Artificial Intelligence for Games
We hope to enhance the gameplay by creating tailor-made experiences and provide personalized games.
Personalization and consumer driven content is a relevant topic in multiple industries and he hope to use this new vision to enhance games and in particular the Artificial Intelligence field. User profiling is already common practice in existing commercial tools like Recommenders Systems, which are able to explicitly or implicitly obtain user preferences and are used to provide more tailor made information, and build in-depth user experiences.
We are harnessing that ability to improve and create more meaningful experiences to gamers and combine it with AI and personalizing users’ gameplay. We are approaching this research at different levels within a game production:
- increase NPC (Non-playable Characters) behavior and reaction to gamers;
- studying and counterbalancing gamer strategies;
- procedural and Personalized Stories and Level Design.
Advanced NPC Behavior
One of the key topics we are working on is Artificial Intelligence of NPCs. We are focusing in building challenging and believable artificial opponents that act as if capable of strategic thinking, and react to gamer’s course of action. Our approach creates adaptive behaviors that take into account long-term and short-term consequences, in order to have:
- hundreds of NPC’s in real time;
- individual NPC’s personalities;
- NPC’s adapt to gamer’s actions.
We are currently able to simulate a virtual city, with 300 NPCs, with different personalities, which are able to survive and adapt to the environment autonomously.
Studying and counterbalancing gamer strategies in real-time
The research we are conducting at the moment, allows game developers to study gamer’s actions and understand common and repeated techniques, and based on that, generate AI behavior to enhance or counter those actions, making games either more pleasant or more challenging to players.
With AI it is possible to recognize common used patterns and allow Game Developers and Designers to understand how gamers behave.
This type of tools is not something new, but, we are proposing clustering methodologies which enables this information to be collected and processed, in such a way that it can be injected back to the Artificial Intelligence system that controls the gameplay.
Procedural and Personalized Stories and Level Design
Procedural design and development is a very interesting field, which is attracting a lot of attention to the game developer’s community. With such tools it is already possible to procedurally generate complete levels.
What if we can use this to provide personalized content to gamers? Nowadays there are several systems providing personalized content to users, we are approaching this systems in a novel way. We are studying the possibility to use gamers’ profiles to customize their experience. By using the players past experience, we are researching to adjust narratives and level design using procedural generated content.
Within this research we hope to completely generate new levels, or section of levels, which can:
- create interesting plots to users;
- level customization with regard of users experience;
- avoid repetition and improve replay ability.
Impacting other fields
The research we are conducting is also proven to be effective in non-game fields, our main contribution have been with regard to:
- personalization Systems (Recommender Systems, Context Aware);
- data mining of user profiles;
- clustering and pattern recognitions.
Carlos Santos (email@example.com)
Giuseppe Magliore (firstname.lastname@example.org)