---
id: 20260402-T0-11
title: "PAR$^2$-RAG：多跳问答需主动检索推理"
title_en: "PAR$^2$-RAG improves multi-hop QA with active retrieval"
url: https://ai.daily.yangsir.net/daily/20260402-T0-11
issue_date: 2026-04-02
publish_date: 2026-04-01T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2603.29085
---

# PAR$^2$-RAG：多跳问答需主动检索推理

PAR$^2$-RAG通过主动检索和推理解决大模型在多跳问答中的脆弱性。现有迭代检索系统可能因早期低召回率锁定错误路径，导致推理失败。该架构在多跳任务中减少错误路径63%，提升推理准确性，让复杂问题分解更可靠。

## English Version

**PAR$^2$-RAG improves multi-hop QA with active retrieval**

PAR$^2$-RAG addresses LLM brittleness in multi-hop QA via active retrieval and reasoning. The method reduces error paths by 63% in complex tasks by preventing early low-recall lock-ins, making multi-step reasoning more reliable for complex queries.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2603.29085)

**详情页**：https://ai.daily.yangsir.net/daily/20260402-T0-11

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*