20005 research outputs found
Sort by
One-shot learning in hybrid system identification: a new modular paradigm
No full textIdentification of hybrid systems requires learning models that capture both discrete transitions and continuous dynamics from observational data. Traditional approaches follow a stepwise process, separating trace segmentation and mode-specific regression, which often leads to inconsistencies due to unmodeled interdependencies. In this paper, we propose a new iterative learning paradigm that jointly optimizes segmentation and flow function identification. The method incrementally constructs a hybrid model by evaluating and expanding candidate flow functions over observed traces, introducing new modes only when existing ones fail to explain the data. The approach is modular and agnostic to the choice of the regression technique, allowing the identification of hybrid systems with varying levels of complexity. Empirical results on benchmark examples demonstrate that the proposed method produces more compact models compared to traditional techniques, while supporting flexible integration of different regression methods. By favoring fewer, more generalizable modes, the resulting models are not only likely to reduce complexity but also simplify diagnostic reasoning, improve fault isolation, and enhance robustness by avoiding overfitting to spurious mode changes
UPPR: universal privacy-preserving revocation
No full textSelf-Sovereign Identity (SSI) frameworks enable individuals to receive and present digital credentials in a user-controlled way. Revocation mechanisms ensure that invalid or withdrawn credentials cannot be misused. These revocation mechanisms must be scalable (e.g., at national scale) and preserve core SSI principles such as privacy, user control, and interoperability. Achieving both is hard, and finding a suitable trade-off remains a key challenge in SSI research.This paper introduces UPPR, a revocation mechanism for One-Show Verifiable Credentials (oVCs) and unlinkable Anonymous Credentials (ACs). Revocations are managed using percredential Verifiable Random Function (VRF) tokens, which are published in a Bloom filter cascade on a blockchain. Holders prove non-revocation via a VRF proof for oVCs or a single Zero-Knowledge Proof for ACs. The construction prevents revocation status tracking, allows holders to stay offline, and hides issuer revocation behavior. We analyze the privacy properties of UPPR and provide a prototype implementation on Ethereum. Our implementation enables off-chain verification at no cost. On-chain checks cost 0.56–0.84 USD, while issuers pay only 0.00002–0.00005 USD per credential to refresh the revocation state
Parametrized statistical appearance and shape modelling strategy to predict proximal and diaphyseal femoral fractures
No full textIntroduction: Femoral loading leading to a fracture is known to vary with anthropometry, and patient-specific finite element models have provided important insights into fracture prediction but are often very time consuming to generate. Additionally, existing parametric models do not simultaneously account for variations in both femur geometry and bone density distribution and remain limited to either the femoral shaft or the proximal femur. This inhibits their ability to predict fractures involving both the shaft and proximal regions.
Methods: In the present study, a novel parametric femur modeling strategy was developed to create whole femur models based on stature, BMI, and age input, including density distribution and geometrical variations, for fracture loading predictions. A statistical shape and appearance femur model was developed based on an input set of CT scans of healthy female femurs (N = 18) between the ages of 50 and 70. Thereafter, multilinear regressions were used to relate principal components to the subject anthropometric characteristics and develop parametric models. The developed parametric models were evaluated using traditional patient-specific models for their potential to represent the influence of changing patient stature, BMI, and age on femoral fractures. Femoral fracture load in three-point bending, axial torsion, and lateral fall cases was predicted using the parametric as well as subject-specific femur models.
Results: The developed parametric model was able to predict femoral fracture load variations due to changing anthropometry and age with an average difference of 4.85% compared with predictions using subject-specific models.
