DynaRAG: Bridging Static and Dynamic Knowledge in Retrieval-Augmented Generation

TL;DR

DynaRAG is a novel retrieval-augmented generation framework that effectively combines static knowledge retrieval with dynamic API invocation to improve accuracy and reduce hallucinations in real-world question answering.

Contribution

It introduces a dynamic knowledge integration approach with an LLM-based reranker, sufficiency classifier, and schema filtering, advancing retrieval-augmented generation capabilities.

Findings

Significantly improves accuracy on dynamic questions

Reduces hallucinations in generated answers

Enhances robustness with schema filtering

Abstract

We present DynaRAG, a retrieval-augmented generation (RAG) framework designed to handle both static and time-sensitive information needs through dynamic knowledge integration. Unlike traditional RAG pipelines that rely solely on static corpora, DynaRAG selectively invokes external APIs when retrieved documents are insufficient for answering a query. The system employs an LLM-based reranker to assess document relevance, a sufficiency classifier to determine when fallback is necessary, and Gorilla v2 -- a state-of-the-art API calling model -- for accurate tool invocation. We further enhance robustness by incorporating schema filtering via FAISS to guide API selection. Evaluations on the CRAG benchmark demonstrate that DynaRAG significantly improves accuracy on dynamic questions, while also reducing hallucinations. Our results highlight the importance of dynamic-aware routing and selective tool use in building reliable, real-world question-answering systems.