25.01.15 - 30.01.15, Seminar 15051
Artificial and Computational Intelligence in Games: Integration
The following text appeared on our web pages prior to the seminar, and was included as part of the invitation.
This seminar will bring together researchers and industry representatives who work at the forefront of artificial intelligence (AI) and computational intelligence (CI) in games. We will assemble practitioners and researchers working on different topics within the broader game AI field with the explicit purpose of investigating how their ideas, algorithms and domains can be combined. The seminar will include experts on topics from game tree search to machine learning to player modeling to content generation, in a wide range of games: from board games to commercial video games and serious games.
The video game industry is the largest of the entertainment industries and is growing rapidly. While modern games offer stunning graphics, the quality of the AI has for a long time been a weak point of many games. There is now a great deal of interest both from the game industry and from academic AI researchers in improving game AI, and in novel applications of AI methods such as procedural content generation, player experience optimization, and automated testing. However, the field is divided, with researchers and practitioners in different subfields not being aware of each others' work. There are a number of recently or not-so-recently developed AI solutions which promise to improve the quality of game AI, but which are not commonly known either in industry or in neighboring academic fields. We expect significant synergies and even new research directions to result from explicitly addressing the integration of these techniques and fields within this Dagstuhl Seminar.
In the previous seminar, groups gathered around a particular problem (e.g. player modeling or content generation) or a particular technique (e.g. search or planning) as applied to games. In this seminar, working groups will explicitly address the intersection of different techniques and problems (e.g.: How can planning be used in content generation? How can search techniques be used for learning human-like behavior?) and the use and modification of these techniques for game design and development (e.g.: How can we build game designs around player modeling? What are the requirements on case-based reasoning for effective use in game engines?). The idea is that leading experts in different techniques and problems assemble in groups in a semi-structured process, and through creative cross-fertilization proceed to solve problems and create new research topics. The main expected outcome of the seminar is a better understanding of how to integrate novel AI into games, game design, and related applications, and how to combine different game AI techniques.