Reports on the AAAI Fall Symposia

Article excerpt

The 1999 American Association for Artificial Intelligence Fall Symposium Series was held Friday through Sunday, 5-7 November 1999, at the Sea Crest Oceanfront Resort and Conference Center. The titles of the five symposia were

* Modal and Temporal Logics-Based Planning for Open Networked Multimedia Systems

* Narrative Intelligence

* Psychological Models of Communication in Collaborative Systems

* Question-Answering Systems

* Using Layout for the Generation, Understanding, or Retrieval of Documents

This article concludes with a previously unpublished report on the 1998 AAAI Fall Symposium on AI and Link Analysis.

Modal and Temporal Logics-Based Planning for Open Networked Multimedia Systems

This symposium provided a forum for researchers involved in using formal methods and in design of networked multimedia systems and adaptive-reactive systems to identify common ground, relevant experiences, applications, open problems, and possible future developments.

To support intelligent and interactive multimedia applications, there's a need to tailor systems to possess and use knowledge about the application domain, user-requirement tasks, the context of interaction, communication, and performance parameters. Specific applications addressed concern video on demand, multimedia presentations, multimedia synchronization protocols, multimedia collaboration, electronic commerce, nomadic services, and so on. Temporal and modal logics have been used to reason about time, action, and adaptive change and to program and verify networked systems.

Several topics were presented and discussed: (1) modal temporal-logic-based management of open distributed-processing systems and active networks; (2) modal-temporal-logic specifications of service quality in multimedia networks; (3) combination of temporal-modal logics with formal languages such as LOTOS and CCS; (4) design of modal-temporal logics: operational models and implementation techniques, programming support and environments, and comparative studies of languages; (5) specification and verification technology: languages and tool support, model checking, and so on; and (6) software processes based on formal methods: design, analysis, verification, refinement, and so on.

Fawzi Daoud GMD

Narrative Intelligence

People are narrative animals. By telling stories, we make sense of the world. We order its events and find meaning in them by assimilating them to more or less familiar narratives. A rich research tradition in a variety of fields suggests that narrative is fundamental to human intelligence. At the Narrative Intelligence Symposium, over 50 researchers gathered to discuss systems at the confluence of AI and narrative, systems that support, model, or investigate human narrative intelligence.

A wide variety of systems were presented: story generation, interactive fiction (including the first public demonstration from Joseph Bates's company Zoesis), systems that support storytelling by children and adults, agents as characters in narratives, agents that can tell stories, and more. These systems were built by researchers working at the intersection of AI and a variety of humanistic disciplines, including literary and cultural theory, art, fiction writing, psychology, and learning theory. Narrative was understood by these researchers from a variety of perspectives, including narrative as story, folk tale, small talk, drama, and world view. Despite this heterogeneity of application areas and approaches, we quickly developed a sense of community and common purpose as we realized we were addressing a set of common questions:

How can we create characters from which interactive narrative emerges? How can we create systems that can understand and generate compelling narratives? How can we build up narrative from a collection of elements, whether language, images, or plot? …