Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation