related gender dynamics. The Masculinity Contest Culture scale, which assesses masculine norms related to emotional resilience, physical superiority, workaholism, and aggressive competitiveness, was employed to collect data from 101 participants representing various startups. The results indicate an overall low prevalence of MCC, with slightly higher scores for strong commitment to work. Individuals from disparate gender and hierarchical status groups exhibited comparable ratings of the intensity of toxic masculinity. While the presence of female founders and supervisors did not affect MCC scores, male dominance within the work environment had a small but significant amplifying effect. Significant differences were observed based on the prevailing leadership style, indicating that a shared leadership approach is associated with the creation of a more inclusive and less toxic work environment. The results challenge commonly held assumptions about tech startups. In addition, the study highlights the need for further research into the impact of leadership dynamics on startup culture.}, subject = {Unternehmensgr{\"u}ndung}, language = {en} } @misc{Schmidt, type = {Master Thesis}, author = {Schmidt, Tim}, title = {The Free Quality School Education Program in Sierra Leone: Achievements and Remaining Challenges}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {163}, abstract = {Sierra Leone, a country in the west of Sub-Saharan Africa, is among the poorest countries in the world, ranking 181 out of 195 on the Human Development Index in 2021. The education system was abandoned for decades and, like the country itself, suffered from various crises such as a devastating eleven-year civil war that ended in 2002 and the Ebola pandemic in 2014. What remained is a dys-functional education system in many respects. When Julius Maada Bio became president in the 2018 democratic elections, he promised an alternative path to development: he put education at the center. The Free Quality School Education Program was born - a program to reshape the fortunes of the country through the transformation of the education system. "The big bet on schools" started. This paper combines first-hand information collected in a four-week fieldtrip with available second-ary data to show that much has changed and that the current administration, along with its partners, has achieved remarkable improvements. Those improvements may sometimes seem small, but it is important to keep the baseline situation in mind. However, there are still major structural challenges that need to be addressed to ensure the sustainability.}, language = {en} } @masterthesis{Nguyen, type = {Bachelor Thesis}, author = {Nguyen, Minh Thu}, title = {Evaluation of the possibility of a ChatGPT plugin in the context of universities}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {46}, abstract = {The popularity of ChatGPT for over half a year has led to transformative changes across various domains. With the ongoing development of models aiming to achieve human-like intelligence, concerns have arisen about the potential impact of artificial intelligence. In this context, OpenAI\&\#039;s Plugin emerges as a third-party service, complementing ChatGPT by accessing and extracting data. Among the 800 verified Plugins, some have demonstrated notable efficiency in assisting users with tasks like mathematics learning and table reservations. The purpose of this thesis is to examine the ChatGPT and ChatGPT Plugin technology. Specifically, it will explore whether the Plugin lives up to its promises, how it functions when integrated into educational settings, and its overall impact. To address these considerations, a Prototype of the Plugin will be developed, and its effectiveness when integrated into an educational management domain will be tested. The research will provide insights into the outcomes of integrating the Plugin within a specific management domain and assess its capabilities and relevance within the university context.}, language = {en} } @masterthesis{Rahmani, type = {Bachelor Thesis}, author = {Rahmani, Parham}, title = {Analysis and Prototypical Implementation of a Database Solution for a Web Chat in a Distributed Application}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {54}, abstract = {The industry is rapidly going towards digitalization and using enterprise software, where scalable and performant database solutions are of important priority while transforming into the digital world. This document presents a comparative analysis, evaluation and prototypical implementation of two database solutions for a chat service in the Remsfal project. One prototype is a fully relational approach using MySQL, and another polyglot persistence approach integrates Apache Cassandra. The motivation behind this research stems from the differences between modern non-relational database technologies and relational database technologies regarding scalability, availability and performance. The study outlines the theoretical foundations, requirements analysis, design and implementation of the chat service and both database implementations, including their API structure, data schema, and integration with the backend. It also explores the incorporation of MinIO as an object storage solution for handling file attachments within the chat service. The advantages and disadvantages of both approaches will be analyzed through load testing and evaluation of the systems under stress, which provides insight into their viability as a database solution. In the end, the findings discuss an optimal prototype for the chat service in the Remsfal project based on the evaluations and the requirements stated in the analysis.