Researchers have developed DESA (Dense Expansion and Sparse Anchoring), a novel method for improving information retrieval using LLM-based query expansion. Unlike previous approaches that fuse fixed top-L dense and sparse rankings, DESA separates the effects of cutoff and fusion by evaluating retrieval effectiveness under complete-list fusion. The method generates complementary reference passages with an LLM, adding new semantic directions to the dense query and incorporating lexical cues into sparse retrieval without broadening the original query's lexical support. Across seven BEIR datasets, DESA demonstrated improvements in nDCG@10 and Recall@20 while significantly reducing the access depths for both dense and sparse retrieval. AI
IMPACT Introduces a new technique for improving information retrieval efficiency and effectiveness by leveraging LLMs for query expansion.
RANK_REASON The cluster describes a novel method presented in a research paper. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
