Table of Contents
Virtual worlds are one of the most compelling sites for research in modeling and reasoning to emerge in recent years. Such worlds as Meadows, and Pueblo are characterized by hundreds of people (e.g. K-12 students) connecting regularly (possibly from all over the Internet) building, playing, interacting (with each other, with "bots") in a persistent, object-oriented, user-extensible, text-rich context. Such worlds are often extraordinarily engaging for their participants: for instance in the last five months on Meadows, a handful of adults and approximately 200 children in Grades 4-9 have built over 10,000 objects (with varying levels of sophistication and fidelity) -- homes (rooms, appliances, toilets, swimming-pools, pets), hotels, cars, car rental agencies, shopping malls, cities, mythical monsters, fish, mazes...
Fundamental questions about modeling, representation and reasoning arise in this context. What kinds of ontologies are a suitable basis for such extraordinarily rich and diverse user-constructed worlds? What kinds of theories of space, time, action, change, causation and common-sense reasoning might one build into such worlds? How can such worlds be designed to facilitate experimentation with different "frameworks for physics"/protocols of interaction? What principles of construction must be followed in order to allow the designs of hundreds of such world participants to inter-operate with each other, within a virtual world and across virtual worlds? How can one design in-situ reasoners that can reason with world-constructions (and learn from past interactions) to help human participants in designing, configuring, building, monitoring, diagnosing/debugging systems and processes embedded in such worlds? How does one build intelligent navigators/browsers that can recognize --- amidst the tens of thousands of objects and processes --- parts of the world "interesting" along given dimensions? What kinds of visual/model-based "construction environments" might one deploy to make it easier to build such worlds? How does one design natural language interaction systems that can leverage the situated/contexted nature of interaction in these spaces? How can (collections of) software agents (bots) be designed to take advantage of the presence of other bots, the structure of the virtual world and interaction with people? What frameworks for ownership and control are needed to foster collaborative work in such worlds, while safeguarding intellectual property? How can multiple modalities of interaction (text, audio, speech, animation) be coordinated to present a coherent experience for participants?
This meeting aims to bring together theoreticians and practitioners from a variety of communities interested in some of these questions --- e.g., qualitative physics and model-based reasoning, model-based planning and diagnosis, hybrid (continuous/discrete)and real-time reasoning, knowledge representation, virtual worlds, constraint and visual programming, interactive theater.
Nine papers will be presented during this meeting.
TUESDAY 13 JANUARY
10:30-12:00, Track A
Special Session: Modelling and Reasoning in Virtual Worlds
Design Requirements for Network Spaces
Vijay Saraswat, AT&T
A Framework for Constructing Virtual Worlds
Seif Haridi, Per Brand, Nils Franzen, Erik Klintskog, Swedish Institute of Computer Science
Patterns and Process of Reasoning in Virtual Worlds
Peg Syverson, U. Texas Austin
13:30-15:00, Track A
Two dimensional shared virtual worlds in middle and elementary schools: lessons learned
Charles Hughes, Michael Moshell, U. Central Florida and Mark Pullen
Collaborative Exploration of Simulated Microworlds
Mark Shirley, Vicki O'Day and Danny Bobrow, Xerox PARC
Modeling motivational behavior in intelligent agents in virtual worlds
Alastair Burt, DFKI
15:30-17:00, Track A
Peter Wegner, Brown U.
Dynamic Semantic Virtual Reality in WAVE
Peter S. Sapaty, U. of Surrey, UK
Extending VRML for intelligent Virtual Agents
Philippe Codognet, INRIA Rocquencourt