Designing a Video Game AI
summary
Reinforcement learning is used to train a game-playing AI by defining a scoring objective rather than coding strategy rules. The speaker describes training an agent for several months to play a puck-and-board game (metal puck on wooden/lead board) against a human, explains how an agent learns, presents three different approaches used, and discusses the pitfall of self-imitation, plus concrete demos and practical advice for developers starting RL game projects.
AI-written summary · about
related talks
24mDesigning a Video Game AI with Reinforcement Learning
56mCan I do AI?
01h 01mIntroduction to Practical AI
39mBuilding self evolving video game AI in Python
30mReinforcement Learning - ChatGPT, Playing Games & More
15mAI Self - Learning Game Playing
22mDeveloping AI in Racing Games Using Reinforcement Learning
13mReinforcement learning and it's growing role in AI