Modern Reinforcement Learning: Actor-Critic Algorithms

Course Overview

Skills and Knowledge:
By taking this course, students will acquire the skills and knowledge to implement policy gradient, actor critic, deep deterministic policy gradient (DDPG), twin delayed deep deterministic policy gradient (TD3), and soft actor critic (SAC) algorithms in a variety of challenging environments from the Open AIgym. Additionally, students will gain a repeatable framework for quickly implementing the algorithms in advanced research papers. Furthermore, students will gain an understanding of the fundamentals of reinforcement learning, including topics such as the Bellman Equation, Markov Decision Processes, Monte Carlo Prediction, Monte Carlo Control, Temporal Difference Prediction TD(0), and Temporal Difference Control with QLearning. Finally, students will gain an understanding of how to read deep reinforcement learning research papers and how to address the explore-exploit dilemma with a deterministic policy, as well as how to deal with overestimation bias and approximation errors in deep neural networks.
Professional Growth:
This course contributes to professional growth by providing students with the knowledge and skills to read and implement deep reinforcement learning research papers. By mastering the content in this course, students will be able to quickly implement advanced algorithms such as policy gradient, actor critic, deep deterministic policy gradient (DDPG), twin delayed deep deterministic policy gradient (TD3), and soft actor critic (SAC) in a variety of challenging environments. This will give them a quantum leap in their capabilities as an artificial intelligence engineer, and will put them in a league of their own among students who are reliant on others to break down complex ideas for them.
Further Education:
This course is suitable for preparing further education, as it covers advanced topics in deep reinforcement learning and provides a repeatable framework for quickly implementing algorithms in research papers. It also reviews core topics such as the Bellman Equation, Markov Decision Processes, Monte Carlo Prediction, Monte Carlo Control, Temporal Difference Prediction TD(0), and Temporal Difference Control with QLearning. Furthermore, it covers topics such as the REINFORCE algorithm, Deep Deterministic Policy Gradients (DDPG), Twin Delayed Deep Deterministic Policy Gradients (TD3), and Soft Actor Critic (SAC). This course is designed for highly motivated and advanced students who have prior course work in college level calculus, reinforcement learning, and deep learning.