Discussion: Therefore, this novel parametric femur model can predict fracture loading while directly incorporating the influence of changing patient anthropometry. In the future, the model could support the development of orthopedic devices tailored to specific patient anthropometries to help mitigate femoral fractures
Beruf und Beruflichkeit - Phänomene der Vergangenheit oder Garanten zukunftsfester Facharbeit
No full textDie mit der Digitalisierung einhergehende hohe Innovationsdynamik führt zu einem Trend der Höherqualifizierung und stellt die Bedeutung gewerblich-technischer Berufe in Frage. Empirisch kann allerdings gezeigt werden, dass gewerblich-technische Berufe einen Beitrag leisten, den Digitalisierungsprozess zu unterstützen. Träger von Berufen verfügen über breit gefächerte Handlungskompetenzen, die die Flexibilität auf dem Arbeitsmarkt deutlich erhöhen. Das Berufskonzept ist allerdings nur stabil, wenn es die Veränderungen selbst mit aufnimmt und Antworten auf veränderte Anforderungen durch das Individuum, die Unternehmen und die Gesellschaft geben kann. Im Beitrag werden die Veränderungen mit ihren Wirkungen auf gewerblich-technische Berufe und Kriterien für eine moderne, innovationsfördernde Beruflichkeit diskutiert. Es werden Thesen für ein zukunftsfestes duales System formuliert, die gewerblich-technische Berufe hervorbringen.The high innovation dynamics associated with digitalization are leading to a trend towards higher qualifications and calling into question the importance of industrial-technical professions. Empirically, however, it can be shown that industrial-technical occupations contribute to supporting the digitalization process. Holders of occupations have a wide range of skills, which results in a high degree of flexibility for employees on the labour market. However, the occupational concept is only stable if it is able to absorb the changes themselves and provide answers to the changing requirements of the individual, society and companies. The article discusses the changes and their effects on industrial-technical occupations and criteria for a modern, innovation-promoting occupation. Theses are formulated for a future-proof dual system that produces such professions
EBPR process dynamics under variable conditions: interaction between PAOs and the microbial community
No full textThe performance of enhanced biological phosphorus removal (EBPR) is influenced by various process conditions, such as the type of carbon source, temperature, COD:P ratio or pH-value. The activity of polyphosphate accumulating organisms (PAOs) and the composition of biocoenosis are strongly influenced by these process conditions. Nevertheless, critical knowledge gaps remain regarding the relevance of microbial diversity and the correlations between PAO/GAO community composition and EBPR process dynamics under different operational conditions. This study demonstrates that Dechloromonas plays a crucial role in the EBPR process and that the type of carbon source has a more sensitive influence on EBPR dynamics than the COD:P ratio, temperature or pH. Using both a microfluidic model biofilm system (MMBS) and a sequencing batch reactor (SBR) system, we found that the consumption of glucose resulted in maximal PAO activity, whereas EBPR performance was limited when using glycerol, ethanol and amino acids. In both systems, Dechloromonas was the dominant PAO, while in MMBS Zoogloea and in SBR Competibacter/Contendobacter were the dominant GAOs. Tessaracoccus (PAO) and Micropruina (GAO) were relevant in the fermentation of glucose and glycerol. No significant effect on EBPR performance was observed between 12 and 20 °C, and PAO activity peaked at pH 7.5. Our findings highlight that PAO-GAO competition is overestimated. We anticipate that our study provides a deeper understanding of the dynamics in the biocoenosis and emphasizes the relevance of biodiversity in EBPR systems, which would help guide relevant applications
Implementation and evaluation of CSBM for intra-satellite communication with cuboid-based signal-space generated symbols
No full textLiFi for intra-satellite communication offers immense advantages like flexible AIT or reduced complexity (harness). However, high bandwidths and redundancies are equally required. Modulation methods that make use of the broad spectrum of light are (Color Space Based Modulations (CSBMs)). However, this requires precise knowledge of the transceivers and environments, as previous methods usually map to the CIE 1931 color scheme. But for intra-satellite communication, various assumptions can be made that favor the use of CSBM within the satellite. This paper presents an automated procedure that generates the symbols for CSBM. In order to ensure high reliability while using the entire color space for the symbols, a method based on cuboids is presented, which guarantees an overlap-free mapping between Transmit- and Signal-Space. In addition, the implementation of a Receiver based on an Field Programmable Gate Array (FPGA) is presented and real world measurements are conducted in detail to show the automatic symbol generation and the evaluation of symbol detection capabilities for communication
Machine unlearning: bias correction in neural network downscaled storms
