The North Carolina Serious Games Initiative held its first event today, a Workshop on Collaboration in the North Carolina Serious Games Space. One idea behind the event was to bring together two audiences -- academic researchers and commercial developers focused on serious games projects -- to enable awareness and cross-pollination between the groups. Another reason for holding the workshop was to demonstrate that North Carolina is already a center of excellence for serious games, and for the assembled attendees to discuss how best to build on this in the future.
The day was structured as a series of short presentations, with the academics speaking in the morning. I focused my note-taking on those sessions, all of which were interesting. As a graduate student from UNC Chapel Hill said of himself and his colleagues at the end of the day, "We're stunned by the breadth of serious games research happening here in North Carolina." I wasn't able to take notes for every talk, but most of them are here.
James Lester, NCSU Computer Science
James' group is focused on human-computer interaction and communication, creating "flexible, adaptive systems that fundamentally enhance human problem-solving. His group is researching "adaptive interaction," defined as "tracking beliefs, preferences, goals, and plans of the user to provide customized advice" within applications such as education, training, analysis, and entertainment. To do this, he's building software that performs plan recognition, misconception detection, evaluation of the user's current knowledge state, and evaluation of usage and learning styles and preferences.
Tiffany Barnes, UNC Charlotte Future Computing Lab
Tiffany's research interests include serious games, advanced learning technologies (educational data mining, intelligent tutoring and agents), and diversity in computing.
The idea behind Tiffany's Game2Learn project is to build a game where students learn computer science as they play. Their programs affect the game world, and later they can create new areas in the game world. She's just beginning to look at educational applications beyond computer science, including English, history, and others. Her students are currently using Unreal Tournament 2004 to build a fantasy role-playing game in which the gameplay involves creating programs to solve problems.
To address intelligent tutoring and agents, Tiffany's group is building virtual students: agents that interact with students like peers. They're able to observe and ask questions of students, as well as answer questions posed by students.
Adriana de Silva e Souza, NCSU Communication
Adriana is investigating mobile communication technologies: what happens to our relationships with spaces and the Internet when we have devices that are continuously connected to the Internet. She's using location-based mobile games to research this. She used the term "hybrid reality games" to describe this space: Botfighters (2001, Sweden), Mogi Mogi (2004, Japan), and Frequency 1550 (2005, Netherlands). In terms of potential users for hybrid reality games, Adriana cited education and tourism, among others.
Len Annetta, NCSU College of Education
Len talked about "HI FIVES: Highly Interactive Fun Internet Virtual Environments in Science" -- building tools that enable Grades 5-9 teachers to build video games themselves to facilitate science education. They started with Active Worlds but have since switched to Half-Life 2 for the forthcoming version of their tool.
"We don't stop playing because we grow old. We grow old because we stop playing." -- Oliver Wendell Holmes
"The worst thing a kid can say about homework is that it is too hard. The worst thing a kid can say about a game is it's too easy." -- Henry Jenkins, MIT
- This is what kids are doing
- Estimates suggest 3.38 billion hours are spent engaged in video games per year
- Game industry was $10 billion last year
- 2000 NAEP results suggest 29% of 4th graders, 32% of 8th graders, and 18% of 12th graders performed at or above proficient levlels in science
Why is this happening (i.e., why do proficiency levels drop so dramatically)? Because we're just lecturing material, memorizing instead of internalizing. Our goal is to make science real to kids -- real to their outside lives.
Anselmo Lastra, UNC Computer Science
Anselmo showed examples of research in virtual environments, including "redirected walking," which uses manipulation of the visual representation of a virtual space to manipulate how a user moves while wearing an immersive display -- in the example shown, a user was placed in a very small room but tricked (through visual clues) into thinking they were walking in a much larger space. Another example shown was procedurally generated house interiors -- using rules to create unlimited numbers of spatially distinct but plausible houses for large virtual environments.
Christopher Healey, NCSU Computer Science
Christopher's research focus is visualization for serious games development and other applications. He showed a variety of examples of innovative techniques for visualizing multiple datasets within a single image or animation. His visualization techniques are designed to be perceptually optimal, multidimensional, real-time, and use dynamic reconfiguration to adapt to viewers' changing interests and needs.
Michael Young, NCSU Computer Science
Michael runs NCSU's Liquid Narrative group. NCSU has a focused effort in serious games, bringing together the Education, Communication, Design, and Computer Science departments, brought together in the Center for Digital Entertainment. The Liquid Narrative group is focused on AI for games. Example AI problems they're looking at include story generation and adaptation, intelligent camera control, and cognitive modeling. A lot of their work involves taking ideas from other disciplines (drama, cognitive psychology, linguistics, film, etc.). The challenge is that most of this work is non-computational, so Michael's group has to build computational models based on this existing knowledge.
Michael described a platform his group is building, Zocalo. Zocalo is a "service-oriented architecture" for game control. It's currently compatible with both Unreal Tournament 2004 and Source. Current services research done within Zocalo includes:
- Task planning
- Cognitive modeling
- Natural language generation
- Camera control
What's all this good for? Integrating existing commercial game engines with complex AI toolkits. Generating tailored novel game experiences. Structuring game experiences so that key features of the story are clear to players. Allowing self-adapting interaction in the face of unexpected user behavior.
Ben Watson, NCSU Computer Science
Ben's group is focused on what he calls Design Graphics:
- Imagery about designed objects: Automating 3D modeling for digital production
- Imagery as designed objects: Automating graphic design to convey information
- Designed display: New types of pictures matching the way we see
- Visual design fundamentals: Applied perceptual experiments
With automated (or procedural) modeling, Ben is addressing the problem of increased content requirements for next-generation game consoles. He showed an impressive-looking demonstration of layering procedurally generated homes on a scanned grid of streets. His future directions include more diversity, higher-level control, changes over time, and lower-level objects like cars, furniture, and appliances.
Ben also showed "adaptive frameless rendering", which is an effort to create a rendering algorithm that functions more like the human visual system -- abandoning the concept of an animation as a series of static frames.