概念获取的框架表征重述方法  

Concept-acquisition by Representation Redescription of Frame

作　　者：施芸虹[1] 危辉[1] 陈燕[1] 

机构地区：[1]复旦大学计算机科学系认知算法模型实验室

出　　处：《计算机应用研究》2007年第1期18-19,30,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60303007);国家"973"计划资助项目(2001CCA03000);中国科学院智能信息处理重点实验室开放基金项目(IIP2002-3)

摘　　要：基于知识的问题求解需要一个丰富而相对完备的信念系统,尤其是当任务具有领域非限定特征时。经典知识工程关于知识获取、表示与使用的方法只能适应领域受限问题,因为它不是从概念发展的角度来建构概念系统的,面临着Framework Problem。针对认知心理学对概念系统发展与表征问题研究有一定的认知深度同时又缺乏系统构造与实现机制研究的情况,提出了一种基于框架表征的概念系统表征与发展方法,详细研究了概念在内的隐水平I、外显水平E1、外显水平E2和外显水平E3上的表征和发展过程。这一研究对于提高基于知识的系统推理能力和问题求解能力具有显著意义。Knowledge-based problem solving requires a concept system that is comparatively rich and complete, especially when the problem is domain unrestrictive. The methods about knowledge acquisition, representation and usage in classical knowledge engineering can adapt themselves merely to domain restricted problems, because it doesn＇t take a developmental view to construct concept system and consequently it is confronted with Framework Problem. In cognitive psychology, the study of concept system has a fairish cognitive depth on issues of development and representation, but lacks investigations on all details of system construction and realization. Stemmed from the theory of developmental psychology, this paper proposes a frame-based representing method for concept system, focusing on the representation and development of concepts on four representing levels ： Implicit （Ⅰ）, Explicit 1 （ E1 ）, Explicit 2 （E2） and Explicit 3 （ E3 ） representations. It will contribute well to the adaptability and flexibility in reasoning and problem solving of knowledge-based systems.

关 键 词：概念化 知识表示 学习 语义系统 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]