OPIEC

OPIEC是一个从英文维基百科全文中提取的大型开放信息抽取语料库，由德国曼海姆大学创建。该数据集包含超过3.4亿个三元组，是迄今为止公开可用的最大的OIE语料库。OPIEC不仅数据量大，还包含丰富的元数据信息，如来源信息、置信度评分、语言标注和语义标注，包括空间和时间信息。这些数据对于下游任务如知识库构建、开放式问答或事件模式归纳非常有价值。OPIEC的创建过程涉及使用Stanford CoreNLP管道和先进的OIE系统MinIE进行文本处理。数据集的应用领域广泛，主要用于自动化知识库构建的研究，旨在解决从自然语言文本中无监督地提取关系及其论据的问题。OPIEC is a large-scale open information extraction (OIE) corpus extracted from the full text of English Wikipedia, created by the University of Mannheim in Germany. This dataset contains over 340 million triples, making it the largest publicly available OIE corpus to date. Apart from its massive data scale, OPIEC also includes rich metadata such as source information, confidence scores, linguistic annotations and semantic annotations including spatial and temporal information. These data hold considerable value for downstream tasks like knowledge base construction, open-ended question answering or event pattern induction. The creation of OPIEC involved text processing using the Stanford CoreNLP pipeline and the state-of-the-art OIE system MinIE. The dataset has a wide range of application fields, and is mainly used in research on automated knowledge base construction, aiming to solve the problem of unsupervised extraction of relations and their arguments from natural language texts.

搜集汇总

背景与挑战

背景概述

OPIEC是一个基于整个英文维基百科构建的开放信息抽取（OIE）语料库，包含超过3.41亿个三元组，每个三元组都带有丰富的元数据，如NLP注释、来源句子和链接信息。该数据集旨在支持知识库构建和NLP研究，提供多个版本（如OPIEC-Clean和OPIEC-Link）以适应不同需求，并采用开放许可证（CC-BY-SA和GNU自由文档许可证）促进学术使用。