AI4Games 2013 - Artificial Intelligence Methodologies for Games

Video games have proven to be a privileged domain for the study of artificial intelligence. A motivational factor is that video games are fun to play and captivating to observe. More interestingly though, they provide increasingly rich and interesting dynamic environments that recreate and model many problems of the real world. As the creatures and entities that inhabit such environments and the environments themselves elicit increasingly higher expectations as a result of the increasingly realistic visual technology used to build them, there is a strong need for methodologies to support the creation of behaviour able to match such expectations and maintain a coherent whole. To that end, methodologies from artificial intelligence are increasingly impacting video game development, while themselves adapting to the constraints of the new medium. Certainly real-time performance is an issue, but other issues such as creating believable and purposely imperfect behaviour to make opponents, individually or as a group, more interesting to interact with are as relevant for the game industry as devising optimal strategies. Artificial intelligence should be able to understand and adapt to the player and know how and when to loose believably, to maintain an optimum gaming experience. Finally, artificial intelligence in video games is not constrained to the fabric of video games themselves. Nowadays, artificial intelligence assists designers and developers alike in the very process of video game design and development, from the inception to the final result, from the supporting story to the simulation of user playthroughs for playtesting. In the quickly evolving medium of digital entertainment, artificial intelligence is taking an increasingly important role.
From an academic perspective, games are an excellent test bed for many AI research fields. Puzzles, classical games and board games are excellent test beds for search algorithms and learning methodologies. Social games and multi-player games are very good test beds for multi agent systems, negotiation algorithms, emotional modelling, user/player modelling. Strategy games and action games are very good for testing real-time decision making, path finding, planning and strategic/tactical reasoning. This interest goes back to the sixties with Arthur Samuel’s checkers program as an example. On the practical side, the game industry is now a multi-billion dollar industry whose main challenges are focused on the creation of richer player experiences. Beyond the audio-visual issues and interaction technologies, player experiences are fundamentally enriched by increasing the believability of the narrative and of the artificial characters, i.e. providing the artificial characters with behavior consistent with their personality. This latter goal is fundamentally attained by using AI methodologies and techniques. As examples we have the use of non-player characters with desires and capabilities in the “Sims 3” game and hinted-execution behavior trees in the “Driver: San Francisco” game.
The Second Thematic Track on AI Methodologies for Games AI4Games’2013 is targeted to both academic researchers and practitioners that use/develop Artificial Intelligent methodologies and techniques for games.
Topics of Interest
The Thematic Track will be structured around the following themes:
- Affective Gaming and Emotional Modeling
- AI games competitions
- AI in classical games and board games
- General game players
- Interactive Storytelling
- Learning and adaptation in games
- Multi-player games
- Natural language for games
- Natural user interfaces
- Negotiation methodologies for games
- Path finding/movement algorithms
- Planning for games
- Game playing bots
- Puzzle solving
- Real-time decision making
- Representation issues in games
- Search algorithms
- Serious Games
- Social Games
- Strategic/tactical games
- User/Player modelling
- Virtual characters and autonomous agents
- Virtual, augmented and alternate realities
- Voice recognition and text to speech
This list of topics is not exhaustive. Papers may address one or more of the listed topics, although authors should not feel limited by them. Unlisted but related topics are also acceptable, provided they fit in the Thematic Track’s main subject.