Verification of Knowledge Bases: a Unifying Logical View

Notions of correctness and completeness of a KB are impossible to capture completely by a formal deenition. However, when the knowledge base is represented in a declarative logical formalism, they can be approached by a logical analysis of its contents. A logical analysis of the knowledge base and constraints that are known to hold on the domain enables us to detect anomalies or discrepancies between the knowledge represented in the KB and the domain. This paper describes a uniied logical framework for the veriication problem of knowledge bases represented by logical rules (i.e., Horn rules and some extensions). We consider several instances of the veriication problem, describe algorithms for veriication, and establish the computational complexity of the veriication problem. In particular, we consider the veriication w.r.t. consistency constraints that are speciied separately on the inputs and on the outputs. Next, we consider dependency constraints on the relations between inputs and outputs. Finally, we consider several extensions to our logical formalism that enable us to express richer classes of constraints. The extensions involve the use of negation and the use of constructors of description logics in the constraints.

[1]  Yehoshua Sagiv,et al.  Optimizing datalog programs , 1987, Foundations of Deductive Databases and Logic Programming..

[2]  R. Van der Meyden,et al.  The complexity of querying indefinite data about linearly ordered domains , 1992, PODS '92.

[3]  Keith E. Williamson,et al.  Inconsistency and redundancy checking for quasi-first-order-logic knowledge bases , 1993 .

[4]  Ashok K. Chandra,et al.  Optimal implementation of conjunctive queries in relational data bases , 1977, STOC '77.

[5]  Anthony C. Klug.  On conjunctive queries containing inequalities , 1988, JACM.

[6]  Dan Suciu,et al.  Deciding Containment for Queries with Complex Objects and Aggregations , 1997, PODS 1997.

[7]  Marie-Christine Rousset,et al.  On the consistency of knowledge bases: the COVADIS system , 1988, Comput. Intell..

[8]  Christine Pierret-golbreich,et al.  Correctness of Methods w.r.t. Problems Specifications , 1996 .

[9]  Alon Y. Halevy,et al.  Queries Independent of Updates , 1993, VLDB.

[10]  Marc Ayel,et al.  SYCOJET and SACCO, two tools for verifying expert systems , 1993 .

[11]  Allen Ginsberg.  Knowledge-Base Reduction: A New Approach to Checking knowledge Bases for Inconsistency and Redundancy , 1988, AAAI.

[12]  Mihalis Yannakakis,et al.  Equivalence among Relational Expressions with the Union and Difference Operation , 1978, VLDB.

[13]  Oded Shmueli,et al.  Equivalence of DATALOG Queries is Undecidable , 1993, J. Log. Program..

[14]  Jeffrey D. Ullman,et al.  Principles of Database and Knowledge-Base Systems, Volume II , 1988, Principles of computer science series.

[15]  M. Rousset,et al.  Formal verification of knowledge bases focused on consistency: two experiments based on ATMS techniques , 1993 .

[16]  Surajit Chaudhuri,et al.  On the equivalence of recursive and nonrecursive datalog programs , 1992, J. Comput. Syst. Sci..