Courses

Reinforcement Learning: From Theory to Algorithm

发布时间:2022-12-18 浏览量:718

Course name:Reinforcement Learning: From Theory to Algorithm

Course code: CS7309

Credit/class hours: 3/48

 

Textbook and References

Richard S. Sutton and Andrew G. Barto, Reinforcement Learning: An Introduction, Second Edition, 2018
  

Course Details

This course will provide a solid introduction to the field of reinforcement learning and students will learn about the core challenges such as how to evaluate and improve policy, how to make model‐free learning and generalization and exploration, also the approaches, including policy iteration, Q‐learning and policy gradient methods. Through a combination of lectures and programming practice, students will become well versed in theory, algorithms and techniques for reinforcement learning.

 

Syllabus

Chapter Topic Lecture and Quizzes Video
1 Introduction
-
2

Markov Decision Processes


-
3

Dynamic Programming


-
4

Model-Free Prediction(1)


-
5

Model-Free Prediction(2)


-
6

Model-Free Control


-
7

Value Function Approximation


-
8

Advanced DQN


-
9

Policy Gradient


-
10

A3C & PPO


-
11

DDPG & Soft AC


-
12

Optimal Control and Planning


-
13

Model-Based Reinforcement Learning


-
14

Variational Inference


-
15

Meta-Reinforcement Learning


-

 

Requirements and Grading

Presentation: 30%

Quizzes: 25%

Attendance: 5%

Final Project : 40%

 

Contact

zhangtianshi@sjtu.edu.cn (TA, Tianshi Zhang)