}, language = {en} } @article{RieckhoffRiesebeckSolderaetal., author = {Rieckhoff, Simon and Riesebeck, Florian and Soldera, Marcos and Mayer-Stillrich, Katja and Wang, Qiong and Ruske, Florian and Lasagni, Andr{\´e}s Fabi{\´a}n and Becker, Christiane}, title = {Texture-Enhanced Mechanical Stability of Transparent Electrodes for Flexible Optoelectronics with Near-Infrared Response}, series = {Advanced Materials Interfaces}, volume = {12}, journal = {Advanced Materials Interfaces}, number = {11}, issn = {2196-7350}, doi = {10.1002/admi.202400922}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21043}, pages = {9}, abstract = {Transparent electrodes with high conductivity and mechanical robustness are essential for flexible opto-electronic applications. Indium tin oxide (ITO) single layers have long been considered as unsuitable for flexible applications due to their brittleness. Here, it is shown that their mechanical stability can be substantially enhanced by texturing the flexible substrate. First, the opto-electronic performance of single ITO layers and ITO/Ag/ITO stacks on polyethylene terephthalate (PET) foils is evaluated numerically by means of Haacke's figure-of-merit. Single ITO layers are found to be the electrode of choice for applications with a near-infrared response due to their superior transparency. Following this, the sheet resistance of ITO layers is experimentally investigated on textured PET upon deformation parallel and perpendicular to a 1D texture grating. An "accordion-like" deformation perpendicular to the grating and high texture aspect ratios are shown to avoid crack formation and loss of conductivity in the ITO. Simulations prove the considerably reduced occurrence of mechanical stress in this case. It is further experimentally demonstrated that texturing foils increase transmittance and haze. The enhanced mechanical robustness and optical performance by using textured foils make single ITO layers promising candidates for flexible opto-electronic applications with a near-infrared response, such as all-perovskite tandem solar cells, thermal sensors, and photodetectors.}, language = {en} } @article{LangeAhrensKrishnakumaretal., author = {Lange, Christoph and Ahrens, Andreas and Krishnakumar, Yadu Krishnan and Grote, Olaf}, title = {Length-Weight Distribution of Non-Zero Elements in Randomized Bit Sequences}, series = {Sensors}, volume = {25}, journal = {Sensors}, number = {12}, editor = {Kim, Hyoungshick}, publisher = {MDPI}, issn = {1424-8220}, doi = {10.3390/s25123825}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21022}, abstract = {Randomness plays an important role in data communication as well as in cybersecurity. In the simulation of communication systems, randomized bit sequences are often used to model a digital source information stream. Cryptographic outputs should look more random than deterministic in order to provide an attacker with as little information as possible. Therefore, the investigation of randomness, especially in cybersecurity, has attracted a lot of attention and research activities. Common tests regarding randomness are hypothesis-based and focus on analyzing the distribution and independence of zero and non-zero elements in a given random sequence. In this work, a novel approach grounded in a gap-based burst analysis is presented and analyzed. Such approaches have been successfully implemented, e.g., in data communication systems and data networks. The focus of the current work is on detecting deviations from the ideal gap-density function describing randomized bit sequences. For testing and verification purposes, the well-researched post-quantum cryptographic CRYSTALS suite, including its Kyber and Dilithium schemes, is utilized. The proposed technique allows for quickly verifying the level of randomness in given cryptographic outputs. The results for different sequence-generation techniques are presented, thus validating the approach. The results show that key-encapsulation and key-exchange algorithms, such as CRYSTALS-Kyber, achieve a lower level of randomness compared to digital signature algorithms, such as CRYSTALS-Dilithium.}, subject = {Burst}, language = {en} } @article{BagackiReinhardtSchlatmannetal., author = {Bagacki, Rory and Reinhardt, Maximilian and Schlatmann, Rutger and Calnan, Sonya and van de Krol, Roel and Browne, Michelle P.