Elaborative Simplification: Content Addition and Explanation Generation in Text Simplification
Elaborative Simplification: Content Addition and Explanation Generation in Text Simplification
复制标题
DOI：
10.18653/v1/2021.findings-acl.455
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Neha Srikanth;Junyi Jessy Li
中科院分区：
文献类型：
--
作者：
Neha Srikanth;Junyi Jessy Li
现代文本简化的研究主要集中在简化层次上，将原始的、更复杂的句子转化为简化的版本。然而，当需要解释困难的概念和推理时，添加内容通常是有用的。在这项工作中，我们提出了第一个数据驱动的研究内容添加在文档简化，我们称之为精心简化。我们引入了一个新的注释数据集的1.3K例精心简化和分析实体，想法和概念是如何通过上下文特异性的透镜详细说明。我们使用大规模预训练的语言模型建立了细化生成的基线，并说明在生成过程中考虑上下文特异性可以提高性能。我们的结果说明了精心简化的复杂性，并为未来的工作提出了许多有趣的方向。
Much of modern day text simplification research focuses on sentence-level simplification, transforming original, more complex sentences to simplified versions. However, adding content can often be useful when difficult concepts and reasoning need to be explained. In this work, we present the first data-driven study of content addition in document simplification, which we call elaborative simplification. We introduce a new annotated dataset of 1.3K instances of elaborative simplification and analyze how entities, ideas, and concepts are elaborated through the lens of contextual specificity. We establish baselines for elaboration generation using large scale pre-trained language models, and illustrate that considering contextual specificity during generation can improve performance. Our results illustrate the complexities of elaborative simplification, suggesting many interesting directions for future work.