融合静态与 Agentic RAG：面向台湾历史问答的后验选择方法
Bridging Static and Agentic RAG for Taiwanese Historical Question Answering
AI 导读
一项新研究在共享生成器与混合检索后端的受控条件下，比较 Agentic RAG 与静态 RAG 在台湾历史问答上的表现。两者总体性能接近，但在 70.83% 的问题上答案不同；逐题择优的 oracle 可将综合得分提升 0.2417。
正文 · 原文
Abstract:Agentic retrieval-augmented generation (RAG) enables language models to adapt retrieval based on previously retrieved evidence, but it remains unclear whether such adaptive orchestration consistently outperforms well-designed static pipelines. We conduct a controlled comparison of agentic and static RAG for Taiwanese historical question answering, sharing the same generator and hybrid retrieval backend. Despite similar aggregate performance, the two pipelines differ on 70.83% of questions, with their advantages largely canceling out when averaged. An oracle that selects the better response per question improves the composite score by 0.2417 over the better individual pipeline, revealing substantial headroom for question-level selection. We therefore introduce a post-hoc selector that compares the two responses and their cited evidence, significantly outperforming either individual pipeline and recovering 60.34% of the oracle headroom. These results show that aggregate comparisons can obscure meaningful question-level differences between retrieval strategies, suggesting that exploiting their complementarity may be more fruitful than seeking a universally superior pipeline.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) |
| Cite as: | arXiv:2609.23056 [cs.CL] |
| (or arXiv:2609.23056v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2609.23056 arXiv-issued DOI via DataCite |
Submission history
From: Kai-Hsin Chen [view email]
[v1]
Sat, 19 Sep 2026 14:43:26 UTC (2,196 KB)
来源：arXiv cs.AI · arxiv.org