MLKI2000 Machine Learning and Artificial Intelligence in Games

By completing the course, the student will have achieved the following learning outcomes: 

Knowledge 

The student 

• understands the complexity of various machine learning (ML) algorithms and their limitations in training
• understands issues and roles of Artificial Intelligence (AI) in the design of games 
• understands tactical and strategic AI used in gaming scenario 

Skills 

The student 

• can explain the general idea behind ML, as well as specific algorithms that are being used in real world scenarios 
• can use ML methodology for research and industrial settings using current trend of ML software libraries 
• can implement and apply ML methods to any chosen game 
• can produce game that the avatars are navigate around and served with the purpose based on goal-oriented/utility-theory 
• is capable of confidently applying common ML algorithms in practice and implementing their own algorithm
• can perform experiments in ML using real-world game scenario
• can program autonomous movement of avatars 
• can design and implement decision making and coordinating action based on heuristic, fuzzy sets or logics
• can read and understand scientific publications on ML/AI and formulate current issues, choice of methods, and results in a short, concise manner 

General competence 

The student 

• is able to plan and carry out varied tasks in accordance with ethical requirements and guidelines 
• is familiar with relevant issues of professional ethics, and is able to make a contribution to a professional community

Central topics: 

• Artificial Neural Network 
• Q-learning Reinforcement Learning 
• Use of popular Machine Learning libraries
• Goal-Oriented Action Planning 
• Utility-based Theory  

The course is organised as a combination of lectures, practical exercises and supervision. 

• 3–5 individual assignments for ML and AI respectively 
• attendance at lectures in accordance with the teaching plan  
• attendance at laboratory teaching in accordance with the teaching plan 

• 1 individual folder assessment on Machine Learning which counts for 50% the final grade 
• 1 individual folder assessment on Artificial Intelligence which counts for 50% the final grade 

To pass the course, both examinations must receive a passing grade. 

Alphabetical grades are used, on a scale from A to F, with E as the lowest passing grade.