Konferenzschrift

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 5 of 105489
  • Publication
    Explaining Face Recognition Through SHAP-Based Pixel-Level Face Image Quality Assessment
    ( 2023)
    Biagi, Clara
    ;
    Rethfeld, Louis
    ;
    ;
    Biometric face recognition models are widely used in many different real-world applications. The output of these models can be used to make decisions that may strongly impact people. However, an explanation of how and why such outputs are derived is usually not given to humans. The lack of explainability of face recognition models leads to distrust in their decisions and does not encourage their use. The performance of face recognition models is influenced by the quality of the input image. In case the quality of a face image is too low, the face recognition system will reject it to avoid compromising its performance. The quality is evaluated by Face Image Quality (FIQ) approaches, which assigned quality scores to the input images. Pixel-level face image quality (PLFIQ) increases the explainability of quality scores by explaining face image quality at the pixel level. This allows the users of face recognition systems to spot low-quality areas and allows them to make guided corrections. Previous works introduced the concept of PLFIQ and proposed evaluation procedures. This work proposes a new way of computing PLFIQ values depending on given FIQ methods using Shapley Values. They score the contribution of each pixel to the overall image quality evaluation. Therefore, Integrating Shapley Values increases the explainability of the FIQ models. Results show that using these methods leads to significantly better and more robust PLFIQ values estimates and thus provide better explainability.
  • Publication
    Particle-Specific Deflection Windows for Optical Sorting by Uncertainty Quantification
    ( 2024)
    Reith-Braun, Marcel
    ;
    Liang, Kevin
    ;
    Pfaff, Florian
    ;
    ; ;
    Bauer, Albert
    ;
    Kruggel-Emden, Harald
    ;
    ; ;
    Hanebeck, Uwe D.
    In current state of the art sensor-based sorting systems, the length of the deflection windows, i.e., the period of nozzle activation and the number of nozzles to be activated, is commonly determined solely by the size of the particles. However, this comes at the cost of the sorting process not accounting for model discrepancies between actual and presumed particle motion, as well as for situations where the available information does not allow for precise determination of nozzle activations. To achieve a desired sorting accuracy, in practice, one is therefore usually forced to enlarge the deflection window to a certain degree, which increases the number of falsely co-deflected particles and compressed air consumption. In this paper, we propose incorporating the uncertainty of the prediction of particle motion of each individual particle into the determination of the deflection windows. The method is based on the predictive tracking approach for optical sorting, which tracks the particles while they move toward the nozzle array based on images of an area-scan camera. Given the extracted motion information from the tracking, we propose an approximation for the distribution of arrival time and location of the particle at the nozzle array assuming nearly constant-velocity or nearly constantacceleration particle motion behavior. By evaluating the quantile function of both distributions, we obtain a confidence interval for the arrival time and location based on prediction uncertainty, which we then combine with the particle size to form the final deflection window. We apply our method to a real sorting task using a pilot-scale chute sorter. Our results obtained from extensive sorting trials show that sorting accuracies can be remarkably improved compared with state-of-the-art industrial sorters and enhanced even further compared with predictive tracking while having the potential to reduce compressed air consumption.
  • Publication
    Comparison of Ethereum Smart Contract Analysis and Verification Methods
    ( 2024)
    Happersberger, Vincent
    ;
    ; ;
    Pignolet, Yvonne Anne
    ;
    Schmid, Stefan
    Ethereum allows to publish and use applications known as smart contracts on its public network. Smart contracts can be costly for users if erroneous. Various security vulnerabilities have occurred in the past and have been exploited causing the loss of billions of dollars. Therefore, it is in the developer’s interest to publish smart contracts that serve their intended purpose only. In this work, we study different approaches to verify if Ethereum smart contracts behave as intended and how to detect possible vulnerabilities. To this end, we compare and evaluate, different formal verification tools and tools to automatically detect vulnerabilities. Our empirical comparison of 140 smart contracts with known vulnerabilities shows that different tools vary in their success to identify issues with smart contracts. In general, we find that automated analysis tools often miss vulnerabilities, while formal verifiers based on model checking with Hoare-style source code annotations require high effort and knowledge to discover possible weaknesses. Specifically, some vulnerabilities (e.g., related to bad randomness) are not detected by any of the tools. Formal verifiers perform better than automated analysis tools as they detect more vulnerabilities and are more reliable. One of the automated analysis tools was able to find only three out of 16 Access Control vulnerabilities. On the contrary, formal verifiers have a hundred percent detection rate for selected tests. As a case study with a smart contract without previously known vulnerabilities and for a more in-depth evaluation, we examine a smart contract using a two-phase commit protocol mechanism which is key in many smart contract applications. We use the presented tools to analyze and verify the contract. Thereby we come across different important patterns to detect vulnerabilities e.g. with respect to re-entrancy, and how to annotate a contract to prove that intended the restriction and requirements hold at any time.
  • Publication
    Das Lieferkettengesetz - Technologien für mehr Transparent in der Supply Chain
    ( 2023-05-09) ;
    Schreynemackers, Pia
    Stellenwert von Nachhaltigkeit in Lieferketten. Hat sich die Bedeutung der Nachhaltigkeit in Ihrer Lieferkette durch die aktuelle Lage (Pandemie, Ukrainekrieg, Lieferkettenstörungen, höhere Gaspreise etc.) verändert? 94% der Teilnehmer geben an, dass der Stellenwert von Nachhaltigkeit in der Lieferkette auch durch aktuelle Krisen nicht nachgelassen hat. Blockchain ermöglicht eine bessere Sichtbarkeit der Lieferkette und eine neue Stufe der Datenintegrität. Künstliche Intelligenz hat die Fähigkeit Unmengen an unsortierten Daten in Echtzeit zu analysieren. Machine learning und KI ermöglichen eine 360° Bewertung der Lieferkette als ein Baustein des LkSG.