Understanding, trust and verification of machine learning
The V3ML project (Verstehen, Vertrauen und Verifikation von Maschinellem Lernen) is an interdisciplinary research initiative that aims to make machine learning (ML) systems more intuitive, trustworthy, and provably reliable. It brings together researchers in computer science and psychology from the University Alliance Ruhr – TU Dortmund, Ruhr University Bochum, and University of Duisburg-Essen.
To evaluate how people interact with and trust ML systems in practice, we use a driving simulator that replicates real-world scenarios involving automated decision-making. This allows us to study trust dynamics in a controlled, yet realistic, setting — helping us design systems that are not only technically sound but also perceived as reliable by users.
Funding
Research Center Trustwothy Data Science and Security