Question

A part of Machine Learning where an agent learns to behave optimally, by performing actions that will either be rewarded or punished

Answer 1

• A part of Machine Learning where an agent learns to behave optimally, by performing actions that will either be rewarded or punished is called Reinforcement learning.
• Reinforcement learning is an important algorithm in Machine Learning.
• Reinforcement learning helps in developing a better model when there is no training dataset and the learning has to happen from experience. This way it is different from supervised machine learning.
• The main components for reinforcement learning are input, output and training.
• Input will be the initial state that the model will start from. Output will have multiple solutions to a specific problem
• During training the model will return an output and the user will decide whether to reward or punish, thereby teaching the model to continuously learn.
• The final best output with maximum reward is chosen as the best solution.