摘要
当代认知科学和人工智能在认知架构上仍然是以计算表征范式为主。在该范式下,“计算”在人工智能中被视为“智能”,在认知科学中被视为“认知”。这意味着,不论是“智能”还是“认知”,在计算表征范式中都是基于“计算”的,根本原因在于只有“计算”(包括算法)在计算机中才能够操作。因此,如何建立人工认知的计算范式、弄清其运作机制是目前人工认知研究的重中之重。米拉的人工计算理论为解决此难题提供了理论依据,但还不够。若将人工计算理论与适应性表征相结合,进一步弄清人工计算的物理层次、神经生理学符号层次和知识层次相互作用的机制和适应性表征特征,则计算范式将成为一个可解释的、有前途的理论。
Contemporary cognitive science and artificial intelligence are still dominated by the paradigm of computationrepresentation in their cognitive architecture.In this paradigm,“computation”is regarded as“intelligence”in AI and“cognition”in cognitive science.This means that both“intelligence”and“cognition”are based on“computation”in the paradigm of computation-representation.The fundamental reason is that only“computation”Including algorithms can be operated in computers.Therefore,how to establish the computational paradigm of artificial cognition and clarify its operational mechanism is the most important task in the current research on artificial cognition.Mira's theory of artificial computation provides a theoretical basis for solving this problem,but it is not enough.If the theory of artificial computation is combined with adaptive representation to further clarify the mechanism of interaction between the physical,neurophysiological symbolic,and knowledge levels of artificial computation and the features of adaptive representation,the computing paradigm will become an explainable and promising theory.
出处
《上海师范大学学报(哲学社会科学版)》
2026年第1期93-101,共9页
Journal of Shanghai Normal University(Philosophy & Social Sciences Edition)
基金
国家社会科学基金重大项目“人工认知对自然认知挑战的哲学研究”(21&ZD061)。
关键词
人工认知
计算范式
人工计算理论
适应性表征
artificial cognition
computing paradigm
theory of artificial computation
adaptive representation
