Hacking Reinforcement Learning
summary
Presents a family of planning algorithms that produce high-quality reinforcement-learning datasets and can achieve superhuman scores with low compute, demonstrated by integrating them with Python RL libraries and quickly generating Ms. Pac-Man games. Provides an intuitive, non-mathematical explanation with visual animations and a four-step "hack" workflow: information gathering, scanning/detecting vulnerabilities (the new algorithms and integration), exploitation (Python proof-of-concept and library examples), and maintenance/evaluation with results and resources.
AI-written summary · about
related talks
01h 31mReinforcement Learning- Pac-Man
31mDesigning game AI with reinforcement learning
01h 01mReinforcement Learning: Pac-Man
28mDeep Reinforcement Learning: theory, intuition, code
30mReinforcement Learning - ChatGPT, Playing Games & More
29mReinforcement Learning Without a PhD: A Python Developer’s Journey
14mHuman-in-the-Loop Reinforcement Learning
02mDesigning a Video Game AI