$.$ Which of the following statements is true about applying function approximation in reinforcement learning? Select all that apply.

a$.$ It can be more memory efficient.

b$.$ Linear function approximation can be applied for any type of function accurately.

c$.$ It allows generalization from the learned states.

d$.$ It can help the agent achieve good discrimination.

e$.$ The method results in higher dimensions of the feature space compared to the state space dimension for better resolution.