Molodezhnaya St. 3, 119296 Moscow, Tel.: +7 495 930-05-46, Fax: +7 495 930-05-06, E-mail: gcras@gcras.ru
Publication date: 11 November 2026. T. Kudryavtseva

November 11, 2026. On November 11, 2026, at the 128(8) Scientific Seminar of the GC RAS, Chief Scientist of the GC RAS, Academician of RAS Alexey Gvishiani and Research Scientist of the Laboratory of Geoinformatics and Big Data of the Arctic of the GC RAS Olga Samokhina presented the report "Application of Large Language Models for System Analysis of Bibliographic Data".

The report was devoted to the application of AI methods in the system analysis of scientific bibliographic information. The researchers formulated an original system-analytical approach to creating structured thematic arrays of bibliographic data using artificial intelligence technologies in the form of large language models.

The approach was tested on data on geological and geophysical studies of the Arctic region. The relevance of the work is due to the growth of publication activity in the field of Arctic research in the 21st century and the need to process large volumes of disparate bibliographic information. As a result of the study, an algorithm was created for the formation and system analysis of bibliographic arrays, based on a combination of the expert approach and the capabilities of neural networks.

A structured database of 3,405 bibliographic units — descriptions of publications by Russian and foreign Arctic researchers — was built.

Control experiments, including on bibliographic materials of Arctic medicine, confirmed the universality, stability, and reproducibility of the proposed approach when changing the subject area, time intervals, and data selection parameters.

It was shown that artificial intelligence can serve as an auxiliary tool of science for automating bibliographic analysis.