摘要
对解决人脑思维机制问题所必须弄清的几个问题作了简要的综述．首先，提出各神经元的基本功能是计算，介绍了人工智能的理论基础的一个假设，其核心是将智能归结为６种符号操作或计算，随之提出了符号计算这个观点存在的一些问题；其次，通过对计算智能与控制智能这两种基础理论的比较和分析，指出要从两个方面衡量智能程度：自适应控制功能、逻辑推理或知识库系统；最后，指出基于抽取和整合的感觉属性检测模型是一个能将网络、符号和反馈机制统一起来的数学模型。将感觉属性检测归结并且为感觉定性映射．
The summary of some problems, which must be worked out for the thinking mechanism of brain, is presented simply. At first, it is proposed that the basic function of neuron is computing, and a hypothesis about theory base of Al is introduced, the core of which is that the intelligence is summed up as 6 kinds of symbol operation or computing, then some problems of this viewpoint are pointed out;secondly, by comparing and analyzing the computing intelligence and control intelligence ,it is pointed out that the intelligence degree must be measured from two aspects: selfadaptable control function,and logic inference or knowledge base system; at last, it is pointed out that the sensational nerve testing model must uniform networks,symbol and feedback mechanisms introduced above, and a sensational attributes testing model based on abstraction and integration for the demand is presented. Sensational attributes testing is summed up as sensational qualitative mapping.
出处
《广西师范大学学报（自然科学版）》
CAS
1997年第3期1-6,共6页
Journal of Guangxi Normal University:Natural Science Edition
基金
国家863高技术
自然科学基金
广西自然科学基金
关键词
思维
人工智能
属性论
神经网络
智能
thinking
artificial intelligence
sensational nerve testing model
attributes theory
