Socially-Legible Reinforcement Learning

Aligning RL policies with pedestrian social conventions