}, title = {Electrochemical hydrogen pumps: a researcher's guide and review}, series = {Chemical Communications}, volume = {61}, journal = {Chemical Communications}, number = {56}, publisher = {The Royal Society of Chemistry}, issn = {1359-7345}, doi = {10.1039/d5cc01815c}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21017}, pages = {10210 -- 10227}, abstract = {Hydrogen is considered an attractive energy vector and an indispensable base chemical for a wide variety of chemical products. As more hydrogen is produced via electrolysis, finding ways to store the H2 will become increasingly important due to the low volumetric energy density at ambient pressure. While high pressure storage is favoured for many applications, compressing hydrogen poses distinct challenges due to its low density and high diffusivity. Electrochemical hydrogen pumps (EHP) present a solution to this challenge by efficiently compressing hydrogen. Hydrogen compression is more efficient using electrochemical hydrogen pumps than conventional mechanical compressors because they operate through isothermal rather than adiabatic compression. Additionally, they can be used to separate hydrogen from gas mixtures, for example from natural gas pipelines supplemented with hydrogen, creating the possibility of integrating them with existing energy transport infrastructure. This paper summarizes recent progress in electrochemical hydrogen pump research and presents a case study on an EHP test cell, test rig and a measurement guide to advance research in this field. Although electrochemical hydrogen pumps offer many advantages, shortcomings remain, including the lack of standardized measurement conditions and procedures, as well as a limited understanding of degradation mechanisms. This review aims to provide insights into these issues and discuss future directions for electrochemical hydrogen pump research.}, language = {en} } @article{BjelobabaWaddingtonPerkinsetal., author = {Bjelobaba, Sonja and Waddington, Lorna and Perkins, Mike and Folt{\´y}nek, Tom{\´a}š and Bhattacharyya, Sabuj and Weber-Wulff, Debora}, title = {Maintaining research integrity in the age of GenAI: an analysis of ethical challenges and recommendations to researchers}, series = {International Journal for Educational Integrity}, volume = {21}, journal = {International Journal for Educational Integrity}, number = {1}, publisher = {Springer Nature Singapore}, address = {Singapore}, issn = {1833-2595}, doi = {10.1007/s40979-025-00191-w}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-21004}, abstract = {Background: This paper is a practice‑informed rapid review that maps the complex ethical challenges arising from the growing use of Generative Artificial Intelligence (GenAI) tools across the research life‑cycle. Earlier research primarily focused on academic integrity concerns related to students' use of GenAI tools; however, limited information is available on the impact of GenAI on academic research. This study aims to examine the ethical concerns arising from the use of GenAI across different phases of research and explores potential strategies to encourage its ethical use for research purposes. Methods: We selected one or more GenAI platforms applicable to various research phases (e.g. developing research questions, conducting literature reviews, processing data, and academic writing) and analysed them to identify potential ethical concerns relevant for that stage. Results: The analysis revealed several ethical concerns, including a lack of transparency, bias, censorship, fabrication (e.g. hallucinations and false data generation), copyright violations, and privacy issues. These findings underscore the need for cautious and mindful use of GenAI. Conclusions: The advancement and use of GenAI are continuously evolving, necessitating an ongoing in-depth evaluation. We propose a set of practical recommendations to support researchers in effectively integrating these tools while adhering to the fundamental principles of ethical research practices.}, subject = {Ethics}, language = {en} } @article{GemesiDoellingerWeinbergeretal., author = {Gemesi, Kathrin and D{\"o}llinger, Nina and Weinberger, Natascha-Alexandra and Wolf, Erik and Mal, David and Keppler, Sebastian and Wenninger, Stephan and Bader, Emily and Wienrich, Carolin and Luck-Sikorski, Claudia and Latoschik, Marc Erich and Israel, Johann Habakuk and Botsch, Mario and Holzapfel, Christina}, title = {Virtual body image exercises for people with obesity - results on eating behavior and body perception of the ViTraS pilot study}, series = {BMC Medical Informatics and Decision Making}, volume = {25}, journal = {BMC Medical Informatics and Decision Making}, number = {1}, publisher = {BioMed Central}, address = {London}, issn = {1472-6947}, doi = {10.1186/s12911-025-02993-x}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20989}, abstract = {Background: A negative body image can have an impact on developing and maintaining obesity. Using virtual reality (VR) to conduct cognitive behavioral therapy (CBT) is an innovative approach to treat people with obesity. This multicenter non-randomized pilot study examined the feasibility and the effect on eating behavior and body perception of a newly developed VR system to conduct body image exercises. Methods: Participants with a body mass index (BMI) ≥ 30.0 kg/m 2 without severe mental diseases attended three study visits in