Reranking
Reordering search results using a more sophisticated model to improve relevance.
Definition
Reranking is a technique that takes initial search results and reorders them using a more sophisticated relevance model. While first-stage retrieval prioritises speed to search through large document collections, reranking applies more computationally intensive analysis to a smaller set of candidates. This two-stage approach balances efficiency with accuracy, significantly improving the quality of top results in RAG and search systems.
Related terms
Semantic Search
Search that understands the meaning of queries rather than just matching keywords.
Retrieval Augmented Generation (RAG)
A technique that grounds AI responses in retrieved documents for accurate, cited answers.
Hybrid Search
Combining keyword and semantic search for more comprehensive results.
More in Data Infrastructure
Chunking
Splitting documents into smaller segments for processing and retrieval.
Embedding
A numerical representation of text that captures its semantic meaning.
Vector Database
A database optimised for storing and querying high-dimensional vector data.
Knowledge Graph
A structured representation of entities and their relationships.
See Reranking in action
Understanding the terminology is the first step. See how Conductor applies these concepts to solve real document intelligence challenges.
Request a demo