A supervised learning algorithm is used when we have a labeled training dataset. Reinforcement learning is used in a scenario where an agent interacts with an environment to observe a basic behavior and change its state to maximize its rewards or goal.
Other differences between them have been given in the following table:
|
Criteria |
Supervised Learning |
Reinforcement Learning |
|
Example |
Digit recognition. |
Chess game. |
|
Works on |
Given labeled dataset. |
Interacting with the given environment. |
|
Decision |
A decision is made on the input given at the beginning. |
Here, the algorithm helps make a decision linearly. |
One of the essential features of RL is that the agent's action may not affect the immediate state of the environment that it is working in, but it can affect the subsequent states. Hence, the algorithm might not learn in the initial state but can learn after a few states are changed.