#ai
#computing
An episode is a sequence of steps which includes a sequence of states and actions in the world
$\tau = (s_0, a_0, s_1, a_1, ...).$
Transitions happens
They are deterministic
$s_{t+1} = f(s_t, a_t)$
Or stochastic
$s_{t+1} \sim P(\cdot|s_t, a_t).$
![[n-step bootstrapping]]