The gap between academic innovation and industry application remains a significant barrier to economic growth, largely due to inefficient and opaque technology transfer mechanisms. While digital platforms exist, they often act as static repositories, and even advanced knowledge graph-based systems fail to model the crucial spatio-temporal dynamics of research evolution. To overcome these limitations, we introduce TechNexus, an intelligent platform that revolutionizes the discovery of collaborative opportunities. TechNexus makes two core contributions: First, it constructs a Spatio-temporal Knowledge Graph (STKG) to create a dynamic, evolving map of the research landscape. Second, it leverages a Retrieval-Augmented Generation (RAG) framework, empowering a Large Language Model (LLM) to reason over this structured, time-and-space-aware data. This synergy enables TechNexus to understand nuanced, natural language queries from industry and generate deeply contextualized recommendations, significantly increasing the likelihood of successful technology transfer. Our platform is demonstrated at https://www.hfutimi.cn/home .

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

TechNexus: An LLM-Powered Platform with Spatio-Temporal KGs for University-Industry Technology Transfer

  • Jun Feng,
  • Xuezhi Yang,
  • Xin Jing,
  • Shuai Fang

摘要

The gap between academic innovation and industry application remains a significant barrier to economic growth, largely due to inefficient and opaque technology transfer mechanisms. While digital platforms exist, they often act as static repositories, and even advanced knowledge graph-based systems fail to model the crucial spatio-temporal dynamics of research evolution. To overcome these limitations, we introduce TechNexus, an intelligent platform that revolutionizes the discovery of collaborative opportunities. TechNexus makes two core contributions: First, it constructs a Spatio-temporal Knowledge Graph (STKG) to create a dynamic, evolving map of the research landscape. Second, it leverages a Retrieval-Augmented Generation (RAG) framework, empowering a Large Language Model (LLM) to reason over this structured, time-and-space-aware data. This synergy enables TechNexus to understand nuanced, natural language queries from industry and generate deeply contextualized recommendations, significantly increasing the likelihood of successful technology transfer. Our platform is demonstrated at https://www.hfutimi.cn/home .