---
格式版本: 2
标题: "AdaPath: Query-Adaptive Path-Finding via Path-Bank for Multi-Hop Implicit Biomedical KGQA"
原文链接: "https://arxiv.org/abs/2608.30556"
发布日期: "2026-08-31"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Mon, 31 Aug 2026 10:32:23 UTC (1,369 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-09-02T03:54:54+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-09-02T03:53:56+08:00"
入库时间: "2026-09-01T19:54:54.658Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=Meta&searchtype=all"
匹配关键词:
  - "AI"
相关厂家:
  - "Meta"
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 5
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "论文内容为生物医学知识图谱问答的路径查找方法，与超节点、AI Rack、机柜级AI基础设施等完全无关，仅来源为arXiv学术平台，但无任何相关技术或商业信息。"
AI质检模型: "zj-deepseek-v4-flash"
AI质检时间: "2026-09-02T03:54:58+08:00"
AI主题相关性: 0
AI来源权威性: 5
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
AI摘要: "研究者提出 AdaPath，一个基于路径库（Path-Bank）的查询自适应路径查找框架，用于多跳隐式生物医学知识图谱问答，针对查询缺少中间推理线索和知识图谱连接密集易出错两大难点进行优化。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-01T23:06:19.080Z"
采集批次: "2026年9月1日23点56分52秒"
采集批次ID: "20260901-235652-761"
去重键: "https://arxiv.org/abs/2608.30556"
---

## Computer Science > Artificial Intelligence

## Title:AdaPath: Query-Adaptive Path-Finding via Path-Bank for Multi-Hop Implicit Biomedical KGQA

Authors:[Jun Hyeong Kim](https://arxiv.org/search/cs?searchtype=author&query=Kim,+J+H), [Dongki Kim](https://arxiv.org/search/cs?searchtype=author&query=Kim,+D), [Yinhua Piao](https://arxiv.org/search/cs?searchtype=author&query=Piao,+Y), [Sung Ju Hwang](https://arxiv.org/search/cs?searchtype=author&query=Hwang,+S+J)

[View PDF](https://arxiv.org/pdf/2608.30556) [HTML (experimental)](https://arxiv.org/html/2608.30556v1)

> Abstract:Path-finding over knowledge graphs has become an effective way to ground LLM reasoning on multi-hop questions. However, biomedical QA introduces two distinct challenges that general-domain methods are not designed for: (i) queries do not expose intermediate reasoning and can be answered through multiple valid pathways, and (ii) biomedical knowledge graphs are densely connected, so path-finding methods easily take wrong turns. To address these challenges, we propose AdaPath, a path-finding framework that retrieves query-adaptive meta-paths from Path-Bank, which captures both query semantics and biomedical knowledge graph structure. AdaPath provides the missing cues in biomedical queries while effectively pruning dense knowledge graph neighborhoods during multi-hop reasoning. We further release BioStrat-QA, a biomedical KGQA benchmark that stratifies multi-hop queries by how much intermediate reasoning they expose. Across biomedical KGQA benchmarks, AdaPath consistently outperforms baselines, sustaining meaningful path-finding even when multi-hop queries expose less surface information. The source code is available at [this https URL](https://github.com/Jun-Hyeong-Kim/AdaPath).

| Comments: |  |
| --- | --- |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | [arXiv:2608.30556](https://arxiv.org/abs/2608.30556) \[cs.AI\] |
|  | (or [arXiv:2608.30556v1](https://arxiv.org/abs/2608.30556v1) \[cs.AI\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.30556](https://doi.org/10.48550/arXiv.2608.30556) |
| Journal reference: | EMNLP 2026 Main Conference |

## Submission history

From: Jun Hyeong Kim \[[view email](https://arxiv.org/show-email/5d5c6194/2608.30556)\]  
**\[v1\]** Mon, 31 Aug 2026 10:32:23 UTC (1,369 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.30556) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))