Are Gradient-based Saliency Maps Useful in Deep Reinforcement Learning?

Understand

Deep Reinforcement Learning (DRL) connects the classic Reinforcement Learning algorithms with Deep Neural Networks.

• A problem in DRL is that CNNs are black-boxes and it is hard to understand the decision-making process of agents.
• In order to be able to use RL agents in highly dangerous environments for humans and machines, the developer needs a debugging tool to assure that the agent does what is expected.
• Currently, rewards are primarily used to interpret how well an agent is learning.