Training intelligent agents in reinforcement learning (RL) is much more
time-consuming than animal learning. This is because agents learn from scratch,
but animals learn with genes inherited from ancestors and are born with some
innate abilities. Inspired by genes in animals, here we c