Deep Reinforcement Learning

Intro

Deep Reinforcement Learning (DRL) combines Reinforcement Learning (learning via trial-and-error rewards) with Deep Learning to enable agents to learn optimal behaviors directly from high-dimensional inputs such as images, video, or sensor data.

Resources

• Using RL to teach robots to fly a drone

Deep RL Methods

• Proximal Policy Optimization (PPO)