Answers are grounded in this paper and cite their sources.
To address the insufficient attention to the technological value potential and diffusion ability of topics in current evolution analysis, this study employs patent data from 2014 to 2023 in the domains of speech and image recognition. A tripartite "keywords-topics-documents" network is constructed using the BERTopic model for evaluation analysis. The evolution patterns of technological value potential and diffusion ability are investigated through the analysis of keyword associations and patent literature related to technical topics. By examining the evolution trajectories of technical topics and integrating value potential and diffusion ability analyses—based on keyword weights calculated using TextRank and patent citation frequencies—this research reveals a trend of cross-fusion in speech and image recognition topics. This trend is characterized by the incorporation of deep learning and multimodal recognition technologies. The value potential of technological topics exhibits an initial decline followed by a subsequent rise, while the diffusion ability demonstrates a continuous downward trend. This study provides intellectual support for technological forecasting and patent analytics.
No takes yet. Share an insight, caveat, or question.
Li et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: