arXiv Open Access 2024

AutoMIR: Effective Zero-Shot Medical Information Retrieval without Relevance Labels

Lei Li Xiangxu Zhang Xiao Zhou Zheng Liu
Lihat Sumber

Abstrak

Medical information retrieval (MIR) is essential for retrieving relevant medical knowledge from diverse sources, including electronic health records, scientific literature, and medical databases. However, achieving effective zero-shot dense retrieval in the medical domain poses substantial challenges due to the lack of relevance-labeled data. In this paper, we introduce a novel approach called \textbf{S}elf-\textbf{L}earning \textbf{Hy}pothetical \textbf{D}ocument \textbf{E}mbeddings (\textbf{SL-HyDE}) to tackle this issue. SL-HyDE leverages large language models (LLMs) as generators to generate hypothetical documents based on a given query. These generated documents encapsulate key medical context, guiding a dense retriever in identifying the most relevant documents. The self-learning framework progressively refines both pseudo-document generation and retrieval, utilizing unlabeled medical corpora without requiring any relevance-labeled data. Additionally, we present the Chinese Medical Information Retrieval Benchmark (CMIRB), a comprehensive evaluation framework grounded in real-world medical scenarios, encompassing five tasks and ten datasets. By benchmarking ten models on CMIRB, we establish a rigorous standard for evaluating medical information retrieval systems. Experimental results demonstrate that SL-HyDE significantly surpasses HyDE in retrieval accuracy while showcasing strong generalization and scalability across various LLM and retriever configurations. Our code and data are publicly available at: https://github.com/ll0ruc/AutoMIR

Topik & Kata Kunci

Penulis (4)

L

Lei Li

X

Xiangxu Zhang

X

Xiao Zhou

Z

Zheng Liu

Format Sitasi

Li, L., Zhang, X., Zhou, X., Liu, Z. (2024). AutoMIR: Effective Zero-Shot Medical Information Retrieval without Relevance Labels. https://arxiv.org/abs/2410.20050

Akses Cepat

Lihat di Sumber
Informasi Jurnal
Tahun Terbit
2024
Bahasa
en
Sumber Database
arXiv
Akses
Open Access ✓