Advanced End-to-End Relation Extraction with Deep Neural Networks

使用深度神经网络进行高级端到端关系提取

• 批准号: 10200889
• 负责人: Venkata Naga Ramakanth Kavuluru
• 金额: $ 33.27万
• 依托单位: UNIVERSITY OF KENTUCKY
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-07-01 至 2024-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10200889
• 关键词: Adverse event; Architecture; Area; Benchmarking; Bioinformatics; Biology; Biomedical Research; Classification; Clinical; Code; Collaborations; Combination Drug Therapy; Communities; Complex; Computer software; Data; Data Set; Dependence; Disease; Distant; Drug Interactions; Encapsulated; Etiology; Evaluation; Fostering; Funding; Future; Generations; Genes; Gold; Growth; Hand; Heart; Information Retrieval; Information Sciences; Intramural Research; Joints; Knowledge Discovery; Label; Language; Lead; Link; Literature; Manuals; Maps; Measures; Methodology; Methods; Modeling; Molecular; Names; Natural Language Processing; Patients; Peer Review; Performance; Periodicity; Pharmaceutical Preparations; Physicians; Process; Psychological Transfer; Reporting; Research; Research Personnel; Resources; Review Literature; Scientist; Semantics; Software Tools; Source; Standardization; Structure; Supervision; System; Terminology; Testing; Text; Training; Translational Research; Trees; base; biomedical data science; clinical care; deep neural network; improved; insight; interest; knowledge base; machine learning method; natural language; neural network; neural network architecture; new therapeutic target; novel; off-label use; protein protein interaction; relating to nervous system; side effect; social media; supervised learning; syntax

ABSTRACT Relations linking various biomedical entities constitute a crucial resource that enables biomedical data science applications and knowledge discovery. Relational information spans the translational science spectrum going from biology (e.g., protein–protein interactions) to translational bioinformatics (e.g., gene–disease associations), and eventually to clinical care (e.g., drug–drug interactions). Scientists report newly discovered relations in nat- ural language through peer-reviewed literature and physicians may communicate them in clinical notes. More recently, patients are also reporting side-effects and adverse events on social media. With exponential growth in textual data, advances in biomedical natural language processing (BioNLP) methods are gaining prominence for biomedical relation extraction (BRE) from text. Most current efforts in BRE follow a pipeline approach containing named entity recognition (NER), entity normalization (EN), and relation classification (RC) as subtasks. They typically suffer from error snowballing — errors in a component of the pipeline leading to more downstream errors — resulting in lower performance of the overall BRE system. This situation has lead to evaluation of different BRE substaks conducted in isolation. In this proposal we make a strong case for strictly end-to-end evaluations where relations are to be produced from raw text. We propose novel deep neural network architectures that model BRE in an end-to-end fashion and directly identify relations and corresponding entity spans in a single pass. We also extend our architectures to n-ary and cross-sentence settings where more than two entities may need to be linked even as the relation is expressed across multiple sentences. We also propose to create two new gold standard BRE datasets, one for drug–disease treatment relations and another first of a kind dataset for combination drug therapies. Our main hypothesis is that our end-to-end extraction models will yield supe- rior performance when compared with traditional pipelines. We test this through (1). intrinsic evaluations based on standard performance measures with several gold standard datasets and (2). extrinsic application oriented assessments of relations extracted with use-cases in information retrieval, question answering, and knowledge base completion. All software and data developed as part of this project will be made available for public use and we hope this will foster rigorous end-to-end benchmarking of BRE systems.

抽象的 连接各种生物医学实体的关系构建了一个至关重要的资源，可以实现生物医学数据科学 应用和知识发现。关系信息跨越转化科学范围 从生物学（例如蛋白质 - 蛋白质相互作用）到翻译的生物信息学（例如，基因 - 疾病酶相关）， 有时进行临床护理（例如，药物与药物相互作用）。科学家报告说，新发现的关系 通过同行评审的文献和医生可以通过临床注释将其传达给乌拉尔语。更多的 最近，患者还报告了社交媒体上的副作用和不良事件。随着指数增长 文本数据，生物医学自然语言处理的进步（BIONLP）方法已获得突出的 文本中的生物医学关系提取（BRE）。 BRE目前的大多数努力遵循包含的管道方法 命名实体识别（NER），实体归一化（EN）和关系分类（RC）为子任务。他们 通常遭受滚雪球错误 - 管道组成部分中的错误导致下游错误 - 导致整体BRE系统的性能较低。这种情况导致评估不同 BRE替代分离进行。在此提案中，我们为严格的端到端评估做出了有力的理由 从原始文本产生关系的地方。我们提出了新颖的深神经网络体系结构 以端到端的方式进行模型BRE，并直接识别单一的关系和相应的实体跨越 经过。我们还将架构扩展到n- ary和跨句子设置，其中有两个以上的实体可能 即使关系在多个句子中表示，也需要链接。我们还建议创建两个 新的黄金标准BRE数据集，一个用于药物 - 疾病治疗关系的数据集，另一个是同类数据集 用于联合药物疗法。我们的主要假设是，我们的端到端提取模型将产生supe- 与传统管道相比，Rior性能。我们通过（1）对此进行测试。基于内在评估 在具有多个黄金标准数据集和（2）的标准性能度量上。面向外部应用 评估与信息检索，问题答案和知识中用用例提取的关系的评估 基础完成。作为该项目一部分开发的所有软件和数据将用于公众使用， 我们希望这将促进BRE系统的严格端到端基准测试。