About this document

RA-ISF: Enhanced Retrieval for QA by Divyesh is a document available to read on EtoBox.

1. The paper proposes a framework called Retrieval Augmented Iterative Self-Feedback (RA-ISF) to address limitations in existing retrieval augmented generation (RAG) models. 2. RA-ISF uses three sub-modules - a self-knowledge module, passage relevance module, and question decomposition module - to iteratively process questions using the model

Author

Divyesh

Language

EN