Objective: The Neurorobotics Lab develops machine learning techniques for robotics and neurotechnology. The lab focuses on data-efficient deep reinforcement learning, spanning offline and inverse RL, robust control, and the integration of model predictive control with learning, together with convex-optimization approaches that make these methods more reliable and interpretable.  We carry these foundations through to safety-critical applications such as high-level decision making for autonomous driving and humanoid robot control, and to machine learning for medicine and neurotechnology.

Research Areas: Foundations of Reinforcement Learning, Offline and Inverse Reinforcement Learning, Robust and Safe Control, Model Predictive Control and Learning, High-Level Decision Making for Autonomous Driving, Behaviour Modelling and Neural Decoding, Machine Learning for Medicine and Neurotechnology