DAI-List Digest Wednesday, 9 October 1991 Issue Number 58 Topics: Reminder - CFP for 11th DAI Workshop DARPA Knowledge Sharing Effort Please send submissions to DAI-List@mcc.com. Send other requests, such as changes in your e-mail address, to DAI-List-Request@mcc.com. ------------------------------------------------------------------------ Date: Fri, 4 Oct 91 11:14:17 -0400 From: Ed Durfee Subject: It's around the corner.... Call for Participation 11th International Workshop on Distributed Artificial Intelligence The Homestead Glen Arbor, Michigan (near Traverse City) February 26-28, 1992 Distributed artificial intelligence (DAI) is concerned with how a decentralized group of intelligent computational agents should coordinate their activities to achieve their goals. When pursuing common or overlapping goals, they should act cooperatively so as to accomplish more as a group than individually; when pursuing conflicting goals, they should compete intelligently. The 11th Workshop on Distributed Artificial Intelligence will bring together researchers studying coordination from diverse perspectives, ranging from distributed computing systems to discourse analysis, from formalisms for representing nested beliefs in agents to cognitive studies of human performance in organizations, from solving inherently distributed problems in applications such as communication network managment to analyzing the evolution of cooperation in populations of artificial systems. The objective is to identify and develop common theories and approaches to coordination in order to improve our conceptual understandings and practical implementations of coordination processes. Participation at the Workshop will be by invitation only and will be limited to approximately 40 people. To participate, please submit a technical paper describing original research or significant applications in DAI to the Workshop chairman. Preference will be given to work that addresses one or more of the four DAI themes listed below. We specifically discourage the submission of papers in areas such as fine-grained parallelism, hardware or language-level concurrency, and connectionism, because we feel that work in these areas is more appropriate for other workshops. A small number of "interested observers" may also be invited to attend. If you would like to be considered for attendance on this basis, please submit a written request justifying your participation. To encourage participants to relate their work to ongoing themes in DAI beforehand, papers are solicited for (but not strictly limited to) the following themes: 1. Coordination/Collaboration Knowledge: The identification, encoding, and use of generic knowledge for coordination and collaboration. This theme focusses on general knowledge about resolving conflicts, compromising, and cooperating. 2. Coordination as Search: When viewing coordination as a search process, decisions are needed regarding algorithms for conducting the search, heuristics for controlling the search, and protocols for exchanging and updating portions of the search space. This theme broadly includes approaches such as distributed constraint satisfaction search, search for compatible distributed plans, search in cooperative problem-solving and design, negotiation search, and search for appropriate organizational designs. 3. Intelligent Agents in Enterprises and Applications: Embedding DAI systems in computer networks used by people to solve problems allows the automation of both cooperative problem-solving activities (such as distributed interpretation or diagnosis) and coordination activities (such as information filtration or resource allocation). This theme includes issues in identifying suitable applications of DAI technology and in developing DAI agents that interact effectively with people and each other. 4. Modeling Through Communication and Observation in Adversarial and Cooperative Systems: Building and maintaining models of other agents' beliefs, abilities, goals, and plans is crucial for intelligent interaction. Topics in this theme include acquiring modeling information (through communication and plan recognition) and using models to make decisions about communication (deciding whether to tell the truth, eliciting more information) and about other actions. These themes are not mutually exclusive, and we welcome papers that integrate insights from more than one of them. As DAI matures, it is appearing more and more in real-world applications. This welcome development raises the need for engineering principles that will help match particular techniques with kinds of problems. We welcome both theoretical and applied papers, and encourage each to contribute to the development of these principles. Specifically, theoretical papers should explain how their principles and methods can be mapped to applications, while applied papers should explain why they use the techniques that they do and why other approaches are less appropriate for the problem at hand. The Workshop will have several types of sessions, including presentation sessions for discussion of distinguished papers, panel discussions for papers sharing common themes, and working sessions on topics such as unifying concepts behind coordination, and methodologies and evaluation criteria for DAI research. VENUE: DAI'92 will be held at the Homestead in Glen Arbor, Michigan, 27 miles from the Traverse City airport, on the shore of Lake Michigan and adjacent to the Sleeping Bear Dunes National Lakeshore. Sculpted by glaciers and hundreds of feet in height, the dunes border the western edge of the peninsula the Chippewa called "Leelanau," "the land of delight." The environment is richly forested and surrounded with relics of frontier America, and the resort has on-site ice skating, down-hill skiing, and cross-country trailheads to 36 kilometers of trails through the National Lakeshore (with equipment rental and instruction available). We'll continue the DAI tradition of a participatory workshop by active practitioners in a setting that offers seclusion, natural beauty, and recreational intermissions. SUBMISSION DETAILS: Papers for review should be a maximum length of 10 pages, in a legible format. Please submit 4 copies to Edmund H. Durfee (address below), and indicate on the title page the theme(s) for which the paper is most relevant. Also, please include an electronic mail address for the appropriate contact person along with the submission. DATES: Deadline for paper submissions (4 copies, 10 page max): November 1, 1991. Notification of acceptance: December 16, 1991. Final papers due (for distribution at the Workshop): January 17, 1992. We expect that revised versions of the best papers from the Workshop will be considered for inclusion in an appropriate journal or in a published