ANALYSIS OF THE ARCTIC'S TECHNOLOGICAL DEVELOPMENT BASED ON SEMANTIC ATTRIBUTE GRAPHS AND MULTI-CRITERIA ATTENTION

Authors:
Publisher website
Google Scholar
Abstract
The technological development of the Russian Arctic is a key factor in ensuring national security, enabling the sustainable exploration of natural resources, and strengthening sovereignty in the face of extreme climatic and geographic challenges. Despite significant public investment, the effectiveness of federal programs remains insufficiently understood, and their impact on technological outcomes is poorly quantified. Existing approaches based on aggregate indices and expert assessments fail to identify which initiatives drive technological transformation or why some regions demonstrate high levels of innovation while others do not. This study aims to develop an interpretable model capable of identifying cause-and-effect relationships between federal programs and technological development indicators. Its novelty lies in the proposal of a new model, MASGN-TT—a semantic graph neural network approach with a multicriteria attention mechanism, applied for the first time in the Russian context to analyze the technological development of the Arctic. The model not only measures program impacts but also visualizes their influence on specific technological domains, ranging from digital infrastructure to environmental monitoring. The results show that two key public policy areas—digital transformation and environmental monitoring—are the dominant drivers of technological growth, while other programs, despite considerable budget allocations, yield limited returns. The analysis also points to a persistent gap between workforce preparation and tangible results. More broadly, the effectiveness of technological development in the Arctic is shaped less by the volume of funding than by how programs are structured and where their efforts are directed. The proposed model provides a tool for evidence-based evaluation of public programs, ensuring transparency, replicability, and adaptability to managerial decision-making. This paves the way for a targeted technological development policy aimed at maximizing systemic impact.

This publication has 12 references indexed in Scilit: