DecoupleSearch: Decouple Planning and Search via Hierarchical Reward Modeling
This work addresses efficiency and accuracy issues in retrieval-augmented generation for AI systems, representing an incremental improvement over existing Agentic RAG methods.
The paper tackles challenges in Agentic RAG systems, such as interdependent planning and search, lack of supervision for intermediate steps, and large candidate spaces, by proposing DecoupleSearch, a framework that decouples these processes using dual value models and hierarchical beam search, achieving improved performance across various policy models.
Retrieval-Augmented Generation (RAG) systems have emerged as a pivotal methodology for enhancing Large Language Models (LLMs) through the dynamic integration of external knowledge. To further improve RAG's flexibility, Agentic RAG introduces autonomous agents into the workflow. However, Agentic RAG faces several challenges: (1) the success of each step depends on both high-quality planning and accurate search, (2) the lack of supervision for intermediate reasoning steps, and (3) the exponentially large candidate space for planning and searching. To address these challenges, we propose DecoupleSearch, a novel framework that decouples planning and search processes using dual value models, enabling independent optimization of plan reasoning and search grounding. Our approach constructs a reasoning tree, where each node represents planning and search steps. We leverage Monte Carlo Tree Search to assess the quality of each step. During inference, Hierarchical Beam Search iteratively refines planning and search candidates with dual value models. Extensive experiments across policy models of varying parameter sizes, demonstrate the effectiveness of our method.