[Submitted on 17 Apr 2026 (v1), last revised 28 Jul 2026 (this version, v2)]

View PDF HTML (experimental)

Abstract: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.

Submission history

From: Lecheng Zheng [view email]
[v1] Fri, 17 Apr 2026 00:09:01 UTC (497 KB)
[v2] Tue, 28 Jul 2026 19:00:01 UTC (498 KB)