KNOWLEDGE-BASED SYSTEMS
In order to make computers intelligent，knowledge should be given to them．Knowledge-based systems are such computer systems that they can effectively store and manage human knowledge，and use the knowledge to solve problems that originally require human intelligence．Examples of such systems are expert systems，decision support systems，knowledge base systems，etc．Knowledge-based systems are of crucial importance in AI applications and in future information technology．
The first basic question we are confronted in the construction of a knowledge-based system is how to represent knowledge，which is referred to as knowledge representation．Knowledge representation is to encode information such objects，goal，actions，and processes into data structures and procedures．It is a fundamental component of any knowledge-based systems．Various schemes for representing knowledge have been proposed，of which the most important ones are production rules，frames，semantic networks，logic，script and etc．
The reasoning mechanism and knowledge management are another two important parts of a knowledge-based system．The function of reasoning mechanism is knowledge utilization．It solves users queries by inferencing using knowledge in the knowledge base．Whereas knowledge management is responsible for performing various operations on the knowledge，such as retrieving and updating．
Knowledge acquisition is the central but most difficult problem．In order to acquire knowledge from various sources，such as experts or text，many stages，such as identification，conceptualization，formalization，implementation，and testing have to be conducted．So far，only a small number of systems currently exist that automate portions of the knowledge acquisition task．
One approach to construct a knowledge base system is to combine a relational database management system and PROLOG，with PROLOG functioning as the inference engine．Some research effort has been devoted to constructing such systems by using commercial DBMSs，while others in an attempt to largely improve the perforrmance，by developing new languages which combine logic programming and relational programming concepts（e. g．，the Logic Data language developed by MCC）and developing high performance parallel database machines[1]．It is expected that these systems will exhibit powerful capabilities both in data management and knowledge management．
NOTES
[1]while引导的比较状语表示前后对比。两个by...介词短语都是方式状语，表示方法、手段。
KEYWORDS
acquisition 获取
decision support systems 决策支持系统
conceptualization 概念化
production rule 产生式规则
formalization 形式化
frame 框架
DBMS-database management system 数据库管理系统
reasoning mechanism 推理机制
EXERCISES
Multiple choices．
（1）In order to make computers intelligent，what should be given to them？
a．keyboard b．I/O port
c．network adapter d．knowledge
（2）What is the first basic question we are confronted in the construction of a knowledge-based system？
a．knowledge representation b．parallelism
c．object-oriented programming d．logic programming
（3）What is the function of reasoning mechanism？
a．knowledge representation b．knowledge management
c．knowledge construction d．knowledge utilization
（4）How to solve users queries using knowledge in the knowledge base？
a．by accessing b．by retrieving
c．by inferencing d．by invoking
答案：
（1）d （2）a （3）d （4）c
翻译：
基于知识的系统
为了使计算机智能化，必须将知识赋予它们，基于知识的系统就是这样的一种计算机系统，它能有效地存储和管理知识并用这些知识来解决原来需要人类智能来解决的问题，这些系统的例子有专家系统、决策支持系统、知识库系统等。基于知识的系统在人工智能应用和将来的信息技术中是至关重要的。
在基于知识的系统构成中，我们所面临的第一个主要问题就是如何表达知识，这称之为知识表达。知识表达是将诸如对象、目标、作用和进程等信息编码成数据结构和过程，它是所有基于知识的系统的基本组成部分。为了表达知识，人们提出了各种各样的方案，其中最为主要的就是产生规则、框架、语义网络、逻辑和脚本等。
推理机制和知识管理是基于知识的系统中另外两个重要部分。推理机制的功能是知识的利用，它通过用知识库中的知识进行推理来解决用户查询，而知识管理则负责对知识进行各种操作，比如检索和更新。
知识获取是最主要的也是最困难的问题。为了从各种途径（比如从专家或书本上）获取知识，必须经过像辨识、概念化、形式化、实现和测试等许多阶段。迄今为止，只有少数几个系统部分地实现了自动知识获取。
构成知识库系统的方法之一是将关系数据库系统和逻辑程序设计语言结合起来，以逻辑程序设计语言驱动推理机。某些研究已试图努力通过商业DBMS来构成这样的系统；此外，为了大幅度提高性能，其他研究则致力于通过开发结合逻辑编程和关系编程概念的新语言（如由MCC开发的逻辑数据语言）以及开发高性能的并行数据库计算机。人们期望这些系统将在数据管理和知识管理两方面展示强大的能力。
文章来源于领测软件测试网 https://www.ltesting.net/
