# Model
A **model** is a reresentation that predicts what will happen inside an environment. This knowledge about the environment and/or task is explicitly encoded, such that instead of directly learning a policy, we learn a model function which helps improve our policy.

Most commonly, they can take the following forms:

1. State Transition model: ![](https://latex.codecogs.com/svg.latex?s_{t+1}=f(s_t,a_t))
2. A Reward model: ![](https://latex.codecogs.com/svg.latex?r_{t+1}=f(s_t,a_t))