Knowledge Engineering for Industrial Expert Systems

The following commentaries on the chapters in this volume start with a few humble premises. The realization is inescapable that the enterprise of psychology is vast, has a long history, and is deeply divided into often non-communicating subfields. The enterprise of artificial intelligence (AI) is much younger than psychology and yet is also divided into areas that barely communicate with each other. The divisions in AI often parallel those in psychology and often seem to be motivated by similar theoretical and methodological concerns. In addition, the enterprise of expert system construction, and in particular knowledge acquisition (KA), is but one tiny corner of AI, yet the activities focused on expertise are expanding beyond the narrow confines of computer science. It is as if a new field concerning itself with expertise, knowledge, and expert system construction—a field as yet unnamed—is trying to be born. It is an event exciting to witness, but it leaves the poor commentator with no fixed or generally agreed-upon platform—no firm Newtonian coordinate system from which to survey and summarize multidisciplinary research.

Building Expert Systems: A Structured Bibliography

One of the greatest challenges in building an expert system is obtaining, representing, and programming the knowledge base. As the size and scope of the problem domain increases, knowledge acquisition and knowledge engineering become more challenging. Methods for knowledge acquisition and engineering for large-scale projects are investigated in this paper. The objective is to provide new insights as to how knowledge engineers play a role in defining the scope and purpose of expert systems and how traditional knowledge acquisition and engineering methods might be recast in cases where the expert system is a component within a larger scale client-server application targeting multiple users.

A New Approach In Learning For Intelligent Multi Agent Systems

Clinical pathology tests generate large volumes of data for critical care clinicians, and ”data overload”, can impede clinical decision making. Bedside arterial blood gas (ABG) analysis is compounding this problem. Expert systems (ES) have great potential to support clinicians' work, but few ES are in routine use. The complexity of knowledge acquisition (KA) leads to poor productivity in the development phase and requires specialist knowledge engineering (KE) skills. This has been termed the knowledge acquisition “bottleneck”. We have devised a novel algorithm, ripple-down rules (RDR), that essentially eliminates the KE bottleneck. The Pathology Expert Interpretative Reporting System (PEIRS), an RDR ES, interprets a broad range of pathology reports (including thyroid function tests, ABG, cardiac enzymes and a range of hormone assays) and has been in routine use for over 4 years. With over 2100 rules, PEIRS is one of the few large medical ES in routine use, and apparently the only large ES built entirely by a pathologist without specialist KE skills or support. A key finding from PEIRS was that its simple KA strategy enabled localisation of its knowledge, e.g. local protocols for interpretation and management. This localisation was essential for gaining the acceptance of the system by the pathologists.

However expert systems (ES) have been in use since the early eighties, but there is a lack of strong theoretical base for handling knowledge elicitation problems. More, the findings of different empirical investigations on the outcomes of ES use are often contradictory. Consequently, there is a requirement in the ES field for theories or explanatory models to formulate propositions, to conduct research and interpret findings in a coherent way. This article reports briefly on a particular part of a work conducted at the University of Lund to investigate the expert system development problems and motives. According to the results of a survey conducted between April 1989–December 1989, seven various knowledge elicitation techniques were identified. The study showed that the most popular elicitation technique used for extracting knowledge from human expert was the structured interview. In this study an attempt is made to outline a conceptual model of knowledge elicitation and puts forward a number of propositions and suggestions which can contribute to our knowledge of expert systems development. The conceptual model is based on non-technical prescriptive guide. The model may allow one to conceptualize KE process in a way that to implement an expert system project more successfully. Furthermore, the result of this study can provide a basis for further research in KE. Because, looking across many of the reports in which knowledge elicitation methods are described, one may find much advice, but few data. One direction that future research might take is, to compare software engineering techniques with knowledge engineering techniques.

The paper considers medical knowledge extraction for the diagnosis of rare hereditary diseases. The specific feature of the proposed approach is the combination of expert estimates of the symptom presence probability with three complementary factors of confidence: time of manifestation, severity and frequency of symptom manifestation in different age ranges. Preliminary textological cards used by experts and cognitive scientists are formed based on the linguistic analysis of related literature.

