Digital Library of Gesellschaft für Informatik e.V.
Not a member yet
40758 research outputs found
Sort by
The Co-Creation of Evoking Change Talk: A Method for Shifting Workplace Culture
No full text"This paper introduces Evoking Change Talk, a gender-responsive method co-created within a Feminist Living Lab in a global tech company. The approach uses short, recurring knowledge impulses on topics like implicit bias as starting points for team-based dialogue. These micro-interventions are embedded into regular meetings and aim to trigger ongoing “change talk”—structured conversations that promote reflection and inclusive action within teams. Grounded in behavioral design and feminist theory, the method shifts the focus from one-time awareness to continuous cultural learning. Initial implementation in some developer and leadership teams suggest increased awareness, surfacing of blind spots, and first steps toward inclusive practices. Reflective diaries capture these shifts and support iterative development. Evoking Change Talk offers a low-threshold, transferable model for embedding equity into everyday tech work and highlights the importance of peer facilitation, psychological safety, and small-scale interventions as levers for structural change
Generative KI in Bildungskontexten: Eine Analyse auftretender Diskriminierung und ihrer Implikationen am Beispiel von ChatGPT
No full textDie zunehmende Verbreitung leistungsfähiger Künstlicher Intelligenz (KI), insbesondere durch Chatbots wie ChatGPT, eröffnet neue Möglichkeiten im Bildungswesen, wirft jedoch auch Fragen hinsichtlich potenzieller Diskriminierungen auf. Dieser Text geht aus einer interdisziplinären Perspektive exemplarisch der Frage nach, inwiefern der Einsatz von ChatGPT im Bildungskontext zu diskriminierenden Effekten führen kann. Die Ergebnisse stammen aus einer Bachelorarbeit und wurden mit dem Ziel aufbereitet, Lesende zu einer kritischen Auseinandersetzung im eigenen Umgang mit LLMs am Beispiel von ChatGPT anzuregen. Der Text richtet sich insbesondere an Lehrkräfte und soll ihre Sensibilität für mögliche Diskriminierungspotenziale beim Einsatz von KI-Systemen wie ChatGPT im Kontext Bildung stärken, sowie dazu anregen, explorativ, die in ChatGPT vorkommenden Verzerrungen zu erkennen. Dazu wird zum einen mit Bildungszenarien gearbeitet, in denen der Einsatz von ChatGPT zu Diskriminierung führen kann. Zum anderen werden Paare aus Promt und zugehöriger Antwort verwendet, die inhärente Verzerrungen verdeutlichen. Abschließend erfolgt eine kurze Sammlung an Erfordernissen, die für einen Einsatz von LLMs im Kontext Bildung wichtig sind
MindSpace: Improving Relaxation Break and Performance Using a Device Exterting Tactile Sensation of Life-like Breathing Movement with Squeeze-like Deep Touch Pressure on Users' Upper Chest Area
No full textWe introduce MindSpace, a pneumatically controlled haptic device that simulates gentle, life-like breathing combined with providing a deep touch sensation. Designed as a focusing agent, MindSpace aims to enhance users' short relaxation breaks, promoting mental clarity, and stress reduction. In a within-subject study (N = 22), we compared rest conditions with and without the device. Our findings provide initial evidence that MindSpace can effectively induce relaxation, increase environmental and bodily awareness, and improve post-break responsiveness. Participants reported that taking short breaks with MindSpace improved physical ease and mental calmness, thus overall relaxation and productivity. Built on prior works, MindSpace integrates slow tactile breathing rhythms with deep pressure, offering a novel multi-sensory approach tailored for micro-break relaxation
A prototype for Simultaneous Localization and Mapping (SLAM) with Fischertechnik Robotics for Education
No full textEmbedded Systems play a crucial role in modern systems development. Due to increasing miniaturization and computational power, they are ubiquitously used in various domains, e.g. automotive, life sciences, and industrial automation. It is therefore essential to include a profound education in this domain in the curricula of computer science and business informatics. A challenge is to provide students with a hands-on experience in the field of embedded systems, i.e. an application that is complex enough to be realistic and requires both conceptual and low-level programming skills, but is also simple enough to be manageable in a semester project. In this paper, we present a prototype for Simultaneous Localization and Mapping (SLAM) with Fischertechnik Robotics for Education. The goal was to evaluate if Fischertechnik robots are suitable to support our goal andto identify the challenges of using them in a SLAM project. With this, we provide an idea for colleagues on how toapply this technology in academic education
Legal Challenges and Technical Solutions to Decentralised Digital Product Passports
No full textThe Digital Product Passport (DPP) constitutes a central element of the European Green Deal. While the Ecodesign for Sustainable Products Regulation (ESPR) introduces decentralisation as a guiding design principle for DPPs, its legal framework remains relatively vague. The paper at hand analyses the regulatory framework set up by the ESPR, examining the relationship between information requirements and DPPs, as well as the legal definition for DPPs introduced by the regulation. Based on this analysis – and after introducing a simplified model for a DPP system –, DPP decentralisation is discussed at three levels: (1) unique identifier minting, (2) data storage, and (3) distributed storage of DPP information. The discussion focusses at emerging legal challenges, while also mentioning applicable technical solutions and pointing at potential benefits of decentralised DPP structures. In particular, the model of distributed DPP storage is shown to be legally and operationally problematic
How Can Attention Guiding Techniques Help to Search Efficiently Without Getting Distracted? A Comparative Study.