an interval of one to four weeks to receive virtual (VR intervention) or traditional (non-VR intervention) body image exercises. Data on anthropometrics, eating behavior (Dutch Eating Behavior Questionnaire, DEBQ), body perception (Body Shape Questionnaire, BSQ; Multidimensional Assessment of Interoceptive Awareness, MAIA), and satisfaction (standardized interview and questionnaire) were collected. Results: In total, 66 participants (VR intervention: 31, non-VR intervention: 35) were included. The majority was female (52/66, 78.8 \%), the mean age was 45.0 ± 12.8 years, and the mean BMI was 36.8 ± 4.3 kg/m 2 . Both intervention groups showed non-significant body weight reduction (VR intervention: 1.7 ± 3.3 \%, non-VR intervention: 0.9 ± 3.0 \%) and showed no statistically significant difference between the groups ( p  = 0.35). Scores of DEBQ, BSQ, and MAIA showed over time no statistically significant changes neither between the two groups nor within the groups (all p  ≥ 0.05). The overall satisfaction of the VR group with the two virtual body image exercises was high (4.1 ± 0.8 on a 5-point Likert scale). Conclusions: The intervention with the developed VR system was feasible and the virtual and traditional body image exercises resulted in statistically non-significant weight loss. It seems that single focus on body image is not successful in improving eating behavior and body perception in people with obesity. Long-term human intervention studies with larger sample sizes are necessary to examine the efficacy of integrating this kind of VR system into standard obesity therapy.}, subject = {Virtual reality}, language = {en} } @article{JassThomas, author = {Jaß, Philipp and Thomas, Carsten}, title = {Using N-Version Architectures for Railway Segmentation with Deep Neural Networks}, series = {Machine Learning and Knowledge Extraction}, volume = {7}, journal = {Machine Learning and Knowledge Extraction}, number = {2}, editor = {Javaheri, Danial}, publisher = {MDPI}, issn = {2504-4990}, doi = {10.3390/make7020049}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20867}, abstract = {Autonomous trains require reliable and accurate environmental perception to take over safety-critical tasks from the driver. This paper investigates the application of N-version architectures to rail track detection using Deep Neural Networks (DNNs) as a means to improve the safety of machine learning (ML)-enabled perception systems. We combine three different neural network architectures (WCID, VGG16-UNet, MobileNet-SegNet) in a 3M1I configuration. In this configuration, we apply two fusion methods to increase accuracy and to enable error detection: Maximum Confidence Voting (MCV), combining the DNN predictions at the image level, and Pixel Majority Voting (PMV), a novel approach for combining the predictions at the pixel level. In addition, we implement a new method for evaluating and combining prediction confidence values in the N-version architecture during runtime. We adjust the overall prediction confidence according to the conformity of all individual predictions, which is not possible with an individual network. Our results show that the N-version architecture not only enables a detection of erroneous predictions by utilizing those adjusted confidence values, but it can also partially improve the predictions by using the PMV combination algorithm. This work emphasizes the importance of model diversity and appropriate thresholds for an accurate assessment of prediction safety. These approaches can significantly improve the practical applicability of ML-based systems in safety-critical domains such as rail transportation.}, subject = {Machine learning}, language = {en} } @article{ZhangBelchevaErmakova, author = {Zhang, Weijie and Belcheva, Veronika and Ermakova, Tatiana}, title = {Interpretable Deep Learning for Diabetic Retinopathy: A Comparative Study of CNN, ViT, and Hybrid Architectures}, series = {Computers}, volume = {14}, journal = {Computers}, number = {5}, editor = {Chowdhury, Rafiqul}, publisher = {MDPI}, issn = {2073-431X}, doi = {10.3390/computers14050187}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20853}, abstract = {Diabetic retinopathy (DR) is a leading cause of vision impairment worldwide, requiring early detection for effective treatment. Deep learning models have been widely used for automated DR classification, with Convolutional Neural Networks (CNNs) being the most established approach. Recently, Vision Transformers (ViTs) have shown promise, but a direct comparison of their performance and interpretability remains limited. Additionally, hybrid models that combine CNN and transformer-based architectures have not been extensively studied. This work systematically evaluates CNNs (ResNet-50), ViTs (Vision Transformer and SwinV2-Tiny), and hybrid models (Convolutional Vision Transformer, LeViT-256, and CvT-13) on DR classification using publicly available retinal image datasets. The models are assessed based on classification accuracy and interpretability, applying Grad-CAM and Attention-Rollout to analyze decision-making patterns. Results indicate that hybrid models outperform both standalone CNNs and ViTs, achieving a better balance between local feature extraction and global context awareness. The best-performing model (CvT-13) achieved a Quadratic Weighted Kappa (QWK) score of 0.84 and an AUC of 0.93 on the test set. Interpretability analysis shows that CNNs focus on fine-grained lesion details, while ViTs exhibit broader but less localized attention. These findings provide valuable insights for optimizing deep learning models in medical imaging, supporting the development of clinically viable AI-driven DR screening systems.