DualRAG: A Dual-Process Approach to Integrate Reasoning and Retrieval for Multi-Hop Question Answering

2025-04-25

AI Summary

Multi-hop question answering (MHQA) faces challenges in dynamically organizing cross-domain knowledge and performing coherent, multi-step reasoning. To address this, we propose DualRAG—a novel framework featuring two tightly coupled, bidirectional processes: Reasoning-augmented Query (RaQ) generation and progressive Knowledge Aggregation (pKA). DualRAG establishes the first tightly integrated, mutually reinforcing loop between reasoning and retrieval. It enables efficient lightweight fine-tuning of small language models, preserving both strong reasoning capabilities and precise retrieval performance. Leveraging LLM-driven iterative query rewriting, knowledge retrieval, and fusion, DualRAG achieves significant improvements in answer accuracy and reasoning coherence across mainstream MHQA benchmarks. Notably, in several settings, it matches or even surpasses the performance of oracle-knowledge baselines—demonstrating the critical role of dynamic knowledge organization in enabling robust multi-hop reasoning.

Technology Category

• Knowledge Representation and Reasoning: Knowledge Acquisition
• Search and Optimization: Metareasoning and Metaheuristics
• Data Mining & Knowledge Management: Intelligent Query Processing

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

Multi-Hop Question Answering (MHQA) tasks permeate real-world applications, posing challenges in orchestrating multi-step reasoning across diverse knowledge domains. While existing approaches have been improved with iterative retrieval, they still struggle to identify and organize dynamic knowledge. To address this, we propose DualRAG, a synergistic dual-process framework that seamlessly integrates reasoning and retrieval. DualRAG operates through two tightly coupled processes: Reasoning-augmented Querying (RaQ) and progressive Knowledge Aggregation (pKA). They work in concert: as RaQ navigates the reasoning path and generates targeted queries, pKA ensures that newly acquired knowledge is systematically integrated to support coherent reasoning. This creates a virtuous cycle of knowledge enrichment and reasoning refinement. Through targeted fine-tuning, DualRAG preserves its sophisticated reasoning and retrieval capabilities even in smaller-scale models, demonstrating its versatility and core advantages across different scales. Extensive experiments demonstrate that this dual-process approach substantially improves answer accuracy and coherence, approaching, and in some cases surpassing, the performance achieved with oracle knowledge access. These results establish DualRAG as a robust and efficient solution for complex multi-hop reasoning tasks.

Problem

• Integrating reasoning and retrieval for multi-hop QA
• Identifying and organizing dynamic knowledge effectively
• Enhancing answer accuracy and coherence in MHQA

Innovation

• Dual-process framework integrates reasoning and retrieval
• Reasoning-augmented Querying navigates reasoning paths
• Progressive Knowledge Aggregation systematically integrates knowledge