从基于逻辑的人工智能到社会智能的发展

Intelligence Development from the Logic Artificial Intelligence to Social Intelligence

从思维科学和系统科学交叉发展的观点回顾人工智能半个世纪发展的历程:它经历了以“物理符号系统”为假设的传统人工智能(逻辑思维的模拟);人工神经网络为代表的分布式人工智能(形象思维的模拟);现场人工智能(逻辑思维、形象思维与运行环境相结合的模拟);智能体技术、人工社会研究所涌现的社会智能(社会思维和群体智慧与运行环境相结合的模拟)。钱学森于1990年发表了开放的复杂巨系统及处理这类系统有关复杂问题的方法论,接着提出:我们要研究的不是智能计算机,是人机结合的智能系统,以此为基础,建立了信息空间综合集成研讨体系(Cyber-space forW orkshop ofMetasynthetic Engineering),为社会智能涌现提供了可操作的技术平台[1]。

This paper summarizes the development of Artificial Intelligence（AI） from the viewpoint of Noetic sciences as well as System sciences as thus： Traditional AI with physical symbol system hypothesis （the simulation of logical thinking） , Artificial Neural Networks representative of distributed AI（ the simulation of thinking in imagery） , Situation AI （ the simulation of logical thinking and thinking in imagery in their environments） , the Emergence of Social Intelligence from the research of agent technology and artificial society （the simulation of social thinking and collective wisdom in their environments）. Qian Xuesen proposed a new discipline of science-the study of Open Complex Giant System and its methodology in 1990, and then pointed out what we study was not intelligent computer, we must pay attention to the human-centered man-computer cooperated intelligent system. On the basis of that, Cyberspace for workshop of Meta-synthetic engineering has been established to provide an operational intelligent platform from which social intelligence can be emerged.