Skip to main content
OnePocketLab
Sign in
onepocketML
Advanced

Reinforcement Learning

Learn by doing: how agents learn from reward and consequence.

A different kind of learning, an agent that takes actions in an environment and learns which choices pay off, from grid-world basics to the ideas behind game-playing AI.

6 of 6 lessons available

NO ANSWER KEY ANYWHEREAGENTpicks a moveWORLDreacts to itactionstate + reward+1ONLY A SCORE, AFTER THE FACT
The one family with no answer key. An action goes out, and a state and a single number come back.

Lessons in this path

  1. Reinforcement Learning
  2. Markov Decision Processes
  3. Exploration vs. Exploitation
  4. Q-Learning in Depth
  5. Deep Q-Networks (DQN)
  6. Policy Gradient Methods
Reinforcement Learning | OnePocketML