Computational AI validation: How to use realistic environment and sensor simulation to make perception in autonomous driving safe

Techniques to capture and render models of the real world have been matured significantly over the last decades. We are now able to synthesize virtual scenes in a visual quality that is hard to distinguish from real photographs for human observers. On the other hand, we have emerging complex technical and particular autonomous systems sensing the real world and aiming at resembling some perceptional tasks formerly only feasible by humans. Because of the complexity of reality and the related increasing complexity of the technical tasks, validation, that makes sure these systems work as intended and are safe are increasingly important. Because of the progress in visual and multi-sensor synthesis, now building systems for validation of these complex systems in the data center becomes not only feasible, but also offers more possibilities for the integration of intelligent techniques in the engineering process of complex applications. One prominent example we focus on in in this context is the validation of automated driving.

Short biography

Oliver Grau works for Intel Labs, Intel’s R&D department in Germany. He was the Intel co-director of the Intel-Visual Computing Institute and now leads research on validation of AI systems for automated driving. Previously he worked as a Lead Technologist for BBC R&D in London, UK on computer vision projects for innovative media production systems.

Oliver published more than 70 scientific publications and holds more than 10 patents. He got a PhD from Hannover University, is a recognized expert on computer vision and visual computing in the research community and served on many scientific program committees. He is a member of the ACM Europe Council, chair of the ACM Europe Technology Policy Committee and a visiting professor of Surrey University, UK.