Practicing esports with AI
Learning esports can be difficult, but AI supported tutorials that adapt to the player can help
When starting a new game, players require an acclimatisation period for them to adjust to the controls and the mechanics of the game. To help players get accustomed, many video games use a tutorial to teach newcomers. Unfortunately, many tutorials can ignore many high level skills for newcomers and be far too simple for experienced players.
This PhD is focused on using modern AI research to help build more intelligent and adaptable tutorials for games. The main aims are: to automatically build skill trees for a game that map out the complexity of skills; to analyse player data which measures the mastery of each skill; and to manipulate in-game AI to set up scenario's in which to practice each skill. Such a system would be thoroughly useful for players in esports games, newcomers and professionals alike, and would allow for a more competitive scene. As well as video games, such a system could be used for more engaging educational, productivity, and creativity support software.
Contact: Joseph Hesketh