New project: Explainable AI in rail transport

Can an autonomous train recognize with sufficient certainty whether a deviation in the camera image is caused by a shadow or perhaps by a child playing on the tracks? And how can humans be sure during safety assessment that the system is not falsely overconfident? How can they ensure that their safety assessment is sufficiently reliable - especially when the technology of the AI system makes it impossible in principle to provide the relevant evidence?

In the XRAISE (Explainable AI for Railway Safety Evaluations) project, we are investigating how explainable artificial intelligence can be used to improve the safety evaluation of AI algorithms. The project is funded 2024-2026 by the Federal Railway Authority and is carried out in cooperation with AI and railway experts from research and practice. 

Contact Romy Müller