Gaming personality - a dual systems approach to reinforcement learning
Most people who have played video games have at some point noticed the lack of intelligence in enemies or other non-player characters (NPCs). In better designed games, the NPC behaviour is often passable but there is rarely a reason for players to believe that game worlds are populated by actual people. With the overarching view of humans as creatures embodied within their environment, we can use the same approach for NPCs as we do for agents that play games. So how do we build AI characters and agents that behave like humans?
My research combines the approaches of psychology and neuroscience with computer science and machine learning. Using data from experiments with human subjects we can model the behaviour of players and use the results to inform what algorithms and parameters to apply when building AI agents. By comparing the behaviour and abilities of our agents with those of human players, we also create a positive feedback loop where findings have potential to both improve AI and further our understanding of ourselves.
An established middle ground between these fields is reinforcement learning; learning by reward. Another important aspect of modelling human behaviour is to consider our individual differences. With more accurate models of human behaviour, we open the door to many possibilities. We could build more interesting NPCs that not only have different personalities but are also able to learn how to act in new environments, enabling game creators to more easily expand the game world with new content without explicitly programming the characters. General game playing agents would also allow for automatic play testing of games for different types of people.
Henrik got hooked on video games in the early ‘90s when he caught his dad playing Tetris instead of working. His interest for the human mind began when Henrik was a teenager and became fascinated with dreams, especially lucid dreaming. He intended to learn more through the combination of mathematics and natural science at Lund University, but later went for a more holistic view of the brain as the better approach. Armed with a BSc in Psychology, an MA in Cognitive Science and a life time of playing video games, Henrik has now come full circle and is using the IGGI PhD programme to launch his quest for world domination through game playing AI based on cognitive modelling of human behaviour.
Home institution: Goldsmiths
Supervisor: Professor Alan Pickering
Ready to apply?
Once you have identified your potential supervisor, we would encourage you to contact them to discuss your research proposal.Learn More