}, subject = {Diabetic retinopathy}, language = {en} } @article{SchultzFenskeOttoetal., author = {Schultz, Christof and Fenske, Markus and Otto, Nicolas and Dion-Bertrand, Laura-Isabelle and G{\´e}linas, Guillaume and Marcet, St{\´e}phane and Dagar, Janardan and Schlatmann, Rutger and Unger, Eva and Stegemann, Bert}, title = {Loss Analysis of P3 Laser Patterning of Perovskite Solar Cells via Hyperspectral Photoluminescence Imaging}, series = {Solar}, volume = {5}, journal = {Solar}, number = {2}, editor = {Boucl{\´e}, Johann}, publisher = {MDPI}, issn = {2673-9941}, doi = {10.3390/solar5020013}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20545}, abstract = {Upscaling perovskite solar cells and modules requires precise laser patterning for series interconnection and spatial characterization of cell parameters to understand laser-material interactions and their impact on performance. This study investigates the use of nanosecond (ns) and picosecond (ps) laser pulses at varying fluences for the P3 patterning step of perovskite solar cells. Hyperspectral photoluminescence (PL) imaging was employed to map key parameters such as optical bandgap energy, Urbach energy, and shunt resistance. The mappings were correlated with electrical measurements, revealing that both ns and ps lasers can be utilized for effective series interconnections with minimal performance losses at optimized fluences. Our findings provide a deeper understanding of fluence-dependent effects in P3 patterning. Moreover, the results demonstrate that the process window is robust, allowing for reasonable cell performance even with deviations from optimal parameters. This robustness, coupled with the scalability of the laser patterning process, emphasize its suitability for industrial module production.}, subject = {Photoluminescence}, language = {en} } @article{MpofuStepulaGuethetal., author = {Mpofu, Nonsikelelo Sheron and Stepula, Elzbieta and G{\"u}th, Uwe and Ehrmann, Andrea and Sabantina, Lilia}, title = {Electrospinning Poly(acrylonitrile) (PAN) Nanofiber Mats with Mushroom Mycelium Powder}, series = {Engineering Proceedings}, volume = {87}, journal = {Engineering Proceedings}, number = {1}, editor = {Morais, Simone}, publisher = {MDPI}, issn = {2673-4591}, doi = {10.3390/engproc2025087045}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20539}, abstract = {Electrospinning is a technique to produce nanofiber mats for diverse applications. In biomedicine in particular, the addition of an antibacterial agent can be advantageous. Here, we report on the needleless electrospinning of nanofiber mats using poly(acrylonitrile) (PAN) blended with different mushroom mycelium powders, which have antibacterial and other functional properties. While PAN blended with Pleurotus ostreatus (oyster mushroom) powder could be electrospun well, PAN blended with Ganoderma lucidum (reishi mushroom) powder was nearly impossible to spin. The PAN/ P. ostreatus nanofiber mats showed a morphology after electrospinning similiar to pure PAN; however, the carbon yield was lower. This indicates the possibility of embedding P. ostreatus powder in PAN nanofiber mats for biotechnological or biomedical applications.}, subject = {Electrospinning}, language = {en} } @misc{Grundmann, type = {Master Thesis}, author = {Grundmann, Max}, title = {Comparing Vision Transformers to Traditional Deep Learning Object Classification Architectures for Automotive Make and Model Recognition}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {102}, abstract = {The Transformer architecture has recently taken over the domain of natural language processing and is now quickly gaining popularity in computer vision, achieving state-of-the-art results on image classification benchmarks like ImageNet. Vision Transformers (ViT) work by dividing an image into fixed-size patches that can be processed as a sequence of tokens. Self-attention is used to compute the importance of these tokens in relation to each other, allowing the model to attend to different regions of the image and model long-range dependencies. This architecture has proven to work well on large datasets and shows better scaling performance than comparable convolutional neural networks (CNN). However, its performance on medium-sized datasets remains unclear. This thesis set out to study the performance potential and limitations of the Vision Transformer architecture compared to the dominant convolutional architecture on a medium-sized dataset for vehicle model classification. For reference, an Inception ResNet v2 model provides a baseline for accuracy on this task. After a review of