EviQE method improves LLM query expansion by selecting relevant documents

Researchers have developed EviQE, a novel method for improving Large Language Model (LLM)-based query expansion by focusing on selecting relevant documents for the model to process. This approach separates the evidence selection from the generation phase, treating different reformulators as complementary retrieval perspectives. Experiments across TREC DL and BEIR benchmarks demonstrated that EviQE, particularly with relevance-based evidence selection using LLM-Score, significantly outperforms direct reformulation and single-source expansion methods. The study found that pooling candidates from multiple reformulators increases relevant document coverage, and subsequent retrieval-generation rounds offer minimal additional benefit after strong conditioning evidence is selected. AI

IMPACT This method could improve the precision and relevance of search results generated by LLM-powered systems.

RANK_REASON The item is an academic paper detailing a new method for information retrieval using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.IR (Information Retrieval) →

• alphaXiv
• arXiv
• Beir
• CatalyzeX Code Finder for Papers
• Connected Papers
• CORE Recommender
• DagsHub
• EviQE
• Gotit.pub
• Hugging Face
• Litmaps
• LLM
• LLM-Score
• ScienceCast
• scite Smart Citations
• TREC DL

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →