Researchers have developed a new method for local citation recommendation (LCR) that does not rely on existing citation graphs, which can introduce bias. They introduced the Local Citation Recommendation Corpus (LCRC), a dataset of 14,735 scientific statements from astrophysics review papers, each with accessible citation targets. This new system achieved an 8.0% improvement in Recall@25 and a 6.9% improvement in NDCG@25 compared to existing dense retrieval baselines, without requiring model fine-tuning. The code and dataset are publicly available to ensure reproducibility. AI
IMPACT This research could streamline the literature review process for researchers by providing more accurate and less biased citation suggestions.
RANK_REASON Academic paper detailing a new dataset and methodology for citation recommendation. [lever_c_demoted from research: ic=1 ai=1.0]
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
