Encouraging the model not to do something with positive rewards, over time it will learn to avoid it in order to maximise its rewards. This setting is known as reinforcement learning
Articles Related
Plugin combo - Component related: Nothing was found.