PAR$^2$-RAG：计划主动检索与推理用于多跳问答
PAR$^2$-RAG: Planned Active Retrieval and Reasoning for Multi-Hop Question Answering
AI总结：
PAR$^2$-RAG通过分阶段的覆盖与承诺分离，结合广度优先锚定和深度优先细化，提升多跳问答的准确率和检索效果。
摘要：
大型语言模型在多跳问答任务中仍存在脆弱性，需要通过检索和推理整合跨文档证据。迭代检索系统可能因早期低召回路径而放大下游错误，而仅规划的方法可能产生无法适应中间证据变化的静态查询集。我们提出计划主动检索与推理RAG（PAR$^2$-RAG），一种两阶段框架，将覆盖与承诺分离。PAR$^2$-RAG首先通过广度优先锚定构建高召回证据前沿，然后在迭代循环中应用深度优先细化并控制证据充分性。在四个多跳问答基准测试中，PAR$^2$-RAG在准确率上优于现有最先进基线，相较于IRCoT，准确率提高23.5%，检索效果在NDCG上提升10.5%。
英文摘要：
Large language models (LLMs) remain brittle on multi-hop question answering (MHQA), where answering requires combining evidence across documents through retrieval and reasoning. Iterative retrieval systems can fail by locking onto an early low-recall trajectory and amplifying downstream errors, while planning-only approaches may produce static query sets that cannot adapt when intermediate evidence changes. We propose \textbf{Planned Active Retrieval and Reasoning RAG (PAR$^2$-RAG)}, a two-stage framework that separates \emph{coverage} from \emph{commitment}. PAR$^2$-RAG first performs breadth-first anchoring to build a high-recall evidence frontier, then applies depth-first refinement with evidence sufficiency control in an iterative loop. Across four MHQA benchmarks, PAR$^2$-RAG consistently outperforms existing state-of-the-art baselines, compared with IRCoT, PAR$^2$-RAG achieves up to \textbf{23.5\%} higher accuracy, with retrieval gains of up to \textbf{10.5\%} in NDCG.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
