Home // ALLDATA 2026, The Twelfth International Conference on Big Data, Small Data, Linked Data and Open Data // View article
Authors:
Xiang-li Zhu
Yifan Wang
Xiaoping Liu
Keywords: Hybrid Intelligence; Frontier Technology Identification; Multi-source Data Fusion; Research Projects; Large Language Models
Abstract:
In the context of the global technological revolution and industrial transformation, the identification of frontier technologies has become a critical component of national strategic competition. However, traditional methods based on citation analysis or patent classification suffer from significant time lags and fail to comprehensively capture the entire lifecycle from technological conception to practical application. To address this, this paper proposes a novel project-linkage paradigm for frontier technology identification, constructing an integrated framework that combines data-driven analysis, intelligent algorithms, and multidimensional assessment. The framework utilizes Large Language Models (LLMs, such as DeepSeek V3) to enhance textual feature extraction and combines Word2Vec vectorization with K-means clustering for technical topic discovery, establishing technology evolution chains through cross-source semantic associations between project requirements and research outputs. Using Defense Advanced Research Projects Agency (DARPA)-funded programs from 2009 to 2025 as empirical subjects, the study finds that: (1) the response rate of research projects to academic publications increased significantly from 80% to 98.3%, indicating that DARPA projects are shifting from following academia to leading academia; (2) three tiers of frontier technologies were identified—mature frontiers, emerging frontiers, and potential frontiers. The results show that the proposed hybrid intelligence framework effectively identifies prospective technological breakthroughs, offering precise support for science and technology decision-making.
Pages: 1 to 7
Copyright: Copyright (c) IARIA, 2026
Publication date: May 24, 2026
Published in: conference
ISSN: 2519-8386
ISBN: 978-1-68558-397-2
Location: Venice, Italy
Dates: from May 24, 2026 to May 28, 2026