Fetching the paper…

Saute RL: Almost Surely Safe Reinforcement Learning Using State Augmentation · Around