Skip to content
Back to archive

[Submitted] Mapping Technological Convergence via LLM-enhanced Analytics

Published:

Research Status & Connection

This manuscript is currently under review. It extends the conference version of the methodology for analyzing technological convergence.


Research Overview

To understand emerging technology markets, we need to examine the historical trajectory of technological supply. This research introduces Semantic Main Path Analysis (SMPA) to trace the evolutionary path of emerging technologies while addressing the “citation inertia” inherent in traditional models.

Methodology

The framework bridges complex network analysis with modern Natural Language Processing (NLP):

  1. Semantic Similarity Integration: Utilizes Large Language Model (LLM)-based sentence embeddings to calculate the contextual similarity between patents, moving beyond purely quantitative citation counts.
  2. Knowledge Mapping: Extracts the Semantic Main Path by applying network algorithms on top of these semantic weights, identifying core technological trajectories.

For the foundational model and initial award details, please refer to the Conference & Award Archive.


Share this research:

Previous
Profiling the User of Real Estate Tax Tech Market (Poster Award)
Next
TRIZ & Graph RAG-based Intelligent Patent Convergence (IPC) Service (Creative Award)