Deep Learning for Games and Simulations (Autumn 2024)

Abstract

The goal of the course is to teach the understanding, design, implementation, and use of modern artificial intelligence (AI) techniques for generating efficient, intelligent behaviors in games and other simulation environments. A particular focus is given to techniques applied to agent-based simulations.

Description

Artificial intelligence has many applications, and many techniques need to be applied in simulated environments, whether during the training (as in robotics or systems simulations) or deployment (such as in computer games). The topics covered in this course apply to games, simulation environments, robotics, and many other areas.

Students will learn a broad understanding of AI algorithms' theoretical, practical, and implementation sides.

The course will partly cover the following topics (AI techniques and problems): 

• neural networks,
• deep reinforcement learning, 
• evolutionary algorithms, 
• neuroevolution,
• Monte Carlo Tree Search (MCTS), 
• hybrid approaches (MCTS+Neural networks), 
• adversarial search,
• multi-agent learning,
• as well as the overview of the latest applications of these techniques in the industry.

Formal prerequisites

Students must have completed a course on programming such as "Introductory Programming" or "Object-Oriented Programming." Furthermore, the student should have familiarity with basic reinforcement learning and graph search algorithms. We will mainly use the Python programming language during the course, so familiarity with the language is an advantage.

Intended learning outcomes

After the course, the student should be able to:

• Theorize about and describe the AI algorithms covered in the class
• Identify tasks that can be tackled through advanced AI techniques and select the appropriate technique for the problem under investigation
• Compare the performance of different AI techniques and reflect on their suitability for different domains
• Design and implement efficient and robust advanced AI algorithms
• Work efficiently in groups

Ordinary exam

Exam type:
D: Submission of written work with following oral, External (7-point scale)
Exam variation:
D2G: Submission for groups with following oral exam supplemented by the submission. Shared responsibility for the report.