the basic concepts and building blocks of both architectures, a selection of promising, improved architectures is assembled, with each tackling one of the most prominent weaknesses of the Transformer: DeiT solves the reliance of ViT on large datasets for pertaining by introducing a distillation strategy that can utilise pre-trained CNNs to inject some inductive biases that Transformers are naturally lacking. Next, Swin introduces an approach using shifted windows and a hierarchical design, inspired by convolutional networks, solving the issue of quadratic scaling with input resolution. Finally, FlexiViT shows how fixed-sized patch embeddings can be made dynamic, making ViTs easy to adapt to specific computational budgets by allowing dynamic patch sizes as inputs. A similar selection was made for top-performing convolutional networks as well as hybrid networks, which have been evaluated on a dataset of 42.000 Mercedes-Benz vehicle images. The evaluation results show very competitive results for nearly all tested architectures, with the Swin-T/8 v2 network achieving the highest accuracy with 96,3\%, closely followed by ConvNeXt-Base, a convolutional model, with an accuracy of 95,6\%. Both models are able to outperform the previous Inception ResNet v2 baseline, with the largest improvements seen in the representation of minority classes in this imbalanced dataset, manifesting in an 8,6\% improvement in the F1-score over the baseline model (81,6\% to 90,2\%). In terms of inference speed, convolutional models are faster (48,1ms for Swin-T/8 v2 versus 29,1ms for ConvNeXt-Base), but trade-offs can be made to bring inference times down, resulting in only slightly reduced accuracy. Training times are also generally slower but more sample-efficient. In conclusion, Transformers have demonstrated competitive performance by offering slightly superior accuracy at the cost of slower image throughput and slower training times. Both the best-performing CNN and Transformer easily surpass the previous baseline model, especially for minority classes. Since the real-world performance of these models has proven to be near-identical, a final decision would slightly favour the convolutional model for its faster inference speed.}, language = {en} } @article{KochBeyer, author = {Koch, Florian and Beyer, Sarah}, title = {A gift from heaven: Google's Environmental Insights Explorer and its tech-down approach to monitor urban sustainability beyond local contexts}, series = {Urban Studies}, journal = {Urban Studies}, publisher = {Sage}, doi = {https://doi.org/10.1177/00420980251340280}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20996}, pages = {1 -- 36}, abstract = {Google's Environmental Insights Explorer (EIE) exemplifies a tech-down approach by leveraging proprietary data, machine learning, and artificial intelligence to analyze urban greenhouse gas (GHG) emissions, solar potential, and tree canopy coverage. While framed as a tool for cities to assess and improve sustainability, the EIE often overlooks localized contexts and depends on opaque data generation processes that follow commercial interests. This study evaluates the EIE's data assemblage through critical urban data studies, focusing on its technical and contextual stacks. It critiques the EIEs lack of transparency, dependence on proprietary data, and marginalization of local expertise. Despite its promise, the EIE raises significant concerns about power imbalances, data validity, and urban governance implications. Cities utilizing the EIE must reconcile its global standardization with specific local needs, navigating a complex landscape shaped by Google's corporate interests. The findings highlight the need for a more inclusive, context-specific approach to urban sustainability metrics that balances innovative data use with transparency and equitable stakeholder engagement. The EIE demonstrates how sustainability indicators and data are defined through a technology-driven process shaped by a private company, while neglecting local knowledge and narratives. This process - referred to as a tech-down approach - differs from both bottom-up and top-down methods of sustainability indicator selection, as well as from data philanthropy initiatives by private companies, due to its exclusive reliance on corporate data and its global scope. This study urges critical reflection on the broader implications of tech-driven urban monitoring tools like the EIE.