No full textAttention Guiding Techniques (AGTs) in augmented reality (AR) assistance systems improve visual search efficiency but may cause important events to be missed. This study investigates both effects in a controlled AR environment using HoloLens 2. Participants searched for virtual target objects at varying search angles and reported changes during the task. Three AGT conditions were tested: Highlight, Highlight with 3D Arrow (+3DArrow), and Highlight with Attention Funnel (+AF). Both +3DArrow and +AF improved search times over Highlight. However, +AF was less effective than +3DArrow for targets behind participants. The condition +3DArrow caused significantly more change blindness than Highlight, likely due to the arrow’s central field of view (FOV) placement. In contrast, AF’s peripheral tunnel rings allowed better monitoring of changes. These results underscore the importance of selecting and designing AGTs according to task context to balance guidance efficiency with situational awareness
TAM System Modeler LuxEst: Visualizing Organizational Complexity Using Tree-Attribute-Matrix Models
No full textWith rising complexity, organizations are increasingly difficult to
manage efficiently. As a first step towards assessing and then
reducing organizational complexity, the structure of an organization,
its elements and their interdependencies, must be modeled
and visualized. Specific software tools can help to capture and
understand the elements contributing to organizational complexity.
This paper describes key features of TAM System Modeler
LuxEst, a software system being developed during more than 100
consulting projects. LuxEst aims at assessing and visualizing the
organizational complexity of an organization, including business
and public administrations. The software enables its users to create
Tree-Attribute-Matrix (TAM) models, to analyze and to manipulate
them
Sharpness Enhancement of Archaeological Images Using Grimaldo-Levin Networks with Unsharp Masking
No full textSharpness in archaeological images is crucial for effectively viewing fine details. Images are sometimes obtained unsharp, blurry, and perceived with unclear information. This paper introduces an improved unsharp masking method using the Grünwald-Letnikov (GL) weights to increase the sharpness of archaeological images. The classical unsharp masking method provides an overshoot effect and unnatural sharpness, leading to an unrealistic representation. GL weights are known for their adequate ability to capture vital attributes, exploiting them to extract significant image details. The extracted details are then used with the unsharp mask filter to provide natural sharpness without generating the overshoot effect. By following this approach, a significant improvement in the clarity of features is achieved, and a more natural sharpness and realistic appearance are delivered
Kreative Python-Programmierung mit Strype
No full textStrype ist eine neue, kostenlose Online-Programmierumgebung für Python, die sich an Programmieranfänger richtet – egal, ob sie völlig neu in der Programmierung sind oder von blockbasierten Systemen umsteigen. Zwei Innovationen zeichnen Strype aus: Erstens ermöglicht der Frame-basierte Editor einen einfacheren Programmstart, vermeidet viele Syntaxfehler und unterstützt Lernende durch bessere Sichtbarkeit und Programmmanipulation. Zweitens ermöglicht die neuartige Grafikunterstützung die Erstellung deutlich visuellerer und interaktiverer Beispiele von Anfang an und sorgt so für einen deutlich ansprechenderen Einstieg in die Programmierung. Zusammen mit der vollständig browserbasierten, installationsfreien Plattform bietet dies das Potenzial, die Art wie wir unterrichten zu verändern
Bericht zum ersten Treffen des Arbeitskreises Artificial Intelligence for Software Architecture
No full textBeim ersten konstituierenden Treffen des Arbeitskreises Artificial Intelligence for Software Architecture ging es darum, konkrete Ziele auszuarbeiten. Der Arbeitskreis innerhalb der Fachgruppe Architekturen widmet sich der Unterstützung und Weiterentwicklung von Softwarearchitekturen mit Hilfe von innovativen KI-gestützten Methoden für Entwurf, Analyse und Optimierung von Architekturen sowie deren Auswertung und der dazugehörigen Forschung