Issue No.04 - July/August (2002 vol.17)
Published by the IEEE Computer Society
Perhaps one of the largest surprises in computing at the end of the 20th century was the extent to which entertainment became a major driving application. The power of the consumer market far outweighs that of the scientific or even defense markets. For example, 3D visualization hardware that would be the envy of any supercomputer center a few years ago is now sold in discount stores, owing to the volume production made possible by the consumer market. Recently, the US computer game industry has been reporting higher revenues than that of the American movie industry, a landmark indicating the widespread popularity and acceptance of computer gaming as a form of entertainment. It seems only natural, then, that interactive entertainment should be considered as a valid applications area for computer science research.
Artificial intelligence in particular has much to offer to computer gaming. The quality of a game's AI is now a leading product differentiator; poor AI can kill a game and great AI becomes an important selling point. From an AI research perspective, game-oriented research offers new, richer simulated worlds and the chance to have significant impact on an important industry. After all, would you rather do your research with Blocks World or Truck World or with the simulated worlds of Civilization, Close Combat, or The Sims? Consequently, there are now numerous efforts to build bridges between the AI research community and the computer game industry.
The articles in this special issue are gleaned from one of those bridging efforts. For the past four years, a series of AAAI symposia have brought together AI researchers and game designers and developers to improve communications and seed collaborations. They represent a sample of how gaming and other forms of interactive entertainment provide productive venues for AI research, and how results in AI can improve existing types of entertainment and even lead to new ones.
In "Character-Based Interactive Storytelling," Marc Cavazza, Fred Charles, and Steven J. Mead propose a way to let viewers talk back to an unfolding story, coaching the characters from the sidelines to make events unfold in more interesting ways. Planning techniques let characters respond dynamically to player interference while trying to achieve their goals. In "How Qualitative Spatial Reasoning Can Improve Strategy Game AIs," James V. Mahoney, Kevin Dill, and Kenneth D. Forbus describe how to use qualitative spatial reasoning techniques to improve the quality of computer opponents in strategy games. They show how using visual reasoning techniques might improve terrain analysis and provide richer conceptual common ground with players. In "Toward a New Generation of Virtual Humans for Interactive Experiences," Jeff Rickel, Stacy Marsella, Jonathan Gratch, Randall Hill, David Traum, and William Swartout discuss their efforts to create virtual humans to support military training. They marshal an impressive array of technologies, including speech I/O, dialogue management, cognitive architecture, and emotion modeling, exploring how to meld them into effective and instructive partners in virtual worlds.
In "A Behavior Language for Story-Based Believable Agents," Michael Mateas and Andrew Stern describe a new behavior language for creating actors in interactive drama. Their language supports both the reactive capabilities needed to respond to player actions and the coordinated actions of multiple actors needed to express the author's narrative. Aaron Khoo and Robert Zubek, in "Applying Inexpensive AI Techniques to Computer Games," show how lightweight AI technologies—behavior-based robotics and Eliza-style natural language processing—can produce interesting computer opponents for first-person tactical combat games. An ingenious aspect of their approach is careful social engineering, creating obnoxious opponents that humans take delight in vanquishing.
Although AI is already a critical factor in interactive entertainment's success, these articles suggest that it can lead to novel entertainment genres. They also suggest new ways that we can use AI to make better software for education and training by creating more engaging (and hence motivating) experiences for learners. More attention to such issues could multiply AI's already potentially revolutionary impact in our society.
Kenneth D. Forbus is a professor of computer science and education at Northwestern University. His research interests include qualitative reasoning, analogy and similarity, sketching and spatial reasoning, cognitive simulation, reasoning system design, articulate educational software, and the use of AI in computer gaming. He received his PhD from MIT in Artificial Intelligence. He is a fellow of the AAAI and serves on the editorial boards of Artificial Intelligence, Cognitive Science, and the AAAI Press. Contact him at the Qualitative Reasoning Group, Northwestern Univ., 1890 Maple Ave., Evanston, IL, 60201; email@example.com.
John Laird is a professor of electrical engineering and computer science at the University of Michigan and Associate Chair of the Computer Science and Engineering Division. He received his PhD in Computer Science from Carnegie Mellon University. He is one of the original developers of the Soar architecture and leads its continued development and evolution. From 1992 to 1997, he led the development of TacAir-Soar, a real-time expert system that simulates military pilots and is used for training in large-scale distributed simulations. His research includes creating human-like AI systems in computer games. He is a fellow of the AAAI. Contact him at the University of Michigan, 1101 Beal Ave., Ann Arbor, MI, 48109; firstname.lastname@example.org.