}, subject = {Smart City}, language = {en} } @article{Kalff, author = {Kalff, Yannick}, title = {Managing Projects as a Mode of Temporal Ordering}, volume = {15}, number = {1}, publisher = {Emerald Insight}, doi = {https://doi.org/10.1108/IJMPB-05-2021-0132}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20781}, abstract = {Purpose: Project studies analyse either managing practices or the temporal nature of project management, which leaves open a research gap: the temporality of managing practice. The contribution demonstrates that performativity theory with a temporal perspective helps us to understand how managing a project organises limited temporal resources by aligning activities, deadlines or milestones to reach a goal in a given time. Design/methodology/approach: The article utilises empirical data and grounded theory methodology. Ten interviews with project managers from two companies support empirically guided theory building and conceptual reasoning. Findings: The article extends John Law's "modes of ordering" to a project-specific mode of temporal ordering. This mode of temporal ordering describes the underlying rationale of project managers who assign, order, and materialise time to generate the temporal structure of the project. Originality: The contribution links managing practice and temporality in a performativity approach. This link focuses the actual actions of the managers and contextualises them in the temporal flow of the project. Managing projects as a mode of temporal ordering describes how project managers enact temporal structures and how they themselves and their activities are temporally embedded. Research limitations: The conceptual nature of the paper and its limited empirical data restrict the generalization of the findings. The article's goal is to initiate further research and to offer a set of tools for such research.}, language = {en} } @inproceedings{FernselKalffSimbeck, author = {Fernsel, Linda and Kalff, Yannick and Simbeck, Katharina}, title = {Audits for Trust: An Auditability Framework for AI-Based Learning Analytics Systems}, series = {Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU}, volume = {2}, booktitle = {Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU}, publisher = {SciTePress}, doi = {10.5220/0013254300003932}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20790}, pages = {51 -- 62}, abstract = {Audits contribute to the trustworthiness of Learning Analytics (LA) systems that integrate Artificial Intelligence (AI) and may be legally required in the future. We argue that the efficacy of an audit depends on the auditability of the audited system. Therefore, systems need to be designed with auditability in mind. We present a framework for assessing the auditability of AI-integrating systems in education that consists of three parts: (1) verifiable claims about the validity, utility and ethics of the system, (2) evidence on subjects (data, models, or the system) in different types (documentation, raw sources and logs) to back or refute claims, (3) means to validate evidence such as technical APIs, monitoring tools, or explainable AI principles must be accessible to auditors. We apply the framework to assess the auditability of the Learning Management System Moodle, which supports an AIintegrating dropout prediction system. Moodle's auditability is limited by incomplete documentation, insufficient monitoring capabilities, and a lack of available test data.}, language = {en} } @article{Kalff, author = {Kalff, Yannick}, title = {The knowledge worker and the projectified self: domesticating and disciplining creativity}, series = {Work Organisation, Labour and Globalisation}, volume = {11}, journal = {Work Organisation, Labour and Globalisation}, number = {1}, publisher = {Pluto Journals}, doi = {10.13169/workorgalaboglob.11.1.0010}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20805}, pages = {10 -- 27}, abstract = {Since the 1970s, projects have become a widespread form of work and labour organisation. They represent complex temporary endeavours and suggest spontaneity and flexibility, albeit reliant on bureaucratic structures. In contrast to this, projects can also be seen as metaphors for late modern biographies and socialisation. Both notions directly affect knowledge work by domesticating knowledge and creativity to fit valorisation processes. Within obscured power relations, a subjectivising antinomy of predictability and flexibility arises: the projectified self. This form of subjectivity domesticates the knowledge worker contributing to new vulnerabilities to precarity due to flexible, spontaneous and uncertain biographies.}, language = {en} } @inproceedings{FernselKalffSimbeck, author = {Fernsel, Linda and Kalff, Yannick and Simbeck, Katharina}, title = {Where Is the Evidence? A Plugin for Auditing Moodle's Learning Analytics}, series = {Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024)}, volume = {2}, booktitle = {Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024)}, publisher = {SciTePress}, doi = {10.5220/0012689800003693}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20818}, pages = {262 -- 269}, abstract = {The paper presents the work-in-progress development of a Moodle plugin to improve the auditability of Moodle's Learning Analytics component. Future legislation, such as the EU AI Act, will require audits and "conformity assessments" of "high-risk" AI systems. Educational applications can be considered high-risk systems due to their important role in individual life and career paths. Therefore, their correctness, fairness, and efficiency must be assessed. However, auditing of the Learning Analytics functions in Moodle is limited. No suitable test-data is available, models and configurations are not persistent and only aggregated quality metrics are returned that are insufficient to assess fairness. The plugin addresses these issues and provides a data interface to extract data for audits. The plugin allows to a) upload and select data for the audit, b) clearly differentiate between model configuration and trained models, c) keep trained models, their configuration and underlying data for future inspections and comparisons, and finally, d) the plugin saves raw predictions for further analysis. The plugin enables the audit of Moodle's Learning Analytics and its underlying AI models and contributes to increased fairness and trustworthiness of Learning Analytics as well as its legally compliant application.