论符号主义人工智能的框架问题及其出路

On the Problems and Solution of the Frame Problem in Artificial Intelligence

在人工智能学科的兴起之初,符号主义作为其中的一个主流流派,为人工智能的发展做出了巨大贡献,而框架问题的出现阻碍了它的发展势头,致使符号主义逐渐走向没落。为解决框架问题,符号主义曾提出过两条主要解决路线,然而均未能成功。人们逐渐意识到,框架问题并非仅仅是一个普通技术问题,其背后牵涉到复杂的深层认识论问题(丹尼特)。为此,德雷福斯等主张从现象学角度对符号主义人工智能进行反思,揭示框架问题技术之外的多重内涵,只有对人类智能有了更深入的全方位理解,塞尔理想中的人工智能才有可能实现。

At the beginning of the artificial intelligence,symbolicism is one of the most important schools of artificial intelligence,but the frame problem played a big impact on its development.In order to give a rational answer,there had been two approaches in symbolicism to solving the problem,but both failed.Recently Dreyfus argued that we should analysis the Frame Problem from a phenomenological point of view,intelligent cognition is not a representation—guided behavior,so we should take it in a multiple situation.