  toggle dark mode toggle dark mode

Play tic-tac-toe against a reinforcement agent

posted on 2020-05-30T11:27:03Z · last modified on 2020-06-10T11:53:21Z · view page on GitHub
tags: ml (7)machine learning rl (3)reinforcement learning game (1) javascript (1)

I finally made an unbeatable tic-tac-toe game in javascript using the qtable obtained in in the first reinforcement learning post. Have Fun going for the tie!


Player O
Player X

Player O's turn


  • part 1: We create the game environment and a simple unbeatable AI based on traditional Q-learning 🤖.
  • part 2: We modify our AI to utilize a neural network: deep Q-learning 👾.
  • part 3: Have some fun and play against the Q-agent 🤓.

If you like this post, consider leaving a comment or star it on GitHub.