finin@prc.unisys.com (Tim Finin) (03/10/90)
CALL FOR PARTICIPATION AAAI-90 Workshop on Knowledge Base Management Systems Boston Hynes Center July 29, 1990 Large knowledge based systems share many of the characteristics of both databases and large software systems. Knowledge base systems and languages need the same sort of capabilities as existing database management systems offer, including persistence, concurrent update, shared access, and large scale. However, the knowledge based setting presents unique requirements which must be addressed, such as a more expressive representation language and a more sophisticated consistency checking capability. At the same time knowledge bases are, in many respects, like large software systems. This leads to requirements for better tools and environments to support their development, validation and maintenance. This workshop will address the issues involved in building knowledge base management systems with a focus on two areas. The first focus area deals with supporting the development, maintenance, and evolution of large knowledge bases. How does the knowledge base development cycle differ from the traditional software development cycle? What are the difficulties in supporting the multi-paradigm environments in which many knowledge based systems are built? Is knowledge base development a design task to which the work in design databases can be applied? The second focus area concerns coupling knowledge bases and databases. Many times it is necessary for a knowledge based system to query or update an existing database which is used by many other applications. In this case the knowledge based system must identify the set of relevant data from the DBMS on which to base its decisions. What happens if this set of data is very large or changing quickly? What modifications must be made to a database management system in order to allow efficient access by knowledge based applications? Are the requirements of knowledge based systems unique in this regard? Another coupling of knowledge bases and databases occurs when a database is used as a backend to store a knowledge base. A third approach is to integrate an inference engine into a DBMS so that a portion of the knowledge based system is moved inside the DBMS. Are either of these two approaches likely to prove helpful in building real applications? This workshop will be open to developers and users interested in these issues. Each participant is asked to submit a two to five page position paper addressing these or closely related issues. Attendance is by invitation only and is limited to roughly 30 participants. About half of the participants will be asked to make presentations over the course of the day. The position papers will be distributed as working notes to all participants and a summary of the workshop will appear in AI Magazine. There will be a $50 registration fee. Send five copies of your position paper by April 15th to: Eric Mays IBM T.J. Watson Research Center P.O. Box 218 (Rt. 134 & Taconic) Yorktown Heights, NY 10598 Phone: 914-945-3587 IMPORTANT DATES: o April 15 - Position papers due o May 15 - Participants determined and notification sent o June 15 - Revised position papers due o June 30 - Schedule and working notes sent to all participants o July 29 - Workshop held at Boston Hynes Center (AAAI is 7/29 - 8/3) ORGANIZING COMMITTEE: o Eric Mays, IBM T.J. Watson Research Center (emays@ibm.com) o Tim Finin, Unisys Center for Advanced Info. Technology (finin@prc.unisys.com) o Doug Lenat, MCC (lenat@mcc.com) o Michael Stonebraker, UC Berkeley (mike@berkeley.edu) -- Tim Finin finin@prc.unisys.com Center for Advanced Information Technology 215-648-2840, 215-648-2288 (fax) Unisys, PO Box 517, Paoli, PA 19301 215-386-1749 (home)