This monograph takes an alternative approach to RL that is different from classic textbooks. Rather than focusing on tabular problems, RL as a generalization of supervised learning is introduced, which is first applied to non-differentiable objectives and later to temporal problems.
Read More
This monograph takes an alternative approach to RL that is different from classic textbooks. Rather than focusing on tabular problems, RL as a generalization of supervised learning is introduced, which is first applied to non-differentiable objectives and later to temporal problems.
Read Less