#ai #computing # Society of RL [[Society of Minds]] argue that our mind is made of many minds, each accomplishing a task. Nature is highly resilient and distributed, why don't we built AIs that is composed of many building blocks of reinforcement learning agents having themselves dense, superficial rewards ? Task: drink a cup of tea - Some agent does control (including many agents for the arm, hand, lips ...) - Some agent does vision - Some agent does touch (temperature check) You can see that as a tree, the higher, the sparser the reward [[Hierarchical Reinforcement Learning]] ![[Society of Minds.png]] Similar topic links: [[Hierarchical Reinforcement Learning]] [[Agency]] [[Humans are in a Red Queen competition with artificial intelligence]] [[Self supervised reinforcement learning]] [[Current bottlenecks of artificial intelligence]]