Logo des Repositoriums

Auflistung nach Schlagwort "explainable artificial intelligence"

1 - 2 von 2
  • Konferenzbeitrag
    The Impact of Explanation Detail in Advanced Driver Assistance Systems: User Experience, Acceptance, and Age-related Effects
    (Mensch und Computer 2023 - Tagungsband, 2023) Hermann, Julia; Nierobisch, Niels; Arndt, Robin; Kubullek, Ann-Kathrin; van Ledden, Sebastian; Dogangün, Aysegül
    User understanding and confidence are critical in the context of advanced intelligent driver assistance systems (ADAS) to ensure the desired response and prevent manual countersteering during automated maneuvers. However, the interventions of advanced ADAS can sometimes be unexpected and disruptive to drivers, especially when the reasons are unclear. In our study, we investigated the effects of differently presented explanations provided by a driver assistance system. We presented participants with three scenarios from the driver’s perspective and created two videos for each scenario with explanations of varying detail. Participants were asked to answer two questionnaires following each video. The results show that more detailed explanations generally lead to a better user experience and higher confidence in the system’s performance. We also discuss the possible influence of age and technology acceptance in our article.
  • Textdokument
    Towards Designing a User-centric Decision Support System for Predictive Maintenance in SMEs
    (INFORMATIK 2021, 2021) Kellner, Domenic; Lowin, Maximilian; von Zahn, Moritz; Chen, Johannes
    In manufacturing, small and medium-sized enterprises (SMEs) face global competition. In the field of predictive maintenance (PdM), artificial intelligence (AI) helps to prevent machine failures and has the potential to significantly reduce costs and increase process efficiency. Even though PdM has several benefits, it also entails considerable challenges for SMEs, especially when it comes to user interactions. In this short paper, we harness the design science methodology and discuss several problems regarding user interactions with predictive maintenance applications. We incorporate two different literature streams, namely, predictive maintenance and decision support systems. Finally, we present necessary design requirements, principles, features, and propose a research design to further develop and evaluate a user-centric PdM decision support system. Thereby, we contribute to making AI tangible in SMEs.