Reinforcement learning (RL) has recently emerged as a core methodology for modelling and analysing complex adaptive systems. Ensuring that learning agents behave safely, robustly, and transparently is ...