deep reinforcement learning is an increasingly popular technique for
synthesising policies to control an agent's interaction with its environment.
There is also growing interest in formally verifying that such policies are
correct and execute safely. Progress has been made in this area