An Introduction to Reinforcement Learning - Snake Your Way out of a Paper Bag

Learn reinforcement learning fundamentals through practical C++ implementation in this 53-minute conference talk from C++ on Sea. Explore how reinforcement learning differs from other AI approaches like LLMs, understanding how an "agent" takes actions in dynamic environments where better outcomes receive rewards and worse outcomes face penalties. Discover real-world applications including targeted marketing, infrastructure optimization, and financial decision-making, while examining how RL has been successfully applied to games like Go and PacMan. Follow along as the speaker implements the classic Snake game and demonstrates how machines can learn to play through vanilla Q-learning using table-based data structures. Start with basic C++ concepts using std::vector and progress through probabilistic move selection and random number generation. Gain hands-on experience with tracking outcomes and reward systems while building a solid foundation in both machine learning concepts and C++ programming techniques. Understand the fundamentals that can later be extended to deep Q-learning with neural networks, emphasizing the educational value of implementing algorithms from scratch to master both the programming language and the underlying concepts.