用医学主题词层级标签提升生物医学检索的语义理解能力
BioHiCL: Hierarchical Multi-Label Contrastive Learning for Biomedical Retrieval with MeSH Labels
- 基于MeSH主题词层级结构设计多标签对比学习
- 在0.1B和0.3B参数模型上实现高效精准检索
- 适合需要细粒度医学语义理解的研究者使用
有效的生物医学信息检索需要建模领域语义及生物医学文本间的层次关系。现有生物医学生成式检索器依赖粗粒度的二元相关性信号，限制了对语义重叠的捕捉能力。我们提出BioHiCL（基于层次化多标签对比学习的生物医学检索），利用层次化的MeSH标注为多标签对比学习提供结构化监督。所提出的模型BioHiCL-Base（0.1B）和BioHiCL-Large（0.3B）在生物医学检索、句子相似性与问答任务上表现优异，同时保持部署时的计算效率。
原文摘要 · Abstract (English)
Effective biomedical information retrieval requires modeling domain semantics and hierarchical relationships among biomedical texts. Existing biomedical generative retrievers build on coarse binary relevance signals, limiting their ability to capture semantic overlap. We propose BioHiCL (Biomedical Retrieval with Hierarchical Multi-Label Contrastive Learning), which leverages hierarchical MeSH annotations to provide structured supervision for multi-label contrastive learning. Our models, BioHiCL-Base (0.1B) and BioHiCL-Large (0.3B), achieve promising performance on biomedical retrieval, sentence similarity, and question answering tasks, while remaining computationally efficient for deployment.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。