An Intuition-Based Approach to Reinforcement Learning

Explore an intuition-based approach to reinforcement learning in this 42-minute talk by Oswald Campesato, co-founding CEO of iQuarkt and author of over 35 technical books. Gain insights into a framework that helps algorithms learn decision-making through environmental feedback, inspired by human and animal intuition. Delve into key concepts such as exploit versus explore, greedy versus epsilon-greedy strategies, discount factors, and reward calculations. Examine practical applications in game playing, robotic control, and autonomous driving. Learn about Q-tables, TD learning versus Monte Carlo methods, and the transition from DRA to MDP. Discover the OpenAI CartPole environment and gain access to useful resources for further exploration of reinforcement learning techniques.

- Introductions
- What is the Goal
- Exploit Versus Explore
- Greedy Versus Epsilon-Greedy
- Discount Factor “g”
- Calculating Rewards
- Pseudo Code
- Working With Q-Tables 1
- Working With Q-Tables 2
- Online Q-Table
- States & Actions
- TD Learning vs Monte Carlo
- From DRA to MDP
- Stochastic Actions
- OpenAI CartPole
- More Terminology
- Useful Links