Grant: Discovery Grants Program - Individual
Simon Fraser University
$20,000.00
Game Design Intelligence
EF 2016-17
Résumé
Video Games provide a unique challenge to Artificial Intelligence (AI) research. They take place in artificial worlds that are more easily sensed and acted upon than the real world. However, they are dynamic and contain sources of uncertainty such as player actions that are not easily modelled. They therefore occupy a middle ground between more static simulated environments on the one hand and robots on the other. Furthermore, they place strict limits on the available computing time for AI algorithms, and require support for controlling and managing these algorithms during game design and game play. Addressing these challenges of video games requires innovation grounded in experience with their restrictions.
The research program I propose here aims to answer some of the challenges of video game AI design. Its first objective is to integrate academic AI algorithms with those commonly used in video games. Specifically, a project will research merging algorithms that perform automated planning with reactive video game algorithms known as behaviour trees. The results will contribute to academic AI research by expanding the use of automated planners to new domains, as well as video game production by making the use of automated planners accessible to non-researchers.
A second objective of the research program is to investigate novel player models that do not require a priori models of the game being played. Formal models of video game worlds are usually not available during game development, but player modelling constitutes an important aspect of game AI. It supports predicting the player's future actions so AI driven characters can anticipate them, and adjusting the game due to knowledge of the player's current state. The algorithms to be investigated will utilize ""keyhole"" techniques that build online player models to perform predictions and offline player models to perform state analysis without relying on prior game models (as if watching the player through a keyhole). Modelling and predicting users without available application models constitutes a novel contribution to AI research and has the potential to transfer to industrial applications easily as it demands less investment from application and game designers.
A final objective of the research program is to explore how gameplay based on AI algorithms can change gameplay when communicated to players visually. Often, the expressivity of AI in a game context is limited by the available means to represent AI states and decisions. Games tend to restrict themselves to means like character animations and audio cues in this context. Displaying abstractions of AI algorithms directly in game greatly enhances the possibilities of communicating their workings to players and enabling new types of gameplay. A few suggestive examples of these techniques currently exist in industry, and they indicate a large realm of possibilities that this project will explore. The results will impact academic research in AI and human computer interaction as they will open new avenues to interact with intelligent characters and environments. The game industry will benefit from this research as it will show ways to solve presentation problems for AI in games, and thus open up novel game design spaces.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2014-05357
Name nom
Gorniak, Peter
Department département
Interactive Arts and Technology, School of (SIAT)
Organization id
5
Institution établissement
Simon Fraser University
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2016
Competition year année de concours
2014
Award amount
20000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
2803
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Knowledge representation
Research subject fr
Répresentation des connaissances
Partie
3
Nb partie
5
Application title
Game Design Intelligence
Keyword
NSERC
Application summary
Video Games provide a unique challenge to Artificial Intelligence (AI) research. They take place in artificial worlds that are more easily sensed and acted upon than the real world. However, they are dynamic and contain sources of uncertainty such as player actions that are not easily modelled. They therefore occupy a middle ground between more static simulated environments on the one hand and robots on the other. Furthermore, they place strict limits on the available computing time for AI algorithms, and require support for controlling and managing these algorithms during game design and game play. Addressing these challenges of video games requires innovation grounded in experience with their restrictions.
The research program I propose here aims to answer some of the challenges of video game AI design. Its first objective is to integrate academic AI algorithms with those commonly used in video games. Specifically, a project will research merging algorithms that perform automated planning with reactive video game algorithms known as behaviour trees. The results will contribute to academic AI research by expanding the use of automated planners to new domains, as well as video game production by making the use of automated planners accessible to non-researchers.
A second objective of the research program is to investigate novel player models that do not require a priori models of the game being played. Formal models of video game worlds are usually not available during game development, but player modelling constitutes an important aspect of game AI. It supports predicting the player's future actions so AI driven characters can anticipate them, and adjusting the game due to knowledge of the player's current state. The algorithms to be investigated will utilize ""keyhole"" techniques that build online player models to perform predictions and offline player models to perform state analysis without relying on prior game models (as if watching the player through a keyhole). Modelling and predicting users without available application models constitutes a novel contribution to AI research and has the potential to transfer to industrial applications easily as it demands less investment from application and game designers.
A final objective of the research program is to explore how gameplay based on AI algorithms can change gameplay when communicated to players visually. Often, the expressivity of AI in a game context is limited by the available means to represent AI states and decisions. Games tend to restrict themselves to means like character animations and audio cues in this context. Displaying abstractions of AI algorithms directly in game greatly enhances the possibilities of communicating their workings to players and enabling new types of gameplay. A few suggestive examples of these techniques currently exist in industry, and they indicate a large realm of possibilities that this project will explore. The results will impact academic research in AI and human computer interaction as they will open new avenues to interact with intelligent characters and environments. The game industry will benefit from this research as it will show ways to solve presentation problems for AI in games, and thus open up novel game design spaces.