Artificial Intelligence in Knowledge Management
Knowledge management (KM) is a topic of growing interest to large organizations. It comprises activities focused on the organization, acquiring knowledge from many sources, including its own experience and that of others, and on the effective application of this knowledge to fulfill the mission of the organization.
The KM community has been eclectic in drawing from many sources for its methodologies and tools. Typical approaches to the management of knowledge are based on concept maps, hypermedia, and object-oriented databases. Techniques developed in AI for knowledge acquisition, representation, and discovery are seen as relevant to KM. However, as yet, KM has no unified underlying theory, and the scale of the problem in large organizations is such that most existing AI tools cannot be applied in their current implementations.
The objective of this symposium was to bring together KM practitioners and applied AI specialists from knowledge acquisition, knowledge representation, and knowledge discovery in databases and attempt to formulate the potential role of various AI subdisciplines in KM. Those attending represented a wide range of industries and areas relevant to AI research and application.
The symposium began with keynote addresses on industrial requirements for KM by Vince Barabba of General Motors and Rob van der Spek of CIBIT. The remainder of the meeting was devoted to intensive group discussion of the role of AI in KM, including a joint session with the Symposium on Ontological Engineering. The sessions were entitled Organizational Knowledge Management, Work-Flow Management, Knowledge Management through Hypermedia and Text Processing, Cognitive Aspects of Knowledge Management, Agents and Multiactor Systems, Knowledge Representation and Reasoning, and Knowledge Discovery.
The accepted papers were made available to participants in advance of the meeting through the World Wide Web and will remain available at ksi.cpsc.ucalgary.ca/AIKM97.
B. R. Gaines University of Calgary, Canada
Mark A. Musen Stanford University
Ramasamy Uthurusamy General Motors
Computational Models for Mixed-Initiative Interaction
This symposium was held as a direct result of discussions at the National Science Foundation Interactive Systems Program Grantees Workshop that was held in Cambridge, Massachusetts, in November 1995. Mixed-initiative interaction is a notion that has concerned members of the interactive systems community for some time, although there is no established vocabulary for discussing it and no established method for achieving it. An important goal of this symposium was to define initiative and consider examples of how systems currently support it.
The symposium brought together researchers in natural language processing, planning, robotics, interactive tutoring systems, and human-computer interfaces. During the symposium, discussion transitioned from theoretical issues to demonstrations of working systems that exhibit varying degrees of initiative. On the first day, participants attempted to define concepts such as having control or taking the initiative. Participants also addressed practical questions such as whether domain task models include enough information to reason about initiative and control or whether discourse-level information is needed. On the second day, participants examined the role that mechanisms such as scripts, constraint satisfaction, blackboards, and top-down planning can play in monitoring and managing an interaction. On the third day, participants watched and discussed demonstrations of interactive systems.
Participants concluded that (1) mixed-initiative behavior is required in any system that aims to tutor users or help them solve problems collaboratively and (2) the problem of getting systems to demonstrate mixed-initiative interaction is a serious one. For some, achieving initiative requires including explicit reasoning about initiative as part of the design, but for others, it can be (and is) an emergent property of the architecture. …