Search papers, labs, and topics across Lattice.
This paper introduces DocNavRAG, a novel framework that enhances complex document question answering by leveraging a navigable graph structure to organize document hierarchies and relationships. By allowing agents to traverse this graph and maintain an evolving evidence state, DocNavRAG significantly improves the quality and sufficiency of answers compared to existing methods. The results demonstrate an average improvement of 7.8% in answer quality and 17.7% in context sufficiency across four long- and multi-document QA benchmarks.
Navigating document structures with a stateful evidence approach boosts answer quality by nearly 8% in complex question answering tasks.
Answering complex questions over large document collections requires assembling complementary evidence across sections and documents. GraphRAG offers structured retrieval but typically uses fixed traversal, while agentic RAG operates over weakly structured interfaces. Our key insight is that agents should navigate document structure within and across documents rather than repeatedly search from scratch. We introduce DocNavRAG, which organizes document hierarchies and cross-region relations into a navigable graph, exposes graph operations for locating, navigating, expanding, and fetching, and maintains an evolving evidence state to guide retrieval until sufficient evidence is collected. Across four long- and multi-document QA benchmarks, DocNavRAG improves answer quality and context sufficiency over the strongest baseline by 7.8\% and 17.7\% on average.
ETH