with details please

Which of the following problems cannot be cast into reinforcement learning? Select one: Finding the way out of a maze. Delivering packages from point A to point B. Grouping students according to their grades into 4 categories. Identifying spam emails based on the email itself as well as the history of spam emails for each user