IEEE Symposium on Computational Intelligence and Games


15 - 18 December 2008

The 2008 CIG symposium will feature a number of free tutorials on topics of interest to the computational intelligence in games communities. Tutorials offer an opportunity for extended instruction on either well-established or "cutting-edge" topics, allowing participants to gain an understanding of important established topics, develop higher skill levels in areas in which they are already knowledgeable, or learn about exciting new areas of research.

Tutorials will be scheduled for the first day of the symposium: Monday, 15 December 2008. Tutorials are free - attendance at the tutorials is included in the symposium registration fee.

Tutorial 1: Learning to Play Games
Presenter: Simon Lucas
Contact: sml AT essex.ac.uk
Slides [pdf]

This tutorial provides a practical introduction to game strategy learning with function approximation architectures, covering two main approaches to learning game strategy: evolution (including co-evolution), and temporal difference learning. The tutorial will show how the selected input features and the function approximation architecture has a critical impact on what is learned, as well as how it is interfaced to the game (e.g. as a value estimator or as an action selector). In addition to standard MLPs, attention is also given to N-Tuple systems and interpolated table-based approximators as they have recently shown great potential to learn quickly and effectively. Each method will be demonstrated with reference to some simple fragments of software, illustrating how the learning algorithm is connected with the game and with the function approximation architecture. Example games will include Othello, Simulated Car Racing, and Ms. Pac-Man.

Simon Lucas is the founding Editor in Chief of the IEEE Transactions on Computational Intelligence and AI in Games and the Chair of the IEEE CIS Games Technical Committee. He was the chair of CIG'05, and has organised numerous games-related competitions at a number of international conferences. Simon is a Reader in Computer Science at The University of Essex in the UK and has published over 130 refereed papers in international journals and conferences, including a number on computational intelligence and games. His homepage can be found at: https://dces.essex.ac.uk/staff/lucas/.

Tutorial 2: Inducing Agent Models from Examples
Presenter: Bobby Bryant
Contact: bdbryant AT cse.unr.edu
Slides [pdf]

In games, it is often useful to derive models of behavior from observations. This is true when creating game content, where non-player characters could be trained to the proper behavior on the basis of demonstrations by subject matter experts, and also during game play, where a game AI could model a player, or other in-game agents, in order to adapt its own responses appropriately. A number of techniques for inducing behavioral models from observations have been published, and interest in that area of research seems to be growing. This tutorial will review the literature on techniques and applications for agent modeling by induction, delineate the state of the art, and enumerate open challenges and potential applications. It will also provide an in-depth treatment of one or more techniques of special interest to the CI community, and provide useful background for this symposium's special session on Player/Opponent Modeling.

Bobby Bryant is an Assistant Professor of Computer Science and Engineering at the University of Nevada, Reno. He has been using a discrete state strategy game to study intelligent agents since 2003, addressing such problems as in situ divisions of labor, policy induction from examples, and visibly intelligent (as opposed to strictly optimal) behavior. He is now the chair for a task force on RTSGs for the IEEE Computational Intelligence Society's Games Technical Committee, and directs the Neuroevolution and Behavior Laboratory at UNR. His homepage can be found at: https://www.cse.unr.edu/~bdbryant/.

Tutorial 3: Measuring and Optimizing Player Satisfaction
Presenters: Georgios Yannakakis and Julian Togelius
Contact: yannakakis AT itu.dk and julian.togelius AT gmail.com
Slides [pdf]

This tutorial focuses on a range of approaches regarding quantitative player satisfaction (cognitive and affective) modeling and artificial intelligence (AI) for improving playing experience. Optimizing player satisfaction is the second research focus of the tutorial. That is, given successful models of player satisfaction how can we adjust interactive systems in order to improve player experience. The purpose of the tutorial is to initiate (or further increase) an interest among CIG'08 participants in this newly established field.
There are indications of high interest from a growing research community and the game industry. In AI and Games research, the status quo is just now beginning to shift toward the perspective, advocated here, of applying AI to model and enhance the player's satisfaction. This tutorial addresses this trend directly by investigating methods of modelling player satisfaction and adapting games to the desires of the individuals playing them. In this respect also, then, the tutorial is timely and significant. Building middleware - from the commercial game development viewpoint - capable of capturing player satisfaction in real-time will deliver products of higher commercial/marketing value and will automate specific game development processes like user testing.
Researchers and game developers will find valuable research results for improving the quality of their games through player satisfaction modelling and optimization techniques covered in the tutorial. So far, there are promising results in small-scale games (both screen-based and real-world physical interaction-based games) that can be used as a starting point. We will give examples from Pac-Man and car racing, and also of optimizing game rules for learnability.

Georgios Yannakakis is Assistant Professor at the Center for Computer Games Research, IT-University of Copenhagen. His primary research interests lie in cognitive (entertainment) modeling, affective computing, neuro-evolution, dynamic game balancing, and real-time learning in video games. Dr. Yannakakis is the chair of the IEEE CIS Task Force on Player Satisfaction Modeling. He has been the chair of two workshops (SAB'06, AIIDE'07) on areas strongly related to player satisfaction in games and program committee member of several game related conferences and workshops. His homepage can be found at: https://www.itu.dk/~yannakakis/.
Julian Togelius is a researcher at the Dalle Molle Institute for Artificial Intelligence (IDSIA) in Lugano, Switzerland. His research interests include evolving game-playing agents, modelling player behaviour, and evolving interesting game content, mainly using evolutionary and coevolutionary techniques. He also co-organizes the well-attended Simulated Car Racing Competitions for the IEEE CIG and CEC conferences. Julian's homepage is available from: https://julian.togelius.com/.