Citation recommendation system improves reference search without bias
Local Citation Recommendation Without Graph Signals
Information RetrievalComputation and LanguageDigital Libraries
Summary
Finding the right references for scientific papers can be slow and biased toward popular or recent works. The authors created a new dataset of scientific sentences with accurate citation references to test methods for suggesting citations. They developed a system that makes better citation recommendations without relying on citation popularity, improving the chance of finding the right references. The authors share their dataset and code so others can build on their work.
What this means in practice
- •For academic search engine developers: Enhance citation suggestions in academic search engines by using citation recommendations that reduce bias toward popular or recent papers.
- •For digital library maintainers: Integrate unbiased citation recommendation to improve reference linking and discovery within scientific digital libraries.
Authors
Benjamin Basseri, Pavlos Protopapas, Kemeng Zhang, Alberto Accomazzi, Stephanie Jarmak, Anna Kelbert
Abstract
Resolving correct citations can be a time-consuming task for research writers. We address local citation recommendation (LCR), where the goal is to retrieve appropriate references for a given sentence. Web tools generally depend on the existing citation graph, which may bias results toward citation count or more recent publications. LCR research using full paper texts has been limited by the lack of publicly available data. We present a new dataset, the Local Citation Recommendation Corpus (LCRC), containing 14,735 scientific statements derived from astrophysics review papers, each with one or more citation targets whose full text is accessible via API. Using this dataset we systematically compare retrieval methods and develop a citation recommendation system that avoids citation graph bias while achieving 8.0% improvement in Recall@25 and 6.9% improvement in NDCG@25 over dense retrieval baselines without model fine-tuning. To support reproducibility, we make our code and dataset publicly available.