Category:
Category:
Query Rewriting
Category:
Retrieval & RAG
Definition
Transforming user queries to improve retrieval accuracy.
Explanation
Query rewriting modifies or expands user queries to increase the chance of retrieving relevant documents. This includes synonym expansion, keyword extraction, semantic reformulation, and multi-query generation. Query rewriting dramatically improves RAG performance when original queries are vague, incomplete, or overly natural.
Technical Architecture
User Query → Rewriting Engine → Expanded Queries → Retrieval Pipeline → LLM
Core Component
Semantic rewriter, keyword extractor, intent parser
Use Cases
Search engines, RAG systems, enterprise assistants
Pitfalls
Over-expansion reduces precision; poor rewrites degrade retrieval
LLM Keywords
Query Rewriting, RAG Query Optimizer
Related Concepts
Related Frameworks
• Retrieval Pipelines
• Reranking
• Hybrid Search
• Query Optimization Architecture