Scholarlink

<p>The interplay between science and technology (S&amp;T) is propelling unprecedented advancements in the field of artificial intelligence (AI). Knowledge network analysis stands as the primary means of identifying S&amp;T connections, typically measuring node and edge coupling between S&amp;T knowledge networks. However, an exclusive focus on network static topology overlooks inherent dynamic synchronization, making it particularly inadequate for increasingly intricate S&amp;T systems. To comprehensively detect S&amp;T interactions within AI, this study introduces a novel dual-perspective framework that incorporates both network structure coupling (NSC) and time series coupling (TSC). First, the S&amp;T knowledge networks are constructed to delineate the static topological structure of S&amp;T knowledge systems. Subsequently, we develop a customized random walk (RW) mapping approach from network to nonlinear time series tailored for S&amp;T knowledge networks. Finally, by integrating the coupling of both static network topologies and dynamic fluctuation sequences, we quantified and evaluated synergies between S&amp;T. Empirical research was conducted on scientific publications and patent data pertaining to AI, and the effectiveness and robustness of our proposed approach were verified in real-world and synthetic network scenarios.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A multi-dimensional coupling model for detecting science-technology interactions within AI

  • 期刊论文

  • 发表日期：

    2026年2月14日

查看全文
  • Zhichao Ba,
  • Leqi Zhu,
  • Yujie Zhang,
  • Kai Meng,
  • Zujun Liu

摘要

The interplay between science and technology (S&T) is propelling unprecedented advancements in the field of artificial intelligence (AI). Knowledge network analysis stands as the primary means of identifying S&T connections, typically measuring node and edge coupling between S&T knowledge networks. However, an exclusive focus on network static topology overlooks inherent dynamic synchronization, making it particularly inadequate for increasingly intricate S&T systems. To comprehensively detect S&T interactions within AI, this study introduces a novel dual-perspective framework that incorporates both network structure coupling (NSC) and time series coupling (TSC). First, the S&T knowledge networks are constructed to delineate the static topological structure of S&T knowledge systems. Subsequently, we develop a customized random walk (RW) mapping approach from network to nonlinear time series tailored for S&T knowledge networks. Finally, by integrating the coupling of both static network topologies and dynamic fluctuation sequences, we quantified and evaluated synergies between S&T. Empirical research was conducted on scientific publications and patent data pertaining to AI, and the effectiveness and robustness of our proposed approach were verified in real-world and synthetic network scenarios.