Deep Reinforcement Learning Overview

Deep RL uses neural networks to approximate policies or value functions in complex spaces.

What does 'deep' reinforcement learning add to standard RL?