Query Expansion
Definition
Query Expansion is an information retrieval technique that augments a query with additional related terms, concepts, entities or representations to increase the likelihood of retrieving relevant information.
Why it matters
A user’s wording may omit synonyms, related concepts or terminology used in relevant documents. Expansion can improve retrieval coverage by broadening the match space, although overly broad expansion can reduce precision.
How it works
Expansion source
- Additional terms can come from thesauri, relevance feedback, embeddings, knowledge graphs, language models or other semantic resources.
Expanded representation
- The original query is supplemented with related words, concepts or entities.
- Some systems preserve the original terms while adding weighted expansion terms.
Retrieval trade-off
- Expansion can increase recall by finding relevant material expressed differently.
- Poorly controlled expansion can introduce unrelated candidates and reduce precision.
How Netsleek uses the term
Netsleek uses Query Expansion as the established IR mechanism and keeps it distinct from Query Fan-Out. Expansion broadens the representation of one information need, while fan-out can create multiple separate retrieval queries or facets.
Comparisons
- Query Expansion vs Query Reformulation: Reformulation changes the query representation. Expansion specifically adds related terms or concepts.
- Query Expansion vs Query Fan-Out: Expansion broadens a query. Fan-out creates multiple queries or retrieval paths.
- Query Expansion vs Multi-Query Decomposition: Decomposition splits a complex information need into multiple sub-questions. Expansion broadens one query representation.
How Query Expansion Connects Across AI Systems
Query Expansion does not operate in isolation. It connects to other established concepts across the AI search and retrieval process:
- Can use semantics: Embeddings and Semantic Retrieval can identify conceptually related expansion terms.
- Can use entities: Entity associations and Knowledge Graph relationships can support entity-based expansion.
- Affects recall: Expansion can improve Recall by increasing the chance that relevant items are retrieved.
- Affects precision: Expansion must be controlled because broader matching can also retrieve irrelevant items.
Related glossary concepts
- Query Reformulation (Query Rewriting)
- Query Fan-Out
- Multi-Query Decomposition
- Recall (Information Retrieval)
- Precision (Information Retrieval)
- Semantic Retrieval
- Knowledge Graph
Common misinterpretations
- Query expansion does not guarantee better retrieval.
- Expansion and fan-out are not synonyms.
- Adding more terms can reduce precision if the added concepts are not relevant.
Summary
Query Expansion broadens a query with related terms, entities or concepts to improve retrieval coverage. It is a classic IR technique that can increase recall while creating a trade-off with precision.