arXiv:2606.13905v1 Announce Type: cross Abstract: LLM-based query expansion improves retrieval by enriching the original query with additional context. Yet most methods remain generation-driven, producing plausible pseudo-documents or expansions without checking how the target corpus responds. This can introduce retrieval drift, amplify misleading vocabulary, or miss terms that distinguish relevant from non-relevant documents. We argue that effective expansion requires retrieval-grounded feedback, not just single-pass generation or unverified iteration. We introduce ADORE (ADapt, Observe, Rele
The proliferation of LLM-based systems makes improving their retrieval capabilities critical for practical applications, addressing existing limitations like retrieval drift.
Improving query expansion through retrieval-grounded feedback means more accurate and reliable information retrieval, which is fundamental to the performance of many AI applications.
Retrieval models will become more adaptive and less prone to generating irrelevant or misleading expansions, leading to enhanced utility in information-intensive tasks.
- · AI developers
- · Search engine companies
- · Data analysis platforms
- · Knowledge management systems
- · Systems relying on unoptimized, generation-only query expansion
More precise and contextually relevant AI-driven search results for users.
Increased efficiency and effectiveness of AI agents and large language models that depend on external knowledge retrieval.
Acceleration of research and development in agentic systems as their access to accurate information significantly improves, expanding their capabilities and applications.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL