Sponsored by the Association for the Advancement of Artificial Intelligence
IAAI-03 was collocated with the International Joint Conference on Artificial Intelligence and held August 12–14, 2003, at the convention center in Acapulco, Mexico.
The Fifteenth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI–03) continued the IAAI tradition of including both case studies of deployed applications and papers on emerging AI applications. The deployed applications papers all included concrete measurements of the benefits provided by AI technology. IAAI–03 was organized as an independent program within the International Joint Conference on Artificial Intelligence, with schedules coordinated to allow attendees to move freely between IJCAI and IAAI sessions. We appreciated the cooperation of the IJCAI organizers since we believe the collocated conferences both benefited from each other. AI applications developers benefited from learning about the latest AI techniques that will enable the next generation of applications. Basic AI research benefited by exposure to the challenges of real-world domains and difficulties and successes in applying AI techniques to real business problems. IAAI–03 addressed the full range of AI techniques, including knowledge-based systems, planning and scheduling, perception and monitoring, knowledge formation, knowledge management, learning, intelligent design, natural language processing, and diagnostic reasoning. Deployed applications are case studies that provide a valuable guide to designing, building, managing, and deploying systems incorporating AI technologies. The 2003 papers addressed applications in a wide variety of domains, including natural language processing, secure mobile agents, sales support, securities fraud detection, and scheduling. These applications provided clear evidence of the impact and value that AI technology has in today’s world. Papers on emerging applications and technologies described efforts whose goal was the engineering of AI applications. They informed AI researchers about the utility of specific AI techniques for applications domains and also informed applications developers about tools and techniques that will enable the next generation of new and more powerful applications.
For 2003, we were very pleased to have three invited talks. Noah Friedland discussed recent progress on a long-term project to build a “Digital Aristotle,” an application capable of answering a broad range of questions in a specific domain. Noah discussed progress in answering questions from an advanced high school chemistry syllabus. Representatives of the specific projects were available to answer questions about the methods they used and the challenges they faced. Luis von Ahn presented the state-of-the-art in CAPTCHA research. CAPTCHAs are simple intelligence challenges that are intended to be easy for humans to solve but difficult for robots. CAPTCHAs are in widespread use on the Internet to keep robots out of “human-only” areas. Larry Hunter discussed emerging new challenges in artificial intelligence applied to biology applications. Biology offers challenging problems — and the opportunity for AI to make a real and direct impact on people’s lives.
John Riedl and Randy Hill
Noah S. Friedland,
Project Halo: Toward a Digital Aristotle
Vulcan Inc. has launched a multistaged effort toward the creation of a digital Aristotle, an application capable of answering and providing cogent explanations to arbitrary questions in an ever-growing number of domains. The pilot phase was a six-month effort intended to investigate the state of the art in question answering, with an emphasis on deep reasoning. Friedland’s talk will discuss the Halo vision and methodology and provide an overview of the pilot phase and its challenge. Friedland will also outline a road map for future phases of their initiative.
University of Colorado Health Sciences Center
AI and Molecular Biology: A Growing Success Story
Applications of artificial intelligence in molecular biology have long been scientifically important and are now growing in industrial significance as well. In application areas as diverse as drug discovery and health care cost containment, AI techniques from machine learning, knowledge representation, and now natural language processing are becoming increasingly part of mainstream commercial practice. In this lecture, Hunter describes some current success stories and outlines the potential for growth.
Luis von Ahn
Carnegie Mellon University
CAPTCHA: Telling Humans and Computers Apart Automatically
A CAPTCHA is a program that can tell whether its user is a human or a computer. More specifically, a CAPTCHA is a program that can generate and grade tests that humans can pass, but current computer programs cannot. CAPTCHAs have several applications in practical security, and some results of this project are currently in use by Yahoo, Hotmail, and other major websites. CAPTCHAs also provide concrete, well-defined challenges to the Artificial Intelligence community. Ahn describes the CAPTCHAproject in detail and surveys the many issues related to AI.