Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-04-09
2002-04-16
Alam, Hosain T. (Department: 2172)
Data processing: database and file management or data structures
Database design
Data structure types
C706S045000, C706S050000, C706S060000
Reexamination Certificate
active
06374261
ABSTRACT:
TECHNICAL FIELD
This invention relates to expert systems.
BACKGROUND OF THE INVENTION
Two types of bottlenecks occur in the life-cycle of an expert system: one is called the “knowledge-acquisition bottleneck”, and the other is called the “knowledge-update bottleneck.” The former refers to how initial knowledge is obtained from domain experts in order to develop an expert system. The latter, also known as knowledge deficiency (or, simply, deficiency), refers to how knowledge evolutions in the domain can be captured to keep an expert system up-to-date for problem solving. The knowledge-update bottleneck is quite critical in domains such as telecommunications where both products and operational knowledge evolve rapidly and the present know-how undergoes constant refinement.
Several methodologies exist for developing and maintaining expert systems and for detecting anomalies, including knowledge deficiency, in a rules database. Not enough work has been done to automate and generalize the process of knowledge-deficiency reduction in expert systems, however. Often, the problem of fixing knowledge deficiency is handled at a high level where methodologies (not tools) provide support for expert-system maintenance. The common approach towards fixing knowledge deficiency in an expert system is to obtain, from time to time, information on the evolving knowledge in a domain and to update the knowledge base manually. Unfortunately, in addition to being tedious and human-effort intensive, the rules-database modification often results in additional anomalies, notably inconsistency in the rules database. The general problem of automating knowledge-database updates to reduce knowledge deficiency is an on-going research effort. It is difficult to provide a generic (that is, a domain-independent) solution to this problem, owing to the widely-varying characteristics of different domains. However, it should be possible to provide a solution tailored to a specific domain that depends on the range and the type of problems that the expert system is designed to solve. Particularly in domains like telecommunications, where operational knowledge plays a critical role in the business, automation to tackle knowledge deficiency for expert systems in the domain could result in considerable savings of cost and effort and could play a vital competitive role.
SUMMARY OF THE INVENTION
This invention is directed to meeting the needs and solving the problems of the prior art. According to the invention, the knowledge-deficiency problem is alleviated by automating the extraction of data relevant to the expert system's diagnostics knowledge database, and automatically updating the knowledge database with the extracted information. More generally according to the invention, a knowledge database that stores information pertaining to a subject (for example, to a target of an expert system) is automatically updated as follows. In response to obtaining a file of information (for example, a product performance recommendation or some other “document” that pertains to the subject), the file is analyzed (e.g., parsed) by computer to identify therein types of information pertaining to the subject (such as different sections of the document, for example). The information of the identified types is then analyzed (e.g., parsed) by computer to identify therein items of information (such as particular fields, or data items) for storage in the knowledge database. The computer extracts found items of information from the file, and arranges them by their types into a database record in the knowledge database. Illustratively, the file is a semi-structured data entity (for example, a document such as the product performance recommendation) that comprises a plurality of types of information that are predefined and at least some of which comprise information expressed in natural language form. Preferably, the computer checks at least some of the identified items of information for consistency between a plurality of the identified types of information, and generates a warning if it determines inconsistency. The file of information may be of one kind or another (for example, a pre-test or a post-test product performance recommendation), and the computer preferably analyzes the file to determine its kind and then performs subsequent analyses using a program-implemented information filter that corresponds to the determined file kind. Having separate filters for different kinds of files greatly simplifies implementation.
The invention encompasses both a method and an apparatus. While the method comprises the steps of the just-characterized procedure, the apparatus effects the method steps. The invention further encompasses a computer-readable medium containing instructions which, when executed in a computer, cause the computer to perform the method steps.
Illustratively, evolving diagnostic knowledge in a domain is represented in the form of semi-structured natural-language reports (generated by, e.g., product development), illustratively called Product Performance Recommendations (PPRs). Every generated PPR represents a fragment of diagnostic knowledge that the expert system should be made aware of. The arrangement reduces the expert system knowledge deficiency by automatically extracting relevant data from the PPRs and updating expert system knowledge databases to keep expert-system diagnostics up-to-date. The arrangement employs intelligent filters that analyze and extract data from the PPRs. Since the extraction process is automated, errors—such as typos and inconsistencies—that can creep in as part of manual insertions and updates to the database, are avoided. This significantly improves the accuracy and reliability of the knowledge-update process. Since the expert system's PPR referrals are kept up-to-date with greater accuracy and reliability, there is a resulting reduction in knowledge deficiency which provides considerable improvement in the overall diagnostic efficiency of the expert system. This in turn helps in fixing product problems faster, better, and cheaper, resulting in improved customer satisfaction.
These and other features and advantages of the invention will become more apparent from the description of an illustrative embodiment of the invention considered together with the drawing.
REFERENCES:
patent: 5542024 (1996-07-01), Balint et al.
patent: 5546507 (1996-08-01), Staub
patent: 5720008 (1998-02-01), McGuinness et al.
patent: 5974405 (1999-10-01), McGuinness et al.
“Automatic E-Mail Distributors”, APTEX,IP Telephony, vol. 6, No. 11, pp. 143-144, Nov. 1998.
P. G. Chander, et al., “Using Goals to Design and Verify Rule Bases”,Decision Support Systems 21, pp. 281-305, 1997.
P. G. Chander, On the Design and Evaluation of Rule-Based Systems,PhD thesis, Department of Computer Science, Concordia University, Montreal, Quebec, Canada, Jun. 1996.
John Debenham, “Expert Systems Designed for Maintenance”,Expert Systems with Applications, vol. 5, pp. 233-244, 1992.
John Keeys Debenham, “A Structure for Expert Systems Maintenance”,Expert Systems with Applications, vol. 11, No. 4, pp. 553-559, 1996.
Uma G. Gupta, “Validation and Verification of Knowledge-Based Systems: A Survey”, Journal of Applied Intelligence vol. 3, No.4, pp. 343-363, 1993.
Alun D. Preece, “A New Approach to Detecting Missing Knowledge in Expert System Rule Bases”,International Journal of Man Machine Studies, vol. 38, No. 4, pp. 661-688, 1993.
Ming Tan, “Telecommunications”,The Handbook of Applied Expert Systems, Edited by Jay Liebowitz, CRC Press, pp. 24-1—24-12 (1997).
A Decentralized AI System for Transcript Applications. H. Hamburger, A. Lodgher, t.Maney, C.Jardine, F.Sift. 1990 IEEE, pp. 48-53.*
Teaching The Computer to Understand. Allport, G.; Jarratt, PI. 1988 IEEE/IEE, p. 8/1-8/2.
Alvarez Miguel O.
Prabhakar Gokul Chander
Alam Hosain T.
Avaya Technology Corp.
Nguyen Tam
Volejnicek David
LandOfFree
Expert system knowledge-deficiency reduction through... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Expert system knowledge-deficiency reduction through..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Expert system knowledge-deficiency reduction through... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2823935