Get Started
Home
Topics
Search
Library
Research questionHow can retrieval systems honor negative constraints without discarding relevant evidence?The excluded concept still appears in the query, so similarity search may retrieve exactly what the user wants to avoid. Penalizing that evidence too strongly can remove documents needed to answer the included part.
AI
Evaluation & Benchmarks
Information Retrieval
Machine Learning
Natural Language Processing
Retrieval-Augmented Generation
Latest papersRecent research connected to this question, newest first.E-SENS: Exclusion-Sensitive Penalization for Negative-Constraint RetrievalThe source presents a training-free reranking strategy that scores the excluded side separately and adjusts dense-retrieval rankings. Evidence is reported on ExcluIR with multiple embedding models, so the observed trade-offs are limited to that benchmark and retrieval setting.research paper · Sep 4, 2026
Related questions
How can NER systems decide when external retrieval helps long-tail entities without adding noise or cost?How should retrieval systems select passages that form coherent evidence chains for multi-hop questions?How can search agents learn when retrieval is necessary and ground answers in evidence without costly supervision?How can high-stakes LLM systems distinguish unsupported claims from novel ones and prioritize expert verification?