A dual-memory architecture for reinforcement learning on neuromorphic platforms

Reinforcement learning (RL) is a foundation of in biological systems and provides framework to address numerous challenges with real-world artificial intelligence applications. Efficient implementations RL techniques could allow for agents deployed edge-use cases gain novel abilities, such as improved navigation, understanding complex situations critical decision making. Towards this goal, we describe flexible architecture carry out reinforcement on neuromorphic platforms. This was implemented using an Intel processor demonstrated solving variety tasks spiking dynamics. Our study proposes usable energy efficient solution applications demonstrates applicability the platforms problems.