SciERC
收藏资源简介:
我们在科学文章中引入了识别实体、关系和共指集群的多任务设置。我们创建了SCIERC,这是一个包含所有三个任务的注释的数据集,并开发了一个名为SCIIE的统一框架,具有共享跨度表示。多任务设置减少了任务之间的级联错误,并通过共指链接利用了跨句关系。实验表明,我们的多任务模型在科学信息提取方面优于以前的模型,而无需使用任何特定领域的特征。我们进一步表明,该框架支持构建科学知识图谱,我们用它来分析信息非科学文献。
We introduce a multi-task setting for identifying entities, relations, and coreference clusters in scientific articles. We created SCIERC, a dataset annotated with all three tasks, and developed a unified framework named SCIIE that features shared span representations. This multi-task setting reduces cascading errors between tasks and leverages cross-sentence relations via coreference links. Experiments show that our multi-task model outperforms previous models in scientific information extraction without utilizing any domain-specific features. We further demonstrate that this framework supports the construction of scientific knowledge graphs, which we use to analyze information from non-scientific literature.




