Simultaneous learning from both equal and explicit preferences enables preference-based reinforcement learning agents to better understand human feedback, leading to improved feedback efficiency and task performance.
