AirRAG: Autonomous Strategic Planning and Reasoning Steer Retrieval Augmented Generation
This addresses the problem of limited solution spaces in RAG for complex reasoning tasks, offering a flexible and lightweight approach that is incremental over existing agentic RAG techniques.
The paper tackles the limitation of existing retrieval-augmented generation (RAG) methods being constrained to a single solution space for complex problems by proposing AirRAG, which integrates autonomous strategic planning with reasoning actions using Monte Carlo Tree Search (MCTS), resulting in significant performance gains on complex question-answering datasets.
Leveraging the autonomous decision-making capabilities of large language models (LLMs) has demonstrated superior performance in reasoning tasks. However, despite the success of iterative or agentic retrieval-augmented generation (RAG) techniques, these methods are often constrained to a single solution space when confronted with complex problems. In this paper, we propose a novel thinking pattern in RAG that integrates autonomous strategic planning with efficient reasoning actions, significantly activating intrinsic reasoning capabilities and expanding the solution space of specific tasks via Monte Carlo Tree Search (MCTS), which we refer to as AirRAG. Specifically, our approach designs five fundamental reasoning actions, which are expanded to a broad tree-based reasoning space using MCTS. The approach also incorporates self-consistency verification to explore potential reasoning paths and inference scaling law. Additionally, computationally optimal strategies are employed to allocate more inference resources to key actions, thereby enhancing overall performance. Experimental results demonstrate the effectiveness of AirRAG, showing significant performance gains on complex question-answering datasets. Furthermore, AirRAG is flexible and lightweight, making it easy to integrate with other advanced technologies and models.