---

page_id: paper-belink kind: paper language: zh-CN source_id: src-cnbben-642525e50f33 source_path: papers/CNBEN/BeLink/BeLink Read Deep.md claim_id: clm-belink-001 claim_status: source_fact

BeLink Read Deep

一句话回忆

Model / method: A BEL pipeline using re-ranking stage through generative models

研究卡

• 研究问题：Biomedical Entity Linking (BEL) with large language models (LLMs) remains computationally inefficient and challenging to deploy in practical settings.
• 核心方法：A BEL pipeline using re-ranking stage through generative models
• 主要结果：Our method demonstrates strong performance on multiple BEL benchmarks, yielding significant improvements in linking accuracy (3%–24%) while reducing inference time compared to the state-of-the-art.
• 数据集：待补证据
• Baseline：待补证据
• 指标：待补证据
• 局限：待补证据

对当前课题的作用

• 待根据原始笔记中明确的 Relation to this project 或研究计划证据补充；当前不作无证据推断。

关键问答

<details><summary>BeLink Read Deep 的核心方法是什么？</summary>

A BEL pipeline using re-ranking stage through generative models

</details>

字段证据

• 一句话回忆：papers/CNBEN/BeLink/BeLink Read Deep.md（line:37；clm-belink-005）
• 研究问题：papers/CNBEN/BeLink/BeLink Read Deep.md（line:38；clm-belink-002）
• 核心方法：papers/CNBEN/BeLink/BeLink Read Deep.md（line:37；clm-belink-003）
• 主要结果：papers/CNBEN/BeLink/BeLink Read Deep.md（line:40；clm-belink-004）

证据

• papers/CNBEN/BeLink/BeLink Read Deep.md（source_id: src-cnbben-642525e50f33）