This lecture from Memorial University's Computer Science 3200/6980 (Winter 2025) covers key reinforcement learning concepts, focusing on Temporal Difference Learning. Learn about on-policy vs off-policy methods, epsilon-soft implementations, SARSA, Q-Learning, and the differences between tabular and deep reinforcement learning through practical examples like the Cliff problem. The second half includes exam preparation guidance and a comprehensive demonstration of Assignment 5, complete with code walkthrough. Taught by Professor David Churchill, this session is part of the Introduction to Artificial Intelligence course that applies algorithmic techniques to game-based problem solving.
Intro to Artificial Intelligence - Temporal Difference Learning - Lecture 19
Overview
Syllabus
00:00 - Intro
01:32 - On Policy vs Off Policy
09:12 - Epsilon-Soft Code / Example
14:08 - Off Policy Methods
16:08 - Temporal Difference Learning
22:46 - Driving Home Example
29:06 - SARSA
33:21 - Q-Learning
36:11 - The Cliff Example
41:48 - Tabular vs 'Deep' RL
42:57 - Exam Questions
43:48 - Assignment 5 Demo
01:02:59 - Assignment 5 Code
Taught by
Dave Churchill
Related Courses
-
Introduction to Artificial Intelligence: Temporal Difference Learning - Lecture 17
-
Sample-based Learning Methods
4.7 -
Reinforcement Learning with Gymnasium in Python
-
Decision Making and Reinforcement Learning
-
Intro to Artificial Intelligence - Temporal Difference Learning - Lecture 19
-
Q-Learning - Model Free Reinforcement Learning and Temporal Difference Learning