collection. PROGRAM COMMITTEE: Edmund H. Durfee (co-chair) H. Van Dyke Parunak (co-chair) EECS Department Industrial Technology Institute University of Michigan P. O. Box 1485 1101 Beal Ave. Ann Arbor, MI 48106 Ann Arbor, MI 48109 (313) 769-4049 (313) 936-1563 van@iti.org durfee@caen.engin.umich.edu Susan Conry, Clarkson University Jacques Ferber, Universite Paris Mike Huhns, MCC Jeff Rosenschein, Hebrew University Katia Sycara, Carnegie Mellon University Marty Tenenbaum, Stanford University Robert Weihmayer, GTE Labs ADVISORY COMMITTEE: Mark Fox, Carnegie Mellon University Les Gasser, Paris IV/Ecole des Mines, France Carl Hewitt, MIT Toru Ishida, NTT, Japan Victor R. Lesser, University of Massachusetts at Amherst Eric Werner, University of Hamburg, Germany ------------------------------------------------------------------------ Date: Wed, 25 Sep 91 10:44:40 EDT From: finin@cs.umbc.edu Subject: Knowledge Sharing Several people have asked me for more information on the DARPA sponsored Knowledge Sharing Effort I talked about recently. Here is a brief overview. Tim THE KNOWLEDGE SHARING EFFORT The Knowledge Sharing Effort is a consortium to develop conventions facilitating sharing and reuse of knowledge bases and knowledge based systems. The goal of the effort is to define, develop, and test infrastructure and supporting technology to enable participants to build much bigger and more broadly functional systems than could be achieved working alone. The output of the effort consists of (1) public-domain specifications and implementations of supporting technology; (2) reports, papers, and technical articles; (3) a reusable public library of proof-of-concept demonstrations. Building new knowledge-based systems today usually entails constructing new knowledge bases from scratch. It could instead be done by assembling together reusable components. System developers would then only need to worry about creating the specialized knowledge and reasoners new to the specific task of their system. Their new system would interoperate with existing systems, using them to perform some of its reasoning. In this way, declarative knowledge, problem solving techniques, and reasoning services could all be shared among systems. This would facilitate building bigger and better systems cheaply. The infrastructure to support such sharing and reuse would lead to greater ubiquity of these systems, potentially transforming the knowledge industry. Definition of conventions enabling sharing among collaborators is the essential first step toward these goals. Technical analyses of knowledge representation technology indicated four complementary areas in which development of common, agreed-upon conventions would enhance leverage between individual research efforts. (These analyses were initiated in a three-day workshop involving over 40 top AI and database researchers from DARPA, NSF, and industry communities.) The Knowledge Sharing Effort is organized around working groups addressing each of these four areas. They are: INTERLINGUA: concerned with translation between different representation languages, with sub-interests in translation at design time and at run-time KRSS (Knowledge Representation System Specification): concerned with defining common constructs within families of representation languages EXTERNAL INTERFACES: concerned with run-time interactions between knowledge based systems and other modules in a run-time environment, with sub-interests in communication protocols for KB-to-KB and for KB-to-DB SHARED, REUSABLE KNOWLEDGE BASES: concerned with facilitating consensus on contents of sharable knowledge bases, with sub-interests in shared knowledge for particular topic areas and in topic-independent development tools/methodologies. As of the end of summer, 1991, the effort had completed the first year of its program. In that first year, an overall model was formulated for the architecture of large knowledge-based systems and the envisioned development methodology for producing such systems. The four working groups were established and began work on their respective tasks. The groups working on translation, common languages, and communication protocols have each produced draft specifications; these are being circulated within the groups for refinement and will be disseminated to the rest of the AI community over the coming year. The group on ontology libraries has been concentrating on experiments to foster a better understanding of requirements; one such experiment has been conducted in the first year, three additional experiments are planned for the second. More specifically, the Knowledge Sharing Effort has in the first year: - Defined a very general and very flexible model for the architecture of large, knowledge-based systems that allows for a development paradigm of composing these systems out of smaller, reusable modules assembled either at design-time or at run-time. - Documented the model in an article to appear in the fall issue of AI Magazine. - Developed a draft specification for a knowledge interchange format, a mechanism for avoiding the cost of building translators between arbitrary languages by providing an intermediate language suitable for bi-directional translation between it and the others. - Developed a small testbed knowledge base and used it to demonstrate the viability of translation into the knowledge interchange format from three different representation languages. Global goals for the upcoming year are to: - Publish specifications for the knowledge interchange format and for a common frame-based, object-oriented terminological representation language. - Disseminate a document on application interface and intermodule communication protocols. - Establish a set of test suites in an on-line generic knowledge base library. - Perform several experiments on collaborative, multisite knowledge-based system construction utilizing the knowledge sharing conventions developed in this effort. - Publish reports from all working groups. The Knowledge Sharing Effort currently involves participants from over a dozen different research centers around the United States, as well as a small number of centers abroad. A representative list of participating research centers includes AT&T, Carnegie-Mellon University, EITech, Hewlett-Packard Research Center, Lockheed, Northwestern University, University of Saarbrucken, USC / Information Sciences Institute, Stanford Research Institute, Stanford University, Unisys Center for Advanced Information Technology, University of Maryland, University of Texas, and Xerox Palo Alto Research Center. For further information, contact: Robert Neches or Ramesh Patil USC / Information Sciences Institute 4676 Admiralty Way Marina del Rey, CA 90292 213 / 822-1511 (after Nov. 5, 310/822-1511) Neches@isi.edu