Robustness of RAG Systems Across Dense and Sparse Retrieval for Long-Tail Scientific Queries

This report synthesises findings from 16 peer-reviewed papers addressing the following research question: How does the robustness of RAG systems vary with different retrieval methods (e.g., dense vs. sparse retrieval) when applied to long-tail scientific queries, evaluated through precision-recall curves. 9 claims were extracted from source literature; 9 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 9.2/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does the robustness of RAG systems vary with different retrieval methods (e.g., dense vs. sparse retrieval) when applied to long-tail scientific queries, evaluated through precision-recall curves?

Autonomous literature synthesis. Automated review score: 9.2/10. Full text and citation available at Assignee Research.