发表机构
Beijing Institute of Technology（北京理工大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对多实体组合问题，传统方法有局限。本文提出HyGRL框架，将文本嵌入知识图谱创建异构网络，通过模仿学习和强化学习进行自适应结构归纳推理，实验证明其在答案准确性、推理保真度等方面表现出色，成本低且推理快。
AI 中文摘要
多实体组合问题给现有的检索增强语言模型带来了重大挑战。传统方法陷入两难境地：标准检索增强生成（RAG）缺乏动态推理，传统图结构的RAG受结构稀疏性限制，基于语言模型（LLM）构建的图结构RAG成本过高。我们提出了HyGRL，一个统一框架，将非结构化文本嵌入到结构化知识图谱中，创建一个异构网络以进行灵活的证据检索。推理被制定为自适应结构归纳，通过一个强大的两阶段过程学习：（1）模仿学习提取启发式专家信号，（2）强化学习使用LLM驱动的偏好奖励来优化策略。实验表明，HyGRL有效地将文本丰富性与结构知识相结合，在答案准确性和推理保真度方面优于当前最优基线，同时保持极低的令牌成本和近实时推理。
英文摘要
Multi-entity compositional questions pose significant challenges to existing retrieval-augmented language models. Conventional methods fall into a dilemma: standard RAG lacks dynamic reasoning, traditional Graph-RAG is limited by structural sparsity, and LLM-constructed Graph-RAG incurs prohibitive costs. We propose \textbf{\fwa}, a unified framework that embeds unstructured text into structured knowledge graphs, creating a heterogeneous network for flexible evidence retrieval. Reasoning is formulated as adaptive structure induction, learned via a robust two-stage process: (1) imitation learning distills heuristic expert signals, and (2) reinforcement learning refines the policy using LLM-driven preference rewards. Experiments demonstrate that {\fwa} effectively merges textual richness with structural knowledge, outperforming SOTA baselines in answer accuracy and reasoning fidelity while maintaining extremely low token costs and near real-time inference((code available at https://github.com/wjywjy123/HyGRL) .
Comments8 pages
