Groups
Temporal Difference (TD) Learning updates value estimates by bootstrapping from the next state's current estimate, enabling fast, online learning.
The policy gradient theorem tells us how to push a stochastic policy’s parameters to increase expected return by following the gradient of expected rewards.