符号奠基问题及其解决策略
发布时间：2019-02-26 17:06
【摘要】：塞尔对强人工智能的批判实际上提出了这样的问题:如果没有外部观察者的解释,形式符号系统的句法属性或物理属性是否能够充分决定语义内容?虽然塞尔给出了否定的答案,但这个问题却启发了人工智能的研究者去研究物理系统内的符号如何能够自动地获得意义的问题。哈纳德把这个问题概括为"符号奠基问题"。为了解决这个问题,目前大致有内在奠基、因果奠基、指号学奠基和基于行动的语义学等几种主要方案,但它们的解法都存在不同程度的缺陷。如何将不同进路的人工智能架构糅合起来,既免于各种进路的缺点而又能发挥其优势,以更好地解决符号奠基问题,是人工智能哲学中的值得研究的方向。
[Abstract]:Searle's criticism of strong artificial intelligence actually raises the question: can the syntactic or physical attributes of the formal symbol system fully determine the semantic content without an outside observer's explanation? Although Searle gave a negative answer, this question has inspired researchers in artificial intelligence to study how symbols in physical systems can automatically acquire meaning. Harnard summed up the problem as "the groundbreaking of symbols." In order to solve this problem, there are several main schemes, such as intrinsic foundation, causal foundation, indexology foundation and action-based semantics, but their solutions have some defects. How to combine the artificial intelligence architecture with different approaches can not only avoid the shortcomings of various approaches but also play its advantages, so as to solve the problem of laying the foundation of symbols better, which is the direction worth studying in the philosophy of artificial intelligence.
【作者单位】： 华南师范大学公共管理学院;北京师范大学哲学学院;
【基金】：国家社科基金“认知科学对当代哲学的挑战”(编号11AZD120) 中央高校基本科研业务费专项资金资助(编号SKZZB2015040)的阶段性成果
【分类号】：TP18-02
，
本文编号：2430969
[Abstract]:Searle's criticism of strong artificial intelligence actually raises the question: can the syntactic or physical attributes of the formal symbol system fully determine the semantic content without an outside observer's explanation? Although Searle gave a negative answer, this question has inspired researchers in artificial intelligence to study how symbols in physical systems can automatically acquire meaning. Harnard summed up the problem as "the groundbreaking of symbols." In order to solve this problem, there are several main schemes, such as intrinsic foundation, causal foundation, indexology foundation and action-based semantics, but their solutions have some defects. How to combine the artificial intelligence architecture with different approaches can not only avoid the shortcomings of various approaches but also play its advantages, so as to solve the problem of laying the foundation of symbols better, which is the direction worth studying in the philosophy of artificial intelligence.
【作者单位】： 华南师范大学公共管理学院;北京师范大学哲学学院;
【基金】：国家社科基金“认知科学对当代哲学的挑战”(编号11AZD120) 中央高校基本科研业务费专项资金资助(编号SKZZB2015040)的阶段性成果
【分类号】：TP18-02
，
本文编号：2430969
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2430969.html
