General Game Playing in AI Education
There are several ways to use General Game Playing in AI education:
- General Game Playing as part of a specialised graduate AI course, e.g. on Intelligent Agents, Knowledge Representation, or Computer Game Playing
- General Game Playing as integral part of a general introductory AI course, both on the undergraduate and the graduate level.
- A specialized graduate course on General Game Playing, where students develop their own players.
Even without reference to General Game Playing, any general introductory course to AI can profit from available tools: Instructors can use the General Game Control Software to let students experiment with any concrete game or agent domain, like e.g. the famous Wumpus World, for which a formalization in GDL exists.
Introductions to General Game Playing
Slides
- A single lecture (2hrs) as part of a graduate course on Intelligent Agents
- General Game Playing (pdf)
- General Game Playing (OpenOffice Impress source)
- Two lectures (6hrs) as part of a graduate course on Knowledge Representation
- General Game Playing 1 (pdf)
- General Game Playing 2 (pdf)
- General Game Playing 1+2 (OpenOffice documents in zip-file)
- A single lecture (2hrs) as part of a graduate course on Artificial Intelligence
- General Game Playing (pdf; in German)
- A single lecture (2hrs) as part of another graduate course on Artificial Intelligence
- General Game Playing (pdf; in German)
Sample Tutorial Questions / Assignments
- GGP tutorial questions for Intelligent Agents course (300769, 11s2 at University of Western Sydney)
- GGP assignment for Knowledge Representation course (COMP4418, 10s2 at UNSW)
- GGP assignment for Knowledge Representation course (COMP4418, 11s2 at UNSW)
General Game Playing as Integral Part of an Introductory AI Course
Slides
- Four lectures (6hrs) as part of an undergraduate Introduction to Artificial Intelligence
- Logic
- Logic Programming
- Planning
- General Game Playing
- Slide source (open office documents in zip-file)
Sample Tutorial Questions / Assignments
- Tutorials for Artificial Intelligence course (COMP3411, 11s1 at UNSW)
Full-Fledged General Game Playing Courses
The goal of a full-fledged course on general game playing is for students to develop their own players and, along the way, learn to apply a broad range of AI methods, including search, reasoning, planning, and learning.
Experiences at Stanford University and TU Dresden over several years suggest that the right size for such a course is about 15-30 participants with groups of 2-4 (ideally, 3) students building one player.
Links
- CS227B - General Game Playing, Stanford University (Michael Genesereth)
- General Game Playing, Technische Universität Dresden (Michael Thielscher)
- Planning and General Game Playing (in German), Universität Bremen (Stefan Edelkamp, Peter Kissmann)