Intuitive RL: Intro to Advantage-Actor-Critic (A2C) | HackerNoon

Abstraction: Intuitive narrative introduction to the Advantage-Actor-Critic RL algorithm

Key points:

• A2C (Advantage Actor-Critic) is a policy gradient method; understanding it is presented as sufficient to understand deep RL broadly
• Uses storytelling rather than pure math to explain RL concepts, motivated by RL's origins in natural learning systems
• References a simple code implementation and an industrial-strength PyTorch A2C based on OpenAI's TensorFlow Baselines
• Recommends Barto and Sutton's Introduction to RL and David Silver's UCL course for deeper study
• Article is introductory; the main technical content is in linked code notebooks rather than the article text itself

Connections: Openai · Actor Critic · Reinforcement Learning · Policy Gradient

Source: https://hackernoon.com/intuitive-rl-intro-to-advantage-actor-critic-a2c-4ff545978752