AI Reinforcement Learning - eLearning

Course timeline

1. Introduction to Reinforcement Learning 

   Lesson 01

   • Three Paradigms of Machine Learning
   • RL Success Stories
   • Elements of an RL Problem
   • Introduction to Gym
   • Training Your First RL Agent Using RLlib

2. Single-Step RL: Multi-Armed Bandits 

   Lesson 02

   • Multi-Armed Bandit Setting
   • Exploration-Exploitation Trade-Off
   • Fundamental Approaches To Trade Off Exploration and Exploitation
   • Advanced Approaches To Trade Off Exploration and Exploitation
   • Introduction to Contextual Bandit Problems
   • A Practical Contextual Bandit Example
   • Deep Contextual Bandits
   • Exploration With Deep Contextual Bandits
   • A Practical Example With Deep Contextual Bandits

3. Multi-Step Reinforcement Learning 

   Lesson 03

   • Introducing Markov Chains
   • Markov Reward Process
   • Markov Decision Process
   • Policy Evaluation and Iteration
   • Tabular Q-Learning
   • Practical Tabular Q-Learning Example
   • Deep Q-Learning
   • Using RLlib To Train a Deep Q Network
   • Policy-Based Methods
   • Using RLib To Train PPO Agent

4. Approaches for Real-World Reinforcement Learning

   Lesson 04

   • Handling Sparse Rewards and Hard Exploration
   • Implement Reward Shaping
   • Disadvantages of Reward Shaping
   • Using Memory To Handle Partial Observability
   • Solving Stateless Cartpole Using LSTM
   • Overcoming Sim-to-Real Gap
   • Introduction to Multi-Agent Reinforcement Learning
   • Training Multiple Agents Using RLib
   • Multi-Agent Reinforcement Learning
   • Offline Reinforcement Learning
   • Conclusion and Other Advanced Topics