P326 - INFORMATIK 2022 - Informatik in den Naturwissenschaften
Auflistung P326 - INFORMATIK 2022 - Informatik in den Naturwissenschaften nach Erscheinungsdatum
1 - 10 von 146
Treffer pro Seite
Sortieroptionen
- TextdokumentHigh-Performance and Quantum Computing for Students(INFORMATIK 2022, 2022) Faber,Peter; Liebelt,HelenaMany technical advances have been applied in the area High-Performance Computing for the first time, and this also seems to hold for Quantum Computing. Traditionally, however, experts in these fields come from a diverse educational background. Recently, Technische Hochschule Deggendorf (Deggendorf Institute of Technology, DIT) introduced a new master's program "`High-Performance Computing / Quantum Computing"' to offer students a direct way into the application specific aspects of these field of expertise. To our knowledge, this is the first course of studies of its kind. This work describes this new program of studies and tentatively identifies some of the first lessons learned so far.
- TextdokumentService Tailoring: Ein Konfigurationsmechanismus für individualisierte After-Sales-Services im Maschinen- und Anlagenbau(INFORMATIK 2022, 2022) Brinker,Jonas; Kammler,Friedemann; Thomas,OliverAußerplanmäßige Instandsetzungen sind eine Herausforderung im Maschinen- und Anlagenbau. Dies betrifft sowohl Nutzer der Maschinen, die aufgrund der Ausfallzeit Abweichungen von ihren Produktionsplänen hinnehmen, aber auch Hersteller, die Serviceressourcen kurzfristig und dynamisch zuordnen müssen. Der Einsatz technischer Innovationen eröffnet dabei neue Serviceoptionen, die beispielsweise vorteilhaft hinsichtlich Zeitaufwands, Kosten oder Nachhaltigkeit sind. Gleichzeitig variieren die kundenindividuellen Anforderungen an die Serviceerbringung, sowohl hinsichtlich der benötigten Ressourcen als auch hinsichtlich individueller Präferenzen. Mit dem „Service Tailoring“ (ST) stellen wir einen Mechanismus vor, wie Serviceoptionen kundenspezifisch konfiguriert werden können. Aus elf Experteninterviews erheben wir zunächst Prozessvarianten in der Instandhaltung und bilden einen generalisierten Lösungsraum. Indem wir Mechanismen der Mass-Customization (MC) als Constraint-Optimization-Problem adaptieren, entwickeln wir einen zweistufigen methodischen Vorschlag und demonstrieren diesen an einem Beispiel aus der Landtechnik.
- TextdokumentMeasuring Gender Bias in German Language Generation(INFORMATIK 2022, 2022) Kraft,Angelie; Zorn,Hans-Peter; Fecht,Pascal; Simon,Judith; Biemann,Chris; Usbeck,RicardoMost existing methods to measure social bias in natural language generation are specified for English language models. In this work, we developed a German regard classifier based on a newly crowd-sourced dataset. Our model meets the test set accuracy of the original English version. With the classifier, we measured binary gender bias in two large language models. The results indicate a positive bias toward female subjects for a German version of GPT-2 and similar tendencies for GPT-3. Yet, upon qualitative analysis, we found that positive regard partly corresponds to sexist stereotypes. Our findings suggest that the regard classifier should not be used as a single measure but, instead, combined with more qualitative analyses.
- TextdokumentGAFAI: Proposal of a Generalized Audit Framework for AI(INFORMATIK 2022, 2022) Markert,Thora; Langer,Fabian; Danos,VasiliosML based AI applications are increasingly used in various fields and domains. Despite the enormous and promising capabilities of ML, the inherent lack of robustness, explainability and transparency limits the potential use cases of AI systems. In particular, within every safety or security critical area, such limitations require risk considerations and audits to be compliant with the prevailing safety and security demands. Unfortunately, existing standards and audit schemes do not completely cover the ML specific issues and lead to challenging or incomplete mapping of the ML functionality to the existing methodologies. Thus, we propose a generalized audit framework for ML based AI applications (GAFAI) as an anticipation and assistance to achieve auditability. This conceptual risk and requirement driven approach based on sets of generalized requirements and their corresponding application specific refinements as contributes to close the gaps in auditing AI.
- TextdokumentYes she can! How to encourage girls to choose a career path in computer science(INFORMATIK 2022, 2022) Dodiya,Janki; Heinisch,Cornelia; Hess,Claudia; Kunz,Sibylle; Rebholz,Sandra; Schlömer,Inga; Steffen,Adrienne; Vaas,SilkeWomen are still underrepresented in studies and in job positions in the fields of STEM. This paper proposes a study and a lecture concept as measures to increase the number of women in Computer Science careers. The study aims to better understand barriers, drivers and influencing factors that affect young women's decision for or against a career in computer science. A lecture series on computer science is proposed, specifically designed for girls and young women to reduce fears and prejudices and increase the visibility of women in IT. This will also provide further research data that will enable adjusting to the requirements of the target group based on the research outcomes. With this article, the authors would like to initiate discussion on how universities can better gear their offerings to a female target group.
