ReOIE2016
收藏资源简介:
ReOIE2016是由上海交通大学计算机科学与工程系的研究人员重新标注的一个用于开放信息提取评估的基准数据集。该数据集包含600个句子,旨在解决原始OIE2016数据集中存在的标注问题,如过度指定提取和错误翻译。ReOIE2016通过手动重新标注,解决了这些问题,并提供了更准确的评估基准。此外,该数据集还标注了上下文信息,尽管这些信息在本研究中不用于评估,但为未来的研究提供了潜在的应用价值。
ReOIE2016 is a benchmark dataset for open information extraction (OIE) evaluation, re-annotated by researchers from the Department of Computer Science and Engineering, Shanghai Jiao Tong University. It contains 600 sentences and aims to address the annotation issues present in the original OIE 2016 dataset, such as over-specified extractions and incorrect translations. Through manual re-annotation, ReOIE2016 resolves these problems and provides a more accurate evaluation benchmark. Additionally, the dataset also annotates contextual information. Although this information is not utilized for evaluation in this study, it offers potential application value for future research.

- 1Span Model for Open Information Extraction on Accurate Corpus上海交通大学计算机科学与工程系 · 2019年