No full textAccurate precipitation at fine spatial resolutions is essential for hydrologic modeling and risk assessment, yet most precipitation data are available at coarse scales. Dynamical downscaling can improve spatial resolution but is computationally expensive while statistical downscaling struggles to reproduce high-resolution characteristics. Machine learning has been shown to offer an operational alternative for transforming coarse data into fine-scale fields, especially honoring spatiotemporal precipitation dependencies. Here, we show that combining machine learning with post-processing bias correction approaches—a form of “machine unlearning”— yields improved performance. We evaluate four machine-learning models—Linear Network (LNet), Fully Connected Network (FCNet), Convolutional Neural Network (UNet), and Wasserstein Generative Adversarial Network (WGAN)—for downscaling precipitation using synthetic benchmark storms with known marginal and spatiotemporal properties. Using synthetic fields provides full control over storm characteristics and enables rigorous evaluation. Raw outputs from all models struggle to reproduce wet/dry boundaries, statistics, and extremes, and only WGAN captures the complex spatiotemporal structure of fine-scale storms. We then apply linear and nonlinear bias corrections to enforce zeros, match the mean of positive values, and align full marginal distributions, including tails. This demonstrates that post-processing is crucial for reliable neural network outputs in operational settings. The results highlight WGAN’s potential for operational downscaling while emphasizing the need for systematic post-processing and careful validation before real-world application
The ethics of analog AI
No full textAI ethics has matured as a field, yet its concepts and tools have been developed almost entirely with digital AI in mind. Emerging “analog” AI approaches, that is, systems that compute over continuous, rather than binary, signals, challenge this digital default. Analog AI promises substantial energy savings and alternative modes of information processing. However, it also departs in ethically significant ways from the widely discussed digital AI systems such that its differences from digital AI could mean that current ethical frameworks may not address analog AI’s unique challenges. This paper highlights how the distinctive characteristics of analog AI raise new ethical questions that require careful consideration. We examine four key areas of AI ethics, namely fairness, privacy, explainability, and safety, and propose a roadmap for the ethical examination of analog AI. Our aim is to lay the foundation for an important but hitherto unrecognized subfield within AI ethics: the ethics of analog AI
A first quantitative assessment of soil health at European scale considering soil genesis
No full textBackground: Soil health degradation is a major threat to European food security, biodiversity, and climate stability. While scientists have debated how to define soil health during recent decades, a quantifiable framework for monitoring, management,
and policy remains lacking.
Aim:We introduce SHERPA (SoilHealth Evaluation, Rating Protocol, and Assessment) as a framework for discussion and present
a first quantitative soil health assessment across Europe.
Methods: All major soil degradation processes (with the exception of organic contamination) were scored, averaged, and
subtracted from the intrinsic soil health resulting in quantitative final scores.
Results: As reported before, cropland soils throughout Europe are highly degraded. Surprisingly, soil health of grasslands is
also very negatively impacted. Soil erosion, nutrient surplus, and pesticide risk are largely driving poor soil health aligning with
reported high biodiversity loss in agricultural land. Forest soils are also surprisingly low in health, mainly because of nitrogen
surplus, reflecting documented widespread forest decline from nutrient imbalances. Interactive maps highlight specific threats to
soil health across Europe, offering valuable insights for targeted action.
Conclusions: SHERPA is able to quantify soil health across Europe. However, at the current state of data availability, soil health
is likely to be overestimated.Monitoring data of soil structure, compact
Soft tissue characterization from miniaturized imaging probes
No full textCharacterizing soft tissue properties is essential for diagnosing and treating diseases. Optical coherence tomography offers high spatial and temporal resolution for analyzing small lesions. However, when applied in the body, image acquisition and interpretation are complicated by relative movements and deformations. This thesis investigates three miniaturized probe designs and image analysis methods, as well as the influence of motion on the assessment of tissue properties, underlying the potential of these probes for improving in-vivo tissue characterization.Die Charakterisierung von Weichgewebeeigenschaften ist entscheidend für die Diagnose und Behandlung von Krankheiten. Die optische Kohärenztomographie bietet eine hohe räumliche und zeitliche Auflösung zur Analyse kleiner Läsionen. Bei der Anwendung im Körper wird die Bildaufnahme und -interpretation jedoch durch relative Bewegungen und Deformationen erschwert. Diese Arbeit untersucht drei miniaturisierte Sondendesigns und Bildanalyseverfahren sowie den Einfluss von Bewegungen auf die Beurteilung von Gewebeeigenschaften und hebt das Potenzial dieser Sonden zur Verbesserung der in-vivo-Gewebecharakterisierung hervor