}, language = {en} } @masterthesis{Fuchs, type = {Bachelor Thesis}, author = {Fuchs, Marie}, title = {The contrast of wabi sabi and modernism and how both concepts combined can act as an approach for sustainable fashion}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {87}, abstract = {This thesis examines the contrasting aesthetics of modernism and wabi sabi. Wabi sabi, a Japanese aesthetic concept encompassing as the wisdom of natural simplicity, is rooted in notion of embracing imperfection and incompleteness. In contrast to wabi sabi is the concept of modernism, which revolves around the idea of functionality and efficiency. The essence of both concepts can be adapted to everyday life today. The obtaining of functionality and efficiency is suggested as worth striving for. The world is becoming increasingly fast-paced, which is also evident in the economic sector, literally epitomized in the term „fast fashion". To create a solution for this problematic is to combine the values of wabi sabi with the general aesthetics of the Western world and modernism to create a more complex, versatile aesthetic that takes into account the environment. To attain this goal, it is necessary to understand the core tenets of both concepts and place them in the context of modern life. This entails examining how wabi sabi can exist in modern life and examine its compatibility with digitization. Not only digital space, but also the correlation with real space will be considered, as well as the broader role of space and time in design. Moreover, the similarities and differences between the two concepts will be researched. In order to create a sustainable collection that not only takes environmental factors into account but also harmonizes the aesthetics of both concepts, it will be examined how they can be adapted to sustainable fashion.}, language = {en} } @masterthesis{Bueltemann, type = {Bachelor Thesis}, author = {B{\"u}ltemann, Marlene}, title = {Sustainable Learning Analytics: Measuring and Understanding the Drivers of Energy Consumption of AI in Education}, school = {Hochschule f{\"u}r Technik und Wirtschaft Berlin}, pages = {43}, abstract = {AIED (AIED) is used increasingly to achieve higher learning outcomes through adaptive personalized learning paths. However, its environmental impact has been disregarded thus far. In this thesis, local energy and resources consumption of the prediction in four AIED models are measured using two different measurement approaches. The results suggest that boolean or mathematically explainable models such as decision tree and logistic regression have a low energy demand. On the other hand, in models that use a complex optimization process or have mulitple interconnected layers like support vector machine (SVM) or multilayer perceptron (MLP), the resource demand rises significantly. A decision tree model is recommended as the model of choice to deploy on the learning platform as it reveals the lowest emissions in both measurements and shows high performance prediction values. The thesis concludes with recommendations for AIED developers and policy ecommendations and pledges for higher attention on the environmental impact of AIED.}, language = {en} } @inproceedings{JohstPerlotHanneetal., author = {Johst, Abraham and Perlot, Nicolas and Hanne, Peter and Rothe, Marcel and Schreier, Andy and N{\"o}lle, Markus and Freund, Ronald}, title = {Bidirectional real-time 10-Gbit/s free-space optical link with a 4-aperture array under strong scintillation}, series = {2023 IEEE International Conference on Space Optical Systems and Applications (ICSOS)}, booktitle = {2023 IEEE International Conference on Space Optical Systems and Applications (ICSOS)}, doi = {10.1109/ICSOS59710.2023.10491221}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20383}, pages = {6}, abstract = {We report on a demonstration of a bidirectional, multi-aperture free-space optics (FSO) system and discuss general system design aspects. A multi-aperture terminal is used for spatial diversity. On both sides, a real-time signal combining takes place on field-programmable gate arrays. The combined signals are weighted by their corresponding channel powers. Small form-factor C-band pluggable modules provide the optoelectrical interfaces of the 10-Gbit/s on/off-keying signals. A custom-built multi-channel bit error tester is used on both terminals for performance determination. Each aperture has an optical tracking system, is coupled to a bidirectional single-mode fiber and emits an optical power of 4 dBm.}, language = {en} }