论文部分内容阅读
在简要回顾符号专家系统和神经网络技术的基础上，综述了基于这两种技术的混合专家系统研究领域当前发展的状况。基于符号推理的专家系统在７０年代取得了长足的进步，但其自身的弱点决定了专家系统技术在发展过程中必然遇到不可克服的困难；８０年代，以联接主义为特征的神经网络技术得到了迅速地发展，一定程度上突破了符号专家系统所面临的障碍。神经计算与符号推理系统之间的互补性，使得基于这两种技术的混合专家系统被认为是当前人工智能领域重要的研究方向之一。
Based on a brief review of the sign expert system and neural network technology, the current development of the hybrid expert system based on these two technologies is reviewed. Expert system based on symbolic reasoning has made great progress in the 1970s, but its own weakness determines the inevitable difficulties that the expert system technology inevitably encountered in the development process. In the 1980s, the neural network technology characterized by the connectionism got The rapid development, to a certain extent, broke through the barriers faced by the symbolic expert system. The complementarities between neural computing and symbolic reasoning systems make hybrid expert systems based on these two technologies considered as one of the most important research directions in current artificial intelligence.