Home▸AI & Machine Learning▸Reinforcement Learning: Q-Learning Gridworld

🤖 Reinforcement Learning: Q-Learning Gridworld

Watch a reinforcement-learning agent explore a gridworld and learn a policy via Q-learning: epsilon-greedy exploration, the Bellman update, and value bars that rise as the agent discovers the shortest path to the goal.

AI & Machine Learning3DModerate60 FPS

⚙ Under the hood

Watch a reinforcement-learning agent explore a gridworld and learn a policy via Q-learning: epsilon-greedy exploration, the Bellman update Q(s,a) ← Q(s,a) + α[r + γ·max Q(s′,a′) − Q(s,a)], and value bars that rise as the agent discovers the shortest path to the goal — the same mechanism behind DQN, PPO, Actor-Critic and RLHF fine-tuning of LLMs.

Three.jsAIReinforcement LearningQ-LearningMachine LearningMDP

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install