Reinforcement Learning

Catalog description

Reinforcement learning (RL) is widely used in many engineering disciplines, such as control systems, autonomous driving, robotics, and optimization. It emphasizes learning by an agent from direct interaction with its environment, without relying on supervision. This course offers an introduction to RL, a powerful framework for sequential decision-making under uncertainty and a cornerstone of modern approaches to autonomous agents. Emphasizing both the computational and statistical challenges unique to RL, the course guides students through essential techniques and tools required to build effective RL systems. Student will understand how to model a sequential decision-making problem as a Markov Decision Process (MDP) and develop a strong conceptual understanding of the most widely used RL methods and gain practical skills for addressing the dynamic and uncertain nature of real-world environments with the flavor of robot navigation application in a standard environment. Prerequisite(s): (ENGR 15100 FOR LEVEL UG WITH MIN. GRADE OF D- OR ECE 15200 FOR LEVEL UG WITH MIN. GRADE OF D-) AND MA 26100 FOR LEVEL UG WITH MIN. GRADE OF D- AND MA 26500 FOR LEVEL UG WITH MIN. GRADE OF D- AND ECE 30200 FOR LEVEL UG WITH MIN. GRADE OF D- Course Learning Outcomes 1. Introduce the fundamentals of reinforcement learning (RL) as a framework for sequential decision-making under uncertainty, highlighting its role across engineering, science, and data-driven domains. 2. Develop conceptual understanding of core RL methods and how they differ from other machine learning paradigms. 3. Equip students with computational and statistical tools necessary to design, implement, and evaluate effective RL systems. 4. Foster problem-solving skills in dynamic and uncertain environments, enabling students to recognize and address RL-specific challenges. 5. Demonstrate real-world applications of RL in data science and beyond, bridging theory with practical case studies. View Class Schedule