摘要
自达尔文以来生物适应环境的观念已深入人心。智能主体,包括人类和机器,其智能的发展也是适应环境的。就人工智能主体而言,它展现的知识表征过程充分体现了人工智能主体适应其环境的特征,尽管这种环境是人类设计者设置的。从认知哲学来看,人工智能的知识表征也与其人类设计者的知识表征极为类似。首先需要给它建立一个能够表征抽象概念的本体论框架,然后在此框架下根据逻辑规则执行各种表征程序,诸如客体范畴化、事件运算、命题态度、范畴推理、缺省信息推理、真理保持系统和基于知识的主体的推理,这些程序在各自环境中适应性地进行表征与推理。人工智能主体Alpha Go和Master是这方面的两个典型例子。
The idea of organism adapting to the environment has been deeply rooted in the hearts of the people since Darwin.Agents,including humans and machines,their intelligent developments are also adapted to the environment.For artificial agents,the process of knowledge representation that agents showed fully reflects the characteristics of artificial agents adapting to their environment,even though the environment is set by human designers.From the perspective of philosophy of cognition,the knowledge representation of artificial intelligence is very similar to that of human designers.First of all,it is necessary to build an ontological framework that can represent abstract concepts.Then various representation programs are executed under this framework according to logical rules such as first-order logic.These programs include object categorization,event operation,propositional attitude,categorical reasoning,default information,truth preserving system,knowledge-based agent,adaptive representation and reasoning in their respective environments.Two artificial agents,Alphago and Master,are two typical examples to illustrate the adaptive knowledge representation and reasoning in artificial intelligence.
出处
《上海师范大学学报（哲学社会科学版）》
CSSCI
北大核心
2019年第1期65-75,共11页
Journal of Shanghai Normal University(Philosophy & Social Sciences Edition)
基金
国家社会科学基金重点项目"科学认知的适应性表征研究"(16AZX006)的阶段性成果
关键词
人工智能
人工主体适应性
知识表征
推理
artificial intelligence
Artifical agent adaptability
knowledge representation
reasoning
