arXiv：cs.AI· Lena Schmidt, Oshin Sharma, Chris Marshall, Sonia Garcia Gonzalez Moral·· 23 小时前
研究：SCANAR 与 AIDOC 可加速医疗领域的 Horizon Scanning
Horizon Scans can be accelerated using novel information retrieval and artificial intelligence tools
AI 导读
开源 Python 工具 SCANAR 和 AIDOC 被用于提升医疗 Horizon Scanning 的检索与筛选效率。SCANAR 通过去重和无监督相关性排序实现新闻文章检索与处理的自动化；AIDOC 借助大语言模型和文本嵌入计算语义相似度，在 95% 召回率下理论上可减少约 65% 的人工审查量，叠加已有开源算法 CMH 后减少 20%。
正文
Abstract:Introduction: Horizon scanning in healthcare assesses early signals of innovation, crucial for timely adoption. Current horizon scanning faces challenges in efficient information retrieval and analysis, especially in grey literature, presenting a need for innovative tools.
Methodology: This article introduces SCANAR and AIDOC, open-source Python-based tools designed to improve horizon scanning. SCANAR automates the retrieval and processing of news articles, using de-duplication and unsupervised relevancy ranking. AIDOC aids filtration by leveraging large language models and text embeddings for calculation of semantic similarity, to reorder data based on relevancy.
Results: Twelve internal datasets from horizon scans and four external benchmarking datasets were used. SCANAR improved retrieval efficiency by automating processes previously dependent on manual labour. AIDOC displayed work-saving potential, achieving around 65% theoretic reduction in manual review efforts at 95% recall or 20% reduction when adding CMH, an existing open-source algorithm to make data-driven decisions on stopping screening early. Comparative analysis with benchmarking data showed AIDOC's performance was similar to existing systematic review automation tools, though performance varied depending on dataset characteristics.
Conclusion: The validation indicates that SCANAR and AIDOC show potential to enhance horizon scanning efficiency by streamlining data retrieval and prioritisation. These tools may alleviate methodological limitations and allow broader, swifter horizon scans. Further studies are suggested to optimize these models and to design new workflows and validation processes that integrate large language models.
| Subjects: | Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2504.01627 [cs.IR] |
| (or arXiv:2504.01627v2 [cs.IR] for this version) | |
| https://doi.org/10.48550/arXiv.2504.01627 arXiv-issued DOI via DataCite |
Submission history
From: Lena Schmidt [view email]
[v1]
Wed, 2 Apr 2025 11:33:08 UTC (1,079 KB)
[v2]
Sun, 11 Oct 2026 11:21:30 UTC (2,402 KB)
来源：arXiv：cs.AI · arxiv.org