arXiv subjects

Dominik Schön

Publications and source records attributed to Dominik Schön.

5 recordsLinked to original sources

arXiv2026

PRISM: Predictive Runtime In-place Scaling and Model Selection for Edge Microservices

Latency-sensitive edge AI services must balance strict deadlines, output quality, and limited compute and energy budgets. However, static CPU provisioning wastes resources because inference cost varies substantially across inputs, model variants, and runtime conditions. We present PRISM, a prediction-guided runtime framework that jointly selects model variants and CPU allocations for containerized edge microservices. Using container-level energy monitoring and lightweight regression models, PRISM adapts each pipeline stage in place and minimizes predicted CPU-package energy under deadline, resource, and offline model-level Quality of Result (QoR) constraints. We evaluate PRISM on more than 52,000 requests in an Automatic License Plate Recognition (ALPR) pipeline with detection and recognition stages. For detection, PRISM reduces energy consumption by 36 % compared to the strongest static configuration while preserving a comparable success rate and using less than half of the average CPU allocation. For recognition, it reaches near-static-best performance with lower average CPU allocation. These results show that predictive in-place adaptation is a practical mechanism for making time-sensitive AI microservice pipelines more energy-efficient at the edge.

cs.DC↗

arXiv2024

"Create a Fear of Missing Out" -- ChatGPT Implements Unsolicited Deceptive Designs in Generated Websites Without Warning

With the recent advancements in Large Language Models (LLMs), web developers increasingly apply their code-generation capabilities to website design. However, since these models are trained on existing designerly knowledge, they may inadvertently replicate bad or even illegal practices, especially deceptive designs (DD). This paper examines whether users can accidentally create DD for a fictitious webshop using GPT-4. We recruited 20 participants, asking them to use ChatGPT to generate functionalities (product overview or checkout) and then modify these using neutral prompts to meet a business goal (e.g., "increase the likelihood of us selling our product"). We found that all 20 generated websites contained at least one DD pattern (mean: 5, max: 9), with GPT-4 providing no warnings. When reflecting on the designs, only 4 participants expressed concerns, while most considered the outcomes satisfactory and not morally problematic, despite the potential ethical and legal implications for end-users and those adopting ChatGPT's recommendations

cs.HC↗

arXiv2024

3DA: Assessing 3D-Printed Electrodes for Measuring Electrodermal Activity

Electrodermal activity (EDA) reflects changes in skin conductance, which are closely tied to human psychophysiological states. For example, EDA sensors can assess stress, cognitive workload, arousal, or other measures tied to the sympathetic nervous system for interactive human-centered applications. Yet, current limitations involve the complex attachment and proper skin contact with EDA sensors. This paper explores the concept of 3D printing electrodes for EDA measurements, integrating sensors into arbitrary 3D-printed objects, alleviating the need for complex assembly and attachment. We examine the adaptation of conventional EDA circuits for 3D-printed electrodes, assessing different electrode shapes and their impact on the sensing accuracy. A user study (N=6) revealed that 3D-printed electrodes can measure EDA with similar accuracy, suggesting larger contact areas for improved precision. We derive design implications to facilitate the integration of EDA sensors into 3D-printed devices to foster diverse integration into everyday objects for prototyping physiological interfaces.

cs.HC↗

arXiv2023

TicTacToes: Assessing Toe Movements as an Input Modality

From carrying grocery bags to holding onto handles on the bus, there are a variety of situations where one or both hands are busy, hindering the vision of ubiquitous interaction with technology. Voice commands, as a popular hands-free alternative, struggle with ambient noise and privacy issues. As an alternative approach, research explored movements of various body parts (e.g., head, arms) as input modalities, with foot-based techniques proving particularly suitable for hands-free interaction. Whereas previous research only considered the movement of the foot as a whole, in this work, we argue that our toes offer further degrees of freedom that can be leveraged for interaction. To explore the viability of toe-based interaction, we contribute the results of a controlled experiment with 18 participants assessing the impact of five factors on the accuracy, efficiency and user experience of such interfaces. Based on the findings, we provide design recommendations for future toe-based interfaces.

cs.HC↗

arXiv2023

UndoPort: Exploring the Influence of Undo-Actions for Locomotion in Virtual Reality on the Efficiency, Spatial Understanding and User Experience

When we get lost in Virtual Reality (VR) or want to return to a previous location, we use the same methods of locomotion for the way back as for the way forward. This is time-consuming and requires additional physical orientation changes, increasing the risk of getting tangled in the headsets' cables. In this paper, we propose the use of undo actions to revert locomotion steps in VR. We explore eight different variations of undo actions as extensions of point\&teleport, based on the possibility to undo position and orientation changes together with two different visualizations of the undo step (discrete and continuous). We contribute the results of a controlled experiment with 24 participants investigating the efficiency and orientation of the undo techniques in a radial maze task. We found that the combination of position and orientation undo together with a discrete visualization resulted in the highest efficiency without increasing orientation errors.

cs.HC↗