- TextdokumentConcept, Implementation and Evaluation of a Virtual Learning Environment for Acquiring Competences in Android App Development(INFORMATIK 2022, 2022) Neumann,Frank; Rodríguez-del-Pino,Juan Carlos; Homer,SebastianThe presented project provides a virtual learning environment for university courses teaching app development using Android and Kotlin. It features both self-guided lessons and associated tests as well as practical app development tasks and associated self-assessed programming exercises. The well-known Moodle plug-in Virtual Programming Lab (VPL) serves as a starting point for the latter part. VPL gets extended by means for the usage of the Gradle build tool, the Android SDK tools and AndroidX tests for unit tests at the level of Android activities and fragments. In particular, the rather excessive resource requirements of the Gradle tool created numerous problems in the sandboxes used by VPL. These issues were addressed by configuring suitable VPL settings. In addition, problems associated with the proper usage of these settings had to be addressed in VPL. In order to evaluate whether the virtual learning environment suits the student’s need for learning app development, a questionnaire was designed that contains questions about the course in general and about the design, handling and the benefits of the exercises The answers of the students to this questionnaire confirm that the proposed virtual learning environment provides a suitable means for learning in the described field.
- TextdokumentD-TOUR: Detour-based point of interest detection in privacy-sensitive trajectories(INFORMATIK 2022, 2022) Schneider,Maja; Gehrke,Lukas; Christen,Peter; Rahm,ErhardData collected through mobile sensors on private and commercial devices can give valuable insights into mobility patterns and facilitate applications such as urban planning or traffic forecasting. At the same time, such data can carry immense privacy risks for the data producers. Stop detection approaches can reveal a person's points of interest (POI) by clustering temporal and spatial features, uncovering private attributes such as home or work addresses. Privacy-preserving mechanisms aim at hiding these POIs, for example via speed smoothing approaches that are able to preserve high data utility. We show experimentally on two real-world data sets that trajectories can contain anomalies that are contained to a certain extent when smoothing a route and are not detected by state of the art stop detection algorithms. We propose a novel attack D-TOUR that reveals POIs based on deviations from the optimal route. Our experiments suggest that our proposed attack has similar performance on unprotected data but outperforms the baseline approach, especially when protection is higher and route features become more sparse.
- TextdokumentUnification of Algorithms for Quantification and Unfolding(INFORMATIK 2022, 2022) Bunse,MirkoQuantification is the supervised learning task of predicting the prevalences of classes in a data sample. Physics literature knows the same task under a different name: unfolding. However, the literature on quantification and the literature on unfolding are largely disconnected from each other. We bridge this interdisciplinary gap by proposing a common framework that integrates algorithms from both fields in a unified form. Instantiations of our framework differ from each other in terms of the loss functions, the regularizers, and the feature transformations they employ.
- TextdokumentFairness in Regression -- Analysing a Job Candidates Ranking System(INFORMATIK 2022, 2022) Markert,Karla; Ahouzi,Afrae; Debus,PascalFairness is one of the pillars of any well-functioning society. Recent law-making in the EU regulates the machine-centered approach and thus increases the necessity for certifiable fairness approaches. In this paper, we adapt previous literature on certifiable fairness in classification systems to a regression model for simplified candidates ranking. This model serves as an illustration for an application that should work fairly even if built upon a biased data set. With our synthetic dataset we are able to analyse the challenges of different fairness notions. Although the fairness training manages to improve the certifiable individual fairness, some of the encoded bias remains. We discuss the challenges we faced, including the selection of suitable parameters and the trade off between accuracy and fairness. We hope to encourage more research into fairness improvement and certification, within and beyond group and individual fairness.
- TextdokumentPotential analysis for the detection of attacks on wireless networks using the Wireless Intrusion Detection System Nzyme(INFORMATIK 2022, 2022) Eisenhut,Maximilian; Honekamp,WilfriedDue to the flexibility and low cost of acquisition compared to wired network connections, wireless networks continue to proliferate. Due to this increasing number and the characteristics of a shared medium, it offers potential attackers a suitable platform to easily gain access to diverse network types. To this end, the range of specialised hardware and software for attacking wireless networks is constantly evolving. Information on the location and other parameters of wireless networks is also documented and updated online in a largely automated manner. Particularly in the economic as well as in the public environment, a special need can thus arise to detect attacks, identify attackers and initiate countermeasures on the basis of this information. This paper describes the evaluation of the possibilities offered by the open-source Wireless Intrusion Detection System (WIDS) Nzyme. For this purpose, the messages that occur during different attacks were examined. Furthermore, real data was recorded and evaluated based on the parameters from the test attacks to draw conclusions about the type and frequency of attacks. The ratio between legitimate reports and false alarms was also determined. Test attacks were successfully detected and could be assigned to possible attacks. Real data was recorded at three locations and compared with the patterns from the test attacks. The evaluation shows that the rate of false alarms in real operations is unacceptable, at over 27%. The causes for this are mostly misconfigurations and atmospheric disturbances. The study further shows, that combined alarm messages allow conclusions to be drawn about the type of attack carried out and thus the number of false alarms can be reduced. The effort and benefit of a WIDS are currently not yet in a meaningful relationship. Nevertheless, use and further development are recommended, taking these circumstances into account.