Artificial intelligence (AI) is reshaping the technological landscape, with patents serving as key indicators of innovation output and technological advancement. Patent assignees play a pivotal role in this process, as their influence reflects both their structural position and functional contributions within innovation ecosystems. However, traditional evaluation methods typically based on patent counts or citation metrics, are limited in capturing the structural complexity and dynamic interactions among innovation entities. To address these limitations, this study constructs a heterogeneous innovation network using AI-related patent data from China spanning 1985 to 2023, integrating multiply types of relationships between patents and assignees. We propose a novel evaluation algorithm—AIHIN, which leverages both the network’s topological structure and fine-grained node attribute features to enable a more comprehensive assessment of assignee influence. Empirical results demonstrate that AIHIN algorithm effectively identifies structurally central and temporally influential assignees in the evolution of AI technologies. This framework provides new insights into the dynamics of innovation within heterogeneous systems and offers methodological contributions to the network-based evaluation of knowledge production and diffusion.
Xipeng Liu (Sat,) studied this question.