COMPUTER-SCIENCE REINFORCEMENT LEARNING GOT REWARDS WRONG

Summary from the TLDR AI newsletter

In reinforcement learning, the reward is part of the agent, not part of the environment. Once we start thinking in terms of the reward translation as being part of the agent, we can also start thinking of ways in which the agent can affect the reward translation mechanism. It could be dynamic and change across time, be tied to policy, or even learned. The shift in perspective creates many possible options, just by changing the notation - because notation matters.