The backbone of science: analysis of citation networks between papers and their sources

2026-07-07

AI Summary

Traditional citation networks treat all references uniformly, making it difficult to identify the core sources that genuinely inspire a study and thereby compromising the accuracy of impact assessment. This work proposes a novel approach that systematically leverages large language models (LLMs) with two prompting strategies to automatically detect seminal citations from full-text articles, constructing a backbone citation network that captures the essential structure of scientific knowledge. Analyses reveal that, although smaller in scale, this backbone network exhibits non-random topology with higher heterogeneity in in-degree distribution. Its topological properties—such as modularity, transitivity, and degree assortativity—systematically differ from those of the full citation network. Nevertheless, rankings of highly cited papers and authors show strong consistency between the two networks, suggesting that despite containing redundancy, the full network remains effective in reflecting relative scholarly influence.

Abstract

The bibliography of scientific papers lists items with variable degree of relevance for the contents of the paper itself. If we could identify the sources, i.e., the works that actually inspired the paper, their citations can help us uncover the genesis of scientific projects and would be more representative of the actual importance of papers and authors than the standard citation counts, when all references are considered. Here we present an analysis of the \textit{backbone of science}, i.e., the network of citations between papers and their sources. The latter are extracted from the full body of papers via Large Language Models (LLMs), which are currently very capable of correctly identifying the context in which a paper is cited. Using two different but related prompts, we find that the LLMs select only a small set of references, not taken at random, and that the resulting backbone networks are quite similar to each other with respect to their in-degree distributions, modularity, transitivity, and degree correlations. Backbone networks have higher heterogeneity in their in-degree distributions, compared to the full network, but the most cited papers are usually the same, with some important exceptions. Citation rankings among authors are also remarkably stable. We conclude that the full citation network, despite its redundancy with respect to the backbones, presents a reliable picture of the relative citation impact of papers and authors.

Problem

• citation networks
• scientific sources
• citation relevance
• backbone of science
• reference analysis

Innovation

• citation backbone
• Large Language Models
• source identification
• scientific impact
• citation network analysis