سیستم های مدیریت پایگاه دانش _ ابزارهایی برای ایجاد سیستم های هوشمند تایید شده
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|5481||2003||7 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 3540 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Knowledge-Based Systems, Volume 16, Issue 3, April 2003, Pages 165–171
As automation of business processes becomes more complex and encompasses less-structured domains, it becomes even more essential that the knowledge used by these processes is verified and accurate. Most application development is facilitated with software tools, but most business rules and expert systems are developed in environments that provide inadequate verification testing. This paper describes an emerging class of applications we refer to as Knowledge Base Management Systems (KBMS). The KMBS provides a full life-cycle environment for the development and verification of business rule and expert systems. We will present an overview of knowledge base verification, the KBMS life-cycle, and the architecture for a KBMS. We then describe building a small expert system in the KBMS, with emphasis on the verification testing at each stage. We conclude with a summary of the benefits of a KBMS.
In many industries, the key to efficiency is automation. The first targets for automation were the most structured problems, such as accounting. Our ability to automate less structured domains is constrained by our ability to verify the knowledge used in the automation. Automation of less-structured domains is achieved with active intelligent components such as expert systems or business rule systems. We will refer to these systems in general as knowledge-based systems. There are many difficulties in building a useful knowledge-based system, including difficulty in capturing deep knowledge, lack of robustness and flexibility, inability to provide deep explanations, difficulties in verification, little learning from experience , and computational efficiency. Knowledge-based systems are harder to build than most people perceive them to be because of the dependencies between rules in the system and the difficulty of verifying them. The importance of verification in knowledge-based systems cannot be overstated. A single bad rule in a medical expert system could kill a patient, just as a single bad rule in a business system could put the company out of business. As we automate more and more processes, the need for verification becomes even more critical. Many automated process can perform incorrectly for a long time, as no person is responsible for checking the process. In a survey of 40 knowledge-based system tools conducted in 1997, Murrell  concludes “The paper provides… areas in which the researcher can provide practitioners with valuable tools for the verification and validation of knowledge-based systems, where currently there are none…” It should also be noted that the systems surveyed were all research systems, and that none are available to the general public. Business rules are usually created in a text editor as program code or database triggers, making the programmer responsible for verifying the program logic. Expert systems are usually built in a vendor-supplied tool that translate specifications into code, but may provide little or no verification testing of the specifications. At least one expert system development tool from a major manufacturer will allow duplicate and conflicting rules to be created, and none verify the application for all 23 verification criteria. However, articles about early examples of Knowledge Base Management Systems (KBMS) software have begun to appear. Aquinas, in production use at The Boeing Company, offers many of the functions desirable in a KBMS. It elicits knowledge directly from the expert into grids, which are analyzed for completeness and consistency. It refines the specifications and generates code for several expert system shells. Complete applications may be created in less than two hours . EULE is a system developed by Swiss Life that has functionality ‘in the triangle of Knowledge Representation, Business Process Modeling, and Knowledge Management.’ This system is designed to automate office tasks in the insurance industry, and uses an extendable High Level Language (HLL) to model characteristics such as laws, regulations, and preconditions for activities. The resulting system is integrated into Swiss Life's Organizational Memory systems, and it is suitable for embedding in business process models . The purpose of KBMS is to offer computerized assistance for building knowledge-based systems. A KBMS: 1. Provides full life-cycle support from knowledge acquisition to delivered code. 2. Guides the user through the development cycle. 3. Detects or prevents verification errors. 4. Algorithmically refines knowledge. 5. Generates code for the knowledge-based system. We will first consider verification of the knowledge-based system because of its influence on KBMS life cycle and the KBMS architecture.
نتیجه گیری انگلیسی
The automation of business processes requires accurate knowledge. A KBMS is a development environment for knowledge-based systems that automates much of the process of building a knowledge-based application. The expert does not need the intervention of a knowledge engineer. Little knowledge of knowledge-based systems is required, as the KBMS guides the user through the process. The KBMS generates code free from syntax-errors, so no implementation language skill is required. The expert is also free to work on the knowledge-based system alone, instead of having to clearly communicate every change to a knowledge engineer. The cost of a knowledge-based system is minimized by dramatically reducing the cost of the knowledge engineer and by enabling the expert to quickly build and maintain a correct knowledge-based system. The knowledge-based systems code generated by the KBMS should represent the state of the art. The knowledge in the project is exhaustively verified, once considered an oxymoron in the expert system field. Refinement with truth-preserving algorithms is used to derive the most concise representation possible. Neither of these tasks can be adequately performed without computers. Testing and debugging demands are reduced. Syntax errors are eliminated. The most dramatic impact of the KBMS is project throughput. An expert, after a little KBMS training, can create a 150-rule knowledge-based system in an hour. Traditional methods would require months of knowledge engineer and expert time to create this system. One topic deserving more research is the transposition of declarative applications into procedural ones. This approach would maintain the system inside the development environment as specifications, but blend the rules and the actions of the inference engine into procedural code, eliminating the inference engine and dramatically increasing run-time performance and lowering implementation cost. Another topic for further research is matching knowledge acquisition tools to types of problems. There are advocates of graphical representations in EULE , hierarchy grids in Aquinas , and production rules as used in the prototype KBMS. It is possible that different knowledge representations and knowledge acquisition tools are suited to different domains, making the availability of multiple approaches desirable in a KBMS. In summary, a KBMS allows a typical end-user to quickly generate verified, refined, high-performance knowledge-based systems. The expectation is that the availability of such a tool will make knowledge-based systems accessible to a far wider audience, and result in the development of better knowledge-based systems.