ECE 414/517
Description
Schedule

UTK ECE 414/517: Reinforcement Learning (Fall 2024)

Time and Location

  • Lectures: Tuesday and Thursday, 12:55 pm - 02:10 pm, in MKB-405

  • Office Hours:

    • Fei: by appointment with emails (in MK-612)

      Anik: each Wed & Thur 3:00 pm - 4:00 pm

Instructors

Course Description

This course offers a comprehensive introduction to reinforcement learning, a type of machine learning where agents learn to make decisions by interacting with their environment. Students will explore core concepts such as Markov Decision Processes, value functions, and policy optimization. Through hands-on projects, students will apply these techniques to solve complex problems in areas like robotics, game AI, and autonomous systems. By the end of the course, participants will have a solid understanding of how to build and implement reinforcement learning models in real-world scenarios.

The students are expected to sign up on Canvas.

Prerequisites

The students are expected to have background in linear system theory, probability theory, and optimization theory, as well as a strong programming background.

Grading

Grading will be based on the following rubric.

Class Participation 5%
Assignments (4*5%) 20%
Project 1 (with report) 20%
Project 2 (with report) 20%
Final Project (with presentation report) 35%

  • Participation (5%; ongoing). Participation points can be earned by actively engaging in discussions during class and on Canvas.

  • Assignments (20%). There will be approximately 4 written assignments throughout the semester. Students are expected to complete all assigned problems individually.

  • Projects (75%). There will be 2 mini-projects + 1 final project. These projects will usually be due two weeks after the description is handed out. The final project will be a project of your choice and will include a presentation. All the project include reports and programming assignments. The first 2 mini-projects are individual while final projects should be done in groups unless given permission otherwise. Therefore, you might want to find a partner sooner than later.


Late Submission Policy: All deadlines are firm unless notified in advance. Late submissions can only be accepted within next 24hr of the deadline but will result in a straight 25% off of grade.

Acknowledgments

The material in this course is inspired and significantly influenced by the teaching of Prof. David Silver, Dr. Amir Sadovnik, and Prof. Weizi Li.