This code contains a very basic introduction to Reinforcement learning using TFJS. The goal is for this to be interactive:
- where the player will choose rock, paper or scizzors.
- then the agent will choose what it thinks is the best action
- then the player will tell the agent if it did bady or poorly
- the agent will then update its ideas on which actions are good and bad
The choice of actions are visualied using plotly.js and we can see them change over time.
open index.html in a browser