人工认知如何创造性地解决问题  

How does Artificial Cognition Creatively Solve Problems

作　　者：魏屹东[1] Wei Yidong

机构地区：[1]山西大学哲学学院,太原030006

出　　处：《南京社会科学》2025年第1期10-19,58,共11页Nanjing Journal of Social Sciences

基　　金：国家社科基金重大项目“人工认知对自然认知挑战的哲学研究”(21&ZD061)的阶段性成果。

摘　　要：一个人工认知系统是否具有创造性,如何拥有创造力,这是新一代人工智能必然遭遇并且必须解决的问题。从认知哲学上审视,一个系统有创造性的内在原因是具有因果推理能力,即一种适应性表征能力。这意味着,人类和智能机器人只要有了适应性表征能力,也就具有了某种程度的创造力。但如何理解创造性,认知创造性的机制是什么,人工认知系统的适应性表征结构是什么,如何为智能体创造性地解决问题建模,这一系列问题是本文要探讨的。从适应性表征方法论来看,基于“时钟隐喻”的创造性齿轮框架可以给出合理的解释,即认为创造力好比转动的齿轮,其实质是一种适应性表征能力。Whether an artificial cognitive system is creative and how to have creativity are questions that the new generation of AI must face and address.Examined from the philosophy of cognition,the intrinsic reason for a system to be creative is to have causal ability,an ability of adaptive representation.This means that as long as humans and intelligent robots have the ability of adaptive representation,they also have some degree of creativity.But how to understand creativity?What are the mechanisms of cognitive creativity?What are the adaptive representational structures of artificial cognitive systems?And how to model creative problem solving by agents?This set of questions is what this paper will explore.From the viewpoint of adaptive representation,the CreaCogs based on the clock metaphor can give a reasonable explanation for the idea that creativity is like turning gears,which in essence is an ability of adaptive representation.

关 键 词：人工认知 人工智能 智能体 适应性表征 创造性 

分 类 号：B023[哲学宗教—哲学理论] TP18[自动化与计算机技术—控制理论与控制工程]