NaviRAG: Towards Active Knowledge Navigation for Retrieval-Augmented Generation

2026-04-14

AI Summary

This work addresses the limitations of traditional retrieval-augmented generation (RAG) systems, which rely on static, flat retrieval mechanisms and struggle with complex tasks requiring dynamic integration of information across multiple granularities. To overcome this, the authors propose an active retrieval framework grounded in a hierarchical knowledge structure, wherein a large language model agent iteratively identifies information gaps and dynamically plans and executes multi-granularity retrieval strategies to achieve conditional, goal-directed knowledge acquisition. Evaluated on long-document question answering benchmarks, the approach significantly improves both retrieval recall and end-to-end answer quality. Ablation studies further confirm the effectiveness of the hierarchical organization and active navigation mechanisms underlying the framework.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Knowledge Representation and Reasoning: Knowledge Acquisition
• Machine Learning: Active Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Retrieval-augmented generation (RAG) typically relies on a flat retrieval paradigm that maps queries directly to static, isolated text segments. This approach struggles with more complex tasks that require the conditional retrieval and dynamic synthesis of information across different levels of granularity (e.g., from broad concepts to specific evidence). To bridge this gap, we introduce NaviRAG, a novel framework that shifts from passive segment retrieval to active knowledge navigation. NaviRAG first structures the knowledge documents into a hierarchical form, preserving semantic relationships from coarse-grained topics to fine-grained details. Leveraging this reorganized knowledge records, a large language model (LLM) agent actively navigates the records, iteratively identifying information gaps and retrieving relevant content from the most appropriate granularity level. Extensive experiments on long-document QA benchmarks show that NaviRAG consistently improves both retrieval recall and end-to-end answer performance over conventional RAG baselines. Ablation studies confirm performance gains stem from our method's capacity for multi-granular evidence localization and dynamic retrieval planning. We further discuss efficiency, applicable scenario, and future directions of our method, hoping to make RAG systems more intelligent and autonomous.

Problem

• Retrieval-Augmented Generation
• active knowledge navigation
• multi-granular retrieval
• complex question answering
• hierarchical knowledge

Innovation

• active knowledge navigation
• hierarchical retrieval
• multi-granular evidence localization
• dynamic retrieval planning
• retrieval-augmented generation