Towards Human-Centered Paradigms in Verification and Validation Assessments (Artikel)

Aus Kompetenzportal
Wechseln zu: Navigation, Suche
Towards Human-Centered Paradigms in Verification and Validation Assessments (Artikel)
Autor Ekaterina Auer, Wolfram Luther
In: Hajian, A., Luther, W., Han Vinck, A. J. (Eds.): Collaborative Technologies and Data Science in Smart City Applications.
Erscheinungsjahr 2018
Seitenzahl 68-81
Review peer

In this paper, we aim at widening the focus of the scientific computations community from one of reliable numerical algorithms and efficient implementations using standardized arithmetic to a broad user-centered system modeling approach and validation design. Modern responsive applications generate huge amounts of heterogeneous input and output data and usually exhibit high complexity. Dealing with these challenges requires novel visual and collaborative approaches to interpret the results, which has to be accounted for in the general V&V procedure. We propose to use reliable visual analytics (VA) as a solution to this task. This analysis needs to go along with an assessment of (meta-)data and code quality, with a choice of adequate data types, with selection of methods to propagate and bound uncertainty and, lastly, with formally rigorous validation efforts. Collaborative outcome analytics done by various stakeholders with multiple expertise should be a precondition for system model evaluation, problem solving and corresponding follow-up actions. We illustrate the general ideas with the help of use cases implementing relevant parts of the proposed enhanced V&V assessment.