Maintenance tools for knowledge-based systems: the MAKE project

1999

Drafting laws is not just translating do's en dont's into obligations and prohibitions. The complex structures of interrelated acts and sections have to be modelled as well. Drafting legislation exists of numerous additions and deletions of texts. When a legal drafter adds an obligation, the alteration may effect the boundaries of a prohibition. This paper discusses modelling norms, and describes methods for regulation drafting, repair and refinement. Each version is tested on a legal knowledge based system in order to compare with the desired behaviour. In this paper the design of an intelligent toolset is presented, both for drafting and refinement, as well as for application simulation.

International Journal of Intelligent Systems, 1994

Verification and validation (V&V) of Knowledge Bases (KBs) are two sides of the same coin: one is intended to assure the structural correctness of the KB, while the other is intended to assure the functional correctness of the domain model embodied in the KB. Knowledge base refinement aims to appropriately revise the KB if a structural or functional error is detected during the V&V process. This paper presents a uniform framework for verification, validation and refinement of KBs represented as sets of production rules, called the VVR system. It incorporates a contradiction-tolerant truth maintenance system (CTMS) for performing both verification and validation analyses, and some simple explanation-based learning techniques for guiding the refinement process. Verification analysis consists of detecting and correcting the main types of structural anomalies: circular rules, redundant rules, inconsistent rules, and inconsistent data, and checks the KB for completeness and violated semantic constraints. In terms of validation, given a set of test cases, the VVR system is capable of detecting and correcting functional errors caused by overgeneralization and/or overspecialization of the KB. If the set of test cases is not available, the VVR system can generate synthetic test cases intended to help the user evaluate KBS performance.

The use of knowledge representation systems for building large complex ontologies to support real-world problems has been demonstrated in various application areas. One such system is Ford's Direct Labor Management System (DLMS) that has been used since 1990 in the very dynamic domain of process planning for vehicle assembly. The long-term maintenance of the DLMS knowledge base has demonstrated both the flexibility and reliability of semantic networkbased approaches in a rapidly changing industrial setting. This maintainability becomes very difficult over time due to changes in all of the following areas: the external business environment, the processes and physical concepts being modeled, and the underlying hardware and software architecture.

Proceedings of the fifth international conference on Artificial intelligence and law - ICAIL '95, 1995

Io this paper, we argue that a representation of legal knowledge should remain completely independent from aspects of the application systems that employ it. This makes it possible to provide isomorphic representations, makes maintenance of the knowledge bases much easier and less error-prone, and is a prerequisite for reusing and sharing the legal knowledge among several knowledge-based systems. We suggest a hybrid knowledge representation approach which integrates a legal representation component with the representation component of an application system by coupling them via their common terminology. This is always possible because the domains of discourse of both representation components necessarily overlap. Sometimes, such an integration does not comply with the control flow requirements of the application system. We propose a solution to this problem, too, which introduces a one-sided dependency only and leaves the knowledge bases to be reused (i.e., the terminology and the legal knowledge) unchanged. We illustrate our approach by discussing its application for the integration of a legal knowledge base with a knowledgebased decision support system.

Applied Logic Series, 1999

We present EULE2, a knowledge-based decision support system for o ce tasks. After motivating the need for such a system, we sketch its functionality and discuss the three main knowledge sources it makes use of: terminological knowledge, knowledge about actions, and knowledge about federal law and company regulations. The terminological knowledge provides all the concepts relevant in the domain of discourse, and allows to introduce instances of those concepts. Knowledge about actions serves two purposes: to determine what changes have to be made when executing an action, and to decide what actions can or cannot be performed at a certain point in an o ce task. A representation formalism that satis es both, quite heterogeneous requirements is described in detail. Knowledge about federal law and company regulations is needed to ensure that all o ce tasks are properly executed { this is a main functionality o ered by EULE2. Subsequently, the paper deals with how to integrate knowledge about law and regulations with the knowledge about actions. We propose a hybrid integration approach that keeps the representation of both kinds of knowledge independent from each other, but ensures that the reasoning takes all relevant pieces of knowledge into account. We argue that the resulting, hybrid integration approach drastically increases maintainability and reusability of knowledge bases as compared to former approaches.

1998

Abstract Knowledge maintenance (KM) is the process of reflecting over some current knowledge based system in order to handle a new situation. Changes that fix new problems must not introduce bugs into old fixes. A common technique for KM is to reflect over dependencies between concepts. In this review, we sample the KM work on modeling dependencies using either a logic-based, a network-based, or a procedural-based approach.

Knowledge-Based Systems, 1989

The development and use of a Knowledge Dictionary, a tool to facilitate the documentation and maintenance of rule based expert systems, is discussed. The Knowledge Dictionary may be used to record heuristics and their component parts, facts and rule actions, in such a way that a knowledge engineer, or end user, may determine the usage of any part of the knowledge, may easily add new parts, and may run the expert system to determine the effect of the maintenance. The Knowledge Dictionary utilizes the relational data model to store the heuristics in a data form rather than in an executable code form. Use of the relational model provides the knowledge engineer with all the power of relational calculus to interrogate the stored knowledge.

The human brain can store countless folds of knowledge. Still, we cannot fully utilize a single human brain to solve a specific problem. Knowledge-based systems (KBSs) are computer programs specifically developed to perform problem solving like human experts. These systems effectively expand efficiency and flawlessly solve problems in various fields. The structure of a KBS can be divided into five standard components, which are described in detail. The component used for storage is called a knowledgebase, while an inference engine is a software module that processes the knowledge stored in the knowledgebase. Different tools, shells and programming languages can be used to develop and utilize KBSs. The most popular development languages are List Programming (Lisp), Prolog, Java Expert Systems Shell (JESS) and C Language Integrated Production System (CLIPS). This paper highlights various known issues in the phases of the development, deployment and maintenance of KBSs. Developers, as well as end users, may face various issues in the life cycle of these systems. The key personalities involved in the development and maintenance of KBSs are knowledge engineers and domain knowledge experts. Most issues are related to the experts’ availability, their behavior and long-term cooperation

1993

Abstract 1 In most real-world applications, knowledge bases grow and change over time. This paper examines two solutions to the problem of knowledge base consistency maintenance as new knowledge is acquired and assimilated in an evolving knowledge base.

1993

As a result of a large-scale invasion of expert systems (ES) in almost all areas of human life, a major problem related to their quality and reliability was identified. This paper presents a possible solution to this problem which consists of extending the functionality of expert systems with a built-in quality assurance facility to allow the ES developer constantly to keep track of the most important structural and functional characteristics of the model.