期刊文献+

论甲骨文是人工智能范式升级的有效媒介及路径

Discussions on Oracle Bone Script as an Effective Medium and Pathway for AI Paradigm Shift
在线阅读 下载PDF
导出
摘要 本文探讨了通用人工智能(AGI)发展的核心路径,提出汉字及甲骨文作为更高效的底层语言媒介,能够推动大语言模型向大图像模型的范式升级。研究表明,汉字的三维拓扑结构(形、音、义)与人类视觉认知机制高度契合,而甲骨文作为象形文字原型,其简约版的图像是大语言模型升级到大图像模型性价比和效率最高的媒介。本文主张以甲骨文为战略支点,结合三进制量子计算与图像符号模型,构建兼具文化根性与计算效率的通用人工智能体系。最后,建议国家把保护和研究甲骨文作为人工智能大图像模型基础的国运级别的文化安全课题来对待,为中国在全球化4.0时代的技术-文明博弈中确立竞争优势。 This paper explores the core path for the development of Artificial General Intelligence(AGI),proposing that Chinese characters and Oracle Bone Inscriptions(OBIs)serve as more efficient underlying linguistic media to promote the paradigm upgrade of large language models(LLMs)to large image models(LIMs).Further analysis indicates that the linear text processing relied on by cur⁃rent AI(such as Byte Pair Encoding,BPE)leads to semantic dimensionality reduction of Chinese char⁃acters,whereas the topological structure and multimodal attributes of OBIs(e.g.,astronomical meta⁃phors,component combinations)can provide a bio⁃inspired design framework for neuro⁃symbolic sys⁃tems.This paper advocates using OBIs as a strategic fulcrum,combining ternary quantum computing with image⁃symbol models,to construct an AGI system that integrates cultural roots and computa⁃tional efficiency.Finally,it suggests that the state should treat the protection and research of OBIs as a national⁃level“culture security”subject for the foundation of AI⁃based large image models,estab⁃lishing a competitive edge for China in the technology⁃civilization game of Globalization 4.0.
作者 周缙峤 岑元达 胡娜 ZHOU Jinqiao;CEN Yuanda;HU Na(Beidouxing Yixin Institute Co.,Ltd.Region:Hong Kong,China;New Era Digital Culture Center City&Postcode:Beijing 100001,China;School of Business Management,Shanghai International Studies University,201620,China)
出处 《上海管理科学》 2026年第1期17-25,共9页 Shanghai Management Science
关键词 通用人工智能 甲骨文 大语言模型 大图像模型 Artificial General Intelligence(AGI) Oracle Bone Inscriptions(OBIs) Large Language Models(LLMs) Large Image Models(LIMs)
  • 相关文献

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部