Intelligent Systems Demonstrations
The IS Demos program will be held Wednesday, July 25, in the Regency Ballroom, 6:30 –10:00 PM
Continuing advances in Artificial Intelligence research are making it possible to develop intelligent systems in a wide range of application areas. The AAAI-07 Intelligent Systems Demonstrations program showcases state-of-the-art AI implementations and provides AI researchers with an opportunity to show applications of their research in action. The program is intended to highlight innovative contributions to the science of AI with an emphasis on the benefits to be gained from developing and using implemented systems in AI research.
This year's demonstrations cover a broad range of domains: game playing, semantic web, distributed optimization, intelligent tutors, and more. System builders will be on hand to present their work. All that is needed to make this evening a big success is your active exploration of these interactive systems!
An Interactive Constraint-Based Approach to Sudoku
Christopher G. Reeson (University of Nebraska-Lincoln), Kai-Chen Huang (USC/Information Sciences Institute), Berthe Y. Choueiry (University of Nebraska-Lincoln and USC/Information Sciences Institute)
We present a Java applet that allows a user to interactively solve a Sudoku puzzle using Constraint Processing (CP) techniques. Our system showcases the power of CP techniques in solving problems through a widely familiar and easily approachable puzzle. It allows the users to apply several consistency algorithms or work independently of our algorithms. The system is available online: sudoku.unl.edu/Solver and sudoku.unl.edu/Constructor.
AURA: Enabling Domain Experts to Construct Declarative Knowledge Bases from Science Textbooks
Ken Barker (University of Texas at Austin), Vinay Chaudhri (SRI international), Shaw-Yi Chaw (University of Texas at Austin), Peter E. Clark (The Boeing Company), Daniel Hansch (Ontoprise), Bonnie John (Carnegie Mellon University), Sunil Mishra (SRI international), John Pacheco (SRI International), Bruce Porter (University of Texas at Austin), Aaron Spaulding (SRI international), Moritz Weiten (Ontoprise)
We will demonstrate the AURA system that can enable students in Physics, Chemistry, and Biology to author knowledge in a college level science textbook in a way that the resulting knowledge base can be used to answer questions from an Advanced Placement examination in the respective discipline. The system supports a knowledge formulation interface based on concept maps, equations, tables and a query formulation interface based on controlled English. We will also demonstrate a semantic Wiki application for collaborative knowledge formulation and show how users can map large amounts of the Wiki’s knowledge into AURA.
A Demonstration of ScriptEase Interruptible and Resumable Behaviors for CRPGs
Maria Cutumisu, Duane Szafron, Jonathan Schaeffer, Kevin Waugh, Curtis Onuczko, Jeff Siegel, Allan Schumacher (University of Alberta)
We developed a behavior model based on generative design patterns for intelligent non-player characters (NPCs) using ScriptEase that solves the difficult problem of interacting NPCs and of interruptible and resumable behaviors. We demonstrate how this model generates complex NPC behaviors from behavior patterns in a commercial game without writing code.
A Deployed Semantically-Enabled Interdisciplinary Virtual Observatory
Deborah McGuinness (Stanford University and McGuinness Associates), Peter Fox (National Center for Atmospheric Research), Luca Cinquini (National Center for Atmospheric Research), Patrick West (National Center for Atmospheric Research), Jose Garcia (National Center for Atmospheric Research), James L. Benedict (McGuinness Associates), Don Middleton (National Center for Atmospheric Research)
We have used semantic technologies to design, implement, and deploy an interdisciplinary virtual observatory. The Virtual Solar-Terrestrial Observatory is a production data framework providing access to observational datasets. It is in use by a community of scientists, students, and data providers interested in the middle and upper Earth’s atmosphere, and the Sun. The data sets span upper atmospheric terrestrial physics to solar physics. The observatory allows virtual access to a highly distributed and heterogeneous set of data that appears as if all resources are organized, stored and accessible from a local machine. The system has been operational since the summer of 2006 and has shown registered data access by over 75% of the active community (last count over 600 of the estimated 800 person active research community). This demonstration will highlight how semantic technologies are being used to support data integration and more efficient data access in a multi-disciplinary setting. A full paper on this work is being published in the IAAI 07 deployed paper track.
Evac-Op: Disaster Evacuation Support
Christopher J. Carpenter, Christopher J. Dugan, Joseph B. Kopena, Robert N. Lass, Gaurav Naik, Duc N. Nguyen, Evan Sultanik, Pragnesh Jay Modi, William C. Regli (Drexel University)
Evac-Op is a prototype system for assisting emergency personnel in monitoring and conducting evacuation and sheltering operations. It is a novel application of distributed constraint optimization combined with mobile wireless networking to sharing situation information and making global decisions on issues such as shelter assignments. In addition to exploring this application and new applications of distributed constraint optimization, Evac-Op is intended as a vehicle for investigating distributed decision making under poor communications, uncertainty, and change.
A Framework for Experimental Research of Autonomous Trading Agents
Eric Sodomka, John Collins, Maria Gini (University of Minnesota)
The Trading Agent Competition for Supply Chain Management is a market simulation in which autonomous agents act as manufacturers in a two- tier supply chain marketplace. The objective of this demonstration is to showcase various methods our research team and others have developed in this domain to run simulations and analyze agent performance in an efficient manner.
Freebase: A Shared Database of Structured General Human Knowledge
Kurt Bollacker and Timothy Kientzle (Metaweb Technologies, Inc.)
Freebase is a practical, scalable, graph-shaped database of structured, general human knowledge, inspired by Semantic Web research and collaborative data communities such as the Wikipedia. Freebase allows public reads and writes through an HTTP-based graph-query API and a sophisticated AJAX for web access. (See www.freebase.com)
Generating and Solving Logic Puzzles through Constraint Satisfaction
Barry O'Sullivan and John Horan (University College Cork, Ireland)
Solving logic puzzles has become a very popular pasttime, particularly since the Sudoku puzzle started appearing in newspapers all over the world. We have developed a puzzle generator for a modification of Sudoku, called Jidoku, in which clues are binary disequalities between cells on a 9x9 grid. Our generator guarantees that puzzles have unique solutions, have graded difficulty, and can be solved using inference alone.
This demonstration provides a fun application of many standard constraint satisfaction techniques, such as problem formulation, global constraints, search and inference. It is ideal as both an education and outreach tool. Our demonstration will allow people to generate and interactively solve puzzles of user-selected difficulty, with the aid of hints if required, through a specifically built Java applet.
The More the Merrier: Multi-Party Negotiation with Virtual Humans
Patrick Kenny, Arno Hartholt, Jonathan Gratch, David Traum, Stacy Marsella, Bill Swartout (USC/Institute for Creative Technologies)
The Institute for Creative Technologies at USC virtual human demo will focus on fully embodied conversational characters that contain task and emotion models. The user will use natural speech to negotiate with the agents to move a clinic. The agents will engage in multi-party dialogue using verbal and non-verbal behavior.
The PhotoSlap Game: Play to Annotate
Tsung-Hsiang Chang, Chien-Ju Ho, and Jane Yung-jen Hsu (National Taiwan University)
PhotoSlap, an intelligent system for semantic annotation of photos, contains a semi-automatic face detector, a bulk annotation tool, and a multi-player online game. By exploring the design principles of gameplay and applying game theoretic analysis, PhotoSlap is designed as a fun and productive game, which adapts itself to different players to produce the desired output.
The original AAAI-07 IS Demos call for papers is available as a printable PDF document
IS Demo Cochairs
Rob Miller (Massachusetts Institute of Technology), Chair
Holger Hoos (University of British Columbia), Cochair