KRDB 1998

Invited Talks

KRDB'98 will feature (at least) the following two invited talks.

OKBC: A Programmatic Foundation for KB Interoperability

by

Prof. Richard Fikes, Stanford University

Open Knowledge Base Connectivity (OKBC) is an application programming interface for accessing knowledge bases stored in knowledge representation systems (KRSs). OKBC is being developed under the sponsorship of DARPA's High Performance Knowledge Base program (HPKB), where it is being used as an intial protocol for the integration of various technology components.

OKBC is a successor of Generic Frame Protocol (GFP) which was primarily aimed at systems that can be viewed as frame representation systems and was jointly developed by Artificial Intelligence Center of SRI International and Knowledge Systems Laboratory of Stanford University.

OKBC provides a uniform model of KRSs based on a common conceptualization of classes, individuals, slots, facets, and inheritance. OKBC is defined in a programming language independent fashion, and has existing implementations in Common Lisp, Java, and C. The protocol transparently supports networked as well as direct access to KRSs and knowledge bases.

OKBC consists of a set of operations that provide a generic interface to underlying KRSs. This interface isolates an application from many of the idiosyncrasies of a specific KRS and enables the development of tools (e.g., graphical browsers, frame editors, analysis tools, inference tools) that operate on many KRSs. It has been successfully used in several ongoing projects at SRI and Stanford University.

From Knowledge Reformulation to Data Warehouse Design

by

Prof. Alon Levy, University of Washington at Seattle

Designing an appropriate representation of an application domain is a critical step in building a knowledge-based application. Similarly, designing an appropriate database (or data warehouse) schema is a key to obtaining efficient performance of a database application. One of the questions addressed by the Artificial Intelligence community almost from its inception is that of problem reformulation. That is, how can a system automatically reformulate a representation of the domain in order to yield better performance for some specific tasks. Reformulations have been considered in a variety of problem-solving settings including automatic programming, constraint satisfaction, design, diagnosis, machine learning, planning, qualitative reasoning, scheduling and theorem proving.

The design of data warehouses poses another instance of the reformulation problem. Broadly speaking, a data warehouse contains a set of views over an operational database. The challenge in warehouse design is to select a set of views such that a given workload of decision support queries can be answered efficiently from the warehouse.

I will discuss the problems of reformulation and data warehouse design from a common perspective. I'll contrast the emphases made in these two lines of research, and examine what each area has to offer the other. I'll touch in detail upon issues of efficiently updating views, rewriting queries using views, and selection of views for a data warehouse.