Start a project

Fig. I — A run, in order

Reinforcement Learning

4 parts, read in order — each one assumes the one before it. Back toall runs orthe full archive.

Start at the beginning

The parts

Supervised learning felt intuitive to me from day one: here's the input, here's the right answer, minimize the difference.

  1. Reinforcement Learning from Zero: Agents, Rewards, and the Loop That Learns from Consequences
  2. DQN: The Moment Reinforcement Learning Met Deep Learning
  3. Policy Gradients and PPO: Learning Behavior Directly
  4. RLHF and Beyond: How Reinforcement Learning Taught Language Models to Behave