arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新，休息一下吧，周末愉快~~
arXiv 2101.01337cs.CLcs.LG
收藏

使用医学知识图谱深度学习整合领域知识用于癌症表型分型

Integration of Domain Knowledge using Medical Knowledge Graph Deep Learning for Cancer Phenotyping

Wayne State University（韦恩州立大学）；Oak Ridge National Laboratory（橡树岭国家实验室）；Louisiana State University Health Sciences Center（路易斯安那州立大学健康科学中心）；University of Kentucky（肯塔基大学）；University of Utah School of Medicine（犹他大学医学院）
Mohammed Alawad, Shang Gao, Mayanka Chandra Shekar, S. M. Shamimul Hasan, J. Blair Christian, Xiao-Cheng Wu, Eric B. Durbin, Jennifer Doherty, Antoinette Stroup… 展开作者

Mohammed Alawad, Shang Gao, Mayanka Chandra Shekar, S. M. Shamimul Hasan, J. Blair Christian, Xiao-Cheng Wu, Eric B. Durbin, Jennifer Doherty, Antoinette Stroup, Linda Coyle, Lynne Penberthy, Georgia Tourassi

阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

摘要：

用于自然语言处理（NLP）的深度学习（DL）的一个关键组件是词嵌入。词嵌入能够有效地捕捉它们所代表的词的含义和上下文，可以显著提高下游DL模型在各种NLP任务中的性能。许多现有的词嵌入技术基于文档和文本中的词共现来捕获词的上下文;然而，它们通常不能捕获概念之间更广泛的特定于领域的关系，这对于手头的NLP任务可能是至关重要的。在本文中，我们提出了一种方法来整合外部知识的医学术语本体到上下文所捕获的词嵌入。具体而言，我们使用医学知识图谱，如统一医学语言系统（UMLS），以找到癌症病理报告中的临床术语之间的联系。这种方法旨在最大限度地减少连接的临床概念之间的距离。我们使用多任务卷积神经网络（MT-CNN）评估所提出的方法，以从约900 K癌症病理报告的数据集中提取六个癌症特征-部位，亚部位，偏侧性，行为，组织学和等级。结果表明，使用我们的域信息嵌入的MT-CNN模型在所有任务中的表现都优于使用标准word 2 vec嵌入的相同MT-CNN，整体微观和宏观F1得分分别提高了4.97%和22.5%。

英文摘要：

A key component of deep learning (DL) for natural language processing (NLP) is word embeddings. Word embeddings that effectively capture the meaning and context of the word that they represent can significantly improve the performance of downstream DL models for various NLP tasks. Many existing word embeddings techniques capture the context of words based on word co-occurrence in documents and text; however, they often cannot capture broader domain-specific relationships between concepts that may be crucial for the NLP task at hand. In this paper, we propose a method to integrate external knowledge from medical terminology ontologies into the context captured by word embeddings. Specifically, we use a medical knowledge graph, such as the unified medical language system (UMLS), to find connections between clinical terms in cancer pathology reports. This approach aims to minimize the distance between connected clinical concepts. We evaluate the proposed approach using a Multitask Convolutional Neural Network (MT-CNN) to extract six cancer characteristics -- site, subsite, laterality, behavior, histology, and grade -- from a dataset of ~900K cancer pathology reports. The results show that the MT-CNN model which uses our domain informed embeddings outperforms the same MT-CNN using standard word2vec embeddings across all tasks, with an improvement in the overall micro- and macro-F1 scores by 4.97\%and 22.5\%, respectively.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

↑