Back to course

The Studio

AI Studio

One lab. Ten days. Every app is an AI capability you build yourself.

7 / 7 apps unlocked

Game Master

Reinforcement learningDay 09
Explore the lesson

Two agents you train on Day 9 with reinforcement learning — no examples, no labels, just rewards.

FrozenLake

Drag to orbit · scroll to zoom
Move 0 / 6Paused · ready to explore

A path through the ice.

A tiny brain: 16 states × 4 actions. Follow the agent’s six-move route, or select a tile to explore its place in the environment.

Agent state00
Q-table size16 × 4

Selected state 00

Start · safe ice

Four available actions
Safe iceHoleAgent route