A method of automating the knowledge acquisition task for process plant expert systems

Eliciting Knowledge from Experts: A Methodological Analysis

This paper will discuss some of the issues involved in building an Expert System that embodies tuning rules for IBM's MVS/XA operating system. To understand the components of an Expert System and their functions, PROLOG on an IBM PC (Turbo-PROLOG from Borland International) was chosen as the development environment. The paper will begin by defining the key concepts about Expert Systems, Knowledge Engineering, and Knowledge Acquisition. The reader will be given a brief overview of PROLOG, from which we can explain how an inference mechanism was developed. Finally, the paper will describe the Expert System that was developed, and additionally will provide a set of key issues that should be addressed in the future. It is our overall objective to provide new insight into the application of AI to CPE.

Expert systems are a well-known and well-received technology. It was thought that the performance of a domain expert could not be duplicated by a machine. Expert systems technologies have shown this to be a false belief, and indeed have demonstrated how experts themselves can come to depend on expert systems. Expert systems enjoy widespread use in industrial domains and further uses are planned. The growth in acceptance has been explosive since about 1986. Continued rampant growth appears to depend on cracking the so-called knowledge acquisition bottleneck.The knowledge acquisition bottleneck limits the scalability of expert systems. While it is relatively straightforward to populate a small-scale knowledge base, it becomes more difficult to maintain consistency and validity as the knowledge base grows. Thus, it is important to automate the knowledge acquisition process. A by-product of this process is that any failure of the expert system will be “soft.”The question is, “What impact can machine learning technologies have on knowledge acquisition in the large?” The true test will be on prospective industrial applications in areas such as biology, education, geology, medicine, and scientific discovery. Machine learning technologies include expert systems, genetic algorithms, neural networks, random seeded crystal learning, or any effective combinations.Relevant subtopics include: Second generation expert systems — progress and prognosisRepertory GridsThe importance of symbolic and qualitative reasoningThe acquisition of fuzzy rulesThe best learning paradigm or combination of paradigmsImpact of machine learning on explanation systemsThe role of toy domains such as chessAutomatic programming revisitedApplications to computer vision, decision support systems, diagnosis, helpdesks, optimization, planning, scheduling, et al.Implementation issues using SIMD and MIMD platformsSources for joint sponsorshipForming industrial partnershipsForming alliances abroad

Focuses on Korean garment manufacturing and the feasibility and benefits of introducing an expert system. Reviews previous literature on knowledge‐based system and describes in detail the stages followed in developing, constructing, testing, evaluating, integrating, implementing and maintaining the expert system. Concludes that an expert system could be used to solve various garment manufacturing problems and would contribute to garment quality improvement through a standardized apparel production process.

SummaryKnowledge acquisition for expert systems typically is a tedious, iterative process involving long hours of consultation between the domain experts and the computer scientists who serve as knowledge engineers. For well-understood domains, however, it may be possible to facilitate the knowledge acquisition process by allowing domain experts to develop and edit a knowledge base directly. Administration of protocol-directed cancer chemotherapy is such a well-understood application area, and a knowledge acquisition system, called OPAL, has been developed for eliciting chemotherapy-protocol knowledge directly from expert oncologists. OPAL’s knowledge acquisition approach is based on the interactive graphics environment available on current generation workstations. The use of graphics improves the interface by reducing typing, avoiding natural language interpretations, and allowing flexibility in entry sequence. The knowledge in OPAL is displayed using an arrangement of hierarchically related, graphical forms. The position of a particular form in the hierarchy defines the context of the knowledge contained in the form. Intelligent editing programs such as OPAL can streamline the knowledge engineering process for highly structured domains requiring repetitive knowledge entry.

Knowledge acquisition based on explicit representation

The authors examine the construction of a connectionist expert system without specifying the network structure before training. The generated connectionist expert system consists of many features, such as operation of forward and backward inference based on partial input information, online learning, noisy data handling, generalization, and the explanation ability. Two sample problems, the Knowledge Base Evaluator 1 and Treatment of Posiboost, are considered in order to illustrate the workings of the connectionist expert system. The training algorithm, which has network generating ability, is presented to build the knowledge base of the connectionist expert system. It provides the abilities needed to realize the described features of the connectionist expert system. This proposed system can be easily used to build expert systems quickly, and the inferencing in the developed systems will be fast. >