Search papers, labs, and topics across Lattice.
ReFormeR learns explicit query reformulation patterns from query-reformulation pairs, creating a library of transferable patterns. It then selects and applies the most appropriate pattern based on the retrieval context of a new query. This approach guides LLMs towards targeted reformulations, improving performance on TREC DL 2019, DL 2020, and DL Hard datasets compared to traditional and LLM-based methods.
Stop prompting LLMs to blindly rewrite queries – ReFormeR distills query transformations into reusable patterns that actually improve retrieval.
We present ReFormeR, a pattern-guided approach for query reformulation. Instead of prompting a language model to generate reformulations of a query directly, ReFormeR first elicits short reformulation patterns from pairs of initial queries and empirically stronger reformulations, consolidates them into a compact library of transferable reformulation patterns, and then selects an appropriate reformulation pattern for a new query given its retrieval context. The selected pattern constrains query reformulation to controlled operations such as sense disambiguation, vocabulary grounding, or discriminative facet addition, to name a few. As such, our proposed approach makes the reformulation policy explicit through these reformulation patterns, guiding the LLM towards targeted and effective query reformulations. Our extensive experiments on TREC DL 2019, DL 2020, and DL Hard show consistent improvements over classical feedback methods and recent LLM-based query reformulation and expansion approaches.
BAIR