人工智能理论:从分立到统一的奥秘

A Secret Making AI Theory from Separation to Unification

由于智能问题的重要性和复杂性,人们曾经分别从结构、功能和行为等多个不同侧面对自然智能问题展开了模拟研究,于20世纪形成了人工智能的结构主义、功能主义和行为主义3个有代表性的理论.这些理论分别都取得了很大的成功,然而它们之间却难以互相沟通和互相认可.为此,提出了一种新的研究途径———从智能生成的共性核心机制入手探讨智能的本质,称为“机制主义”.研究发现:一般情形下,智能生成的核心机制表现为“信息—知识—智能的转换”;而且,结构主义、功能主义、行为主义三者能在“机制主义”框架下得到和谐的统一.这可为人工智能理论的未来发展提供新的认识和新的方向.

Due to the high importance and complexity in artificial intelligence （AI） study, there have been various schools of AI theory that were formed in the 20th century among which the structuralism, functionalism and behaviorism approaches are the representatives. All the three approaches have made appreciable successes. However, there have been also controversies among the three. A new approach, named mechanism approach, is thus proposed here to attack the nucleus of intelligence through the investigation on the core mechanism of its formation. It is our discovery that the core mechanism of intelligence formation appeared to be a series of transformations--from information to knowledge and further to intelligence. In addition, it is also found that the three mentioned schools of AI theory could well be harmoniously unified within the framework of the mechanism approach. The mechanism approach and the harmonious unification of AI theories may hopefully provide a new understanding and hence a new direction for the future development of AI theory.