Category: Game AI with Reinforcement Learning
-
Advanced Game AI: Multi-Agent RL, Curriculum Learning, and Self-Play
Multi-agent RL, self-play, curriculum learning, and population training for competitive game AI. When one agent isn't enough.
-
Policy Gradient Methods: PPO and A3C for Complex Game Environments
Why DQN fails for continuous control, how actor-critic methods reduce variance, and why PPO became the de facto standard for game AI and robotics.
-
Deep Q-Networks (DQN): Training AI to Play Atari Games
Build a Deep Q-Network from scratch to master Atari games. Covers experience replay, target networks, Double DQN, and practical training tips.
-
Q-Learning for Grid Worlds: Building Your First Game AI Agent
Build a functional game AI with just the Bellman equation and a dictionary. No neural networks, no gradients โ pure Q-learning for grid worlds.
-
Game AI with Reinforcement Learning: Why RL Beats Traditional Methods
Why reinforcement learning beats rule-based game AI: the exploration-exploitation tradeoff, credit assignment, and when to use RL vs traditional methods.