Guest
Discover more about data collaboration at In this talk, you'll learn about “An Intuition-Based Approach to Reinforcement Learning”. A framework that helps algorithms learn how to make decisions by receiving feedback from their environment. It's based on the idea that humans and animals use intuition to make decisions, and this process can be used to create more efficient and effective learning algorithms. By emphasizing simpler, more intuitive strategies, this approach has been applied to game playing, robotic control, and autonomous driving, and shows promise in enabling more efficient and effective learning compared to traditional reinforcement learning methods. Oswald Campesato is the co-founding CEO of iQuarkt and the author of more than 35 technical books. He has 20 years of experience as a software developer and has worked in numerous management roles. Campesato recently completed books about TensorFlow 2/Keras and Angular/machine learning and he's currently working on an NLP/machine learning book. He has designed a unique curriculum for new natural language processing and deep reinforcement learning courses and also teaches machine learning and deep learning courses. Key Topics: 0:00 - Introductions 3:40 - What is the Goal 5:59 - Exploit Versus Explore 8:18 - Greedy Versus Epsilon-Greedy 9:05 - Discount Factor (“g”) 10:50 - Calculating Rewards 16:00 - Pseudo Code 18:51 - Working With Q-Tables (1) 20:38 - Working With Q-Tables (2) 21:41 - Online Q-Table 22:15 - States & Actions 27:51 - TD Learning vs Monte Carlo 29:05 - From DRA to MDP 30:45 - Stochastic Actions 32:18 - OpenAI CartPole 36:40 - More Terminology 38:20 - Useful Links Related Blogs: Churn Prevention with Reinforcement Learning: Pieter Abbeel on Deep Reinforcement Learning: 7 Reinforcement Learning Use Cases in 2022: Reinforcement Learning for Anyone: Open AI Gym and Ray: Follow ODSC on: LinkedIn - Twitter - Facebook -
