CHAIR OF

COMPUTATIONAL

INTELLIGENCE

This page contains information about the seminar "Hot Topics in Computational Intelligence in Games". The seminar is held in winter term 2021/2022 by Dr.-Ing. Alexander Dockhorn.

If you would still like to join the course, please let me know by writing a short e-mail to alexander.dockhorn@ovgu.de.

The seminar can be held online on request!

Language: 

The seminar will be held in English. Only if all participants agree, the seminar can also be given in German.

Participants and Topics:

Master student of the courses Informatik/Computer Science, Visual Computing and Data and Knowledge Engineering are eligible to visit this course. It is recommended for participants to have experience in game-development and artifical intelligence. Courses such as Introduction to Deep Learning, Computational Intelligence in Games or Intelligent Systems cover background on some of the topics that can be discussed in this course but are not required. The recommended background knowledge depends on the topic chosen by each student.

The following topics might be adressed in the course:

• (General) Game Learning
• Decision-Making in Complex Decision Spaces
• Action and Game-State Abstraction
• Deep Reinforcement Learning
• Procedural Content Generation
• Player Modelling
• Believable Agent Behavior
• Path-Finding versus Steering, Context Steering, and Multi-Objective Steering

Additional Topics on request

Organization:

The course will consist of three parts. First, a short overview of available topics will be given in the first meeting and each participant will be able to select a topic which they would like to study in more detail.

The second part will be structured by presentations given by the participants. Each presentation will cover the topic chosen by the participant (listed above or additional topics that have been approved). Presentations should have a length of about 30 minutes and will be followed by about 15 minutes of discussion.

In the final part, each participant is asked to apply and analyse one of the techniques learned in a personal project. This does not need to be the same topic as presented by the student but can be any of the topics seen during presentations. The result of each project is a working demonstration of the chosen topic in a game or a research framework with commented source code and a written documentation indicating the general concept of the solution and a "How to" to start the demo.

Both, the presentation as well as the project, will be part of this course's examination.

Timeline: 

If you still want to participate, please write a short e-mail to alexander.dockhorn@ovgu.de.