Dr. André Ofner

Managed the the day-to-day operations of the decision making AI team working on Level 4 autonomous driving (full autonomy within operational design domain). Guided the development, validation and deployment of reinforcement learning and active inference models for real-time interactive planning and rule aware decision making. Led the design of a SOTIF compatible safety architecture and simulation environment. Led the internal research areas focused on active inference & decision-making AI.

Conducted research on per-design explainable AI, developing cognitively inspired and inherently interpretable model architectures under the Free Energy Principle. Designed and evaluated predictive coding, active inference and generative models based on iterative inference and biologically plausible local learning rules. Authored and co-authored peer-reviewed publications and presented results at international conferences and workshops.

Researched cognitive-science-grounded machine learning models, linking human perceptual processes to variational inference and predictive coding. Developed and analysed predictive processing and brain-inspired ML models, including applications to audio, neural and behavioural data.