Definition
Iterative retrieval is a multi-step search strategy where an AI agent or RAG pipeline performs successive rounds of querying, using the insights or missing information identified in previous retrieval steps to refine and launch subsequent searches.
Not to be confused with 'Single-shot RAG'; iterative retrieval involves a feedback loop where the agent decides if it has enough information to answer or if it needs to search again.
"A researcher in a library who finds one book, checks its bibliography, and then goes back to the stacks to find the referenced sources to fill in knowledge gaps."
- Query Decomposition(Component)
- Multi-hop Reasoning(Contextual Framework)
- Self-RAG(Optimization Pattern)
Conceptual Overview
Iterative retrieval is a multi-step search strategy where an AI agent or RAG pipeline performs successive rounds of querying, using the insights or missing information identified in previous retrieval steps to refine and launch subsequent searches.
Disambiguation
Not to be confused with 'Single-shot RAG'; iterative retrieval involves a feedback loop where the agent decides if it has enough information to answer or if it needs to search again.
Visual Analog
A researcher in a library who finds one book, checks its bibliography, and then goes back to the stacks to find the referenced sources to fill in knowledge gaps.