Rag Fusion - RFF Algorithm

This system is designed to enhance search results by generating multiple queries from a user's initial query, conducting vector-based searches, and re-ranking the results using the Reciprocal Rank Fusion (RRF) algorithm.

Architecture

Workflow

1. Initial Query: User provides an initial search query.
2. Query Generation: The system generates multiple related queries using the GPT model.
3. Vector Search: Each generated query is used to perform a vector-based search, retrieving documents from a predefined set.
4. Reciprocal Rank Fusion: Retrieved documents are re-ranked based on their relevance across multiple queries using the RRF algorithm.
5. Final Output: A re-ranked list of documents is produced, offering the most relevant search results.

License

For more information, please refer to the full research paper here.