Puzzle game testing agent
- Predicting level difficulty, estimate churn
- Experiment with techniques such as Deep Reinforcement Learning and Monte Carlo Tree Search
- Company collaboration with Tactile Games
- Inserted in a larger research project
- Open to other games
Player modelling: churn prediction
- Analytics/Game design or AI
- Free-to-play games
- Understand when, and why players churn
- Identify strategies for retention
Player modelling: GAIL
- Create realistic agents from play traces
- Experiment with Generative Adversarial Imitation Learning
- Work with Vizdoom or any other platform you like
Control vs emergence in PCG
- Procedural content generation
- AI algorithm as co-creator
- Investigate novel algorithms (generative adversarial networks, transformers, etc…)
- How to control these algorithms
- Balance between emergence and control
Modelling player cognition and emotions
- Potential collaboration with the augmented Cognition Lab
- Analysing physical signals
- Connect player behaviour with physiology
Computational cinematography
- Collaboration with Kaspar.AI
- Automated film editing
- Identifying scenes and cuts
- Build scene sequences
- Generating narratives
- Make movies from videogames (automated machinima)
- Automated film classification
Sport analytics (internship opportunity)
- Better data visualisation
- Modelling of career projections
- Strategy and tactics modelling
- Collaboration with Brøndby IF