Inquisitive Question Generation for High Level Text Comprehension
Inquisitive Question Generation for High Level Text Comprehension
复制标题
DOI：
10.18653/v1/2020.emnlp-main.530
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Wei-Jen Ko;Tengyang Chen;Yiyan Huang;Greg Durrett;Junyi Jessy Li
中科院分区：
文献类型：
--
作者：
Wei-Jen Ko;Tengyang Chen;Yiyan Huang;Greg Durrett;Junyi Jessy Li
在各种环境中，探究性的探索问题对人类来说是很自然的，但对于自动系统来说是一项具有挑战性的任务。一种自然类型的问题试图填补文本理解过程中的知识空白，就像阅读一篇新闻文章：我们可能会问背景信息，事情发生背后的更深层次的原因，或者更多。尽管最近在数据驱动方法方面取得了进展，但生成此类问题超出了在现有数据集上训练的模型的范围。我们介绍INQUISITIVE，这是一个包含约19 K个问题的数据集，这些问题是在一个人阅读文档时引发的。与现有的数据集相比，INQUISITIVE问题更多地针对文本的高层次（语义和语篇）理解。我们发现，读者从事一系列的语用策略，以寻求信息。最后，我们评估了基于GPT-2的问题生成模型，并表明我们的模型能够生成合理的问题，虽然任务是具有挑战性的，并强调了上下文的重要性，以产生询问性问题。
Inquisitive probing questions come naturally to humans in a variety of settings, but is a challenging task for automatic systems. One natural type of question to ask tries to fill a gap in knowledge during text comprehension, like reading a news article: we might ask about background information, deeper reasons behind things occurring, or more. Despite recent progress with data-driven approaches, generating such questions is beyond the range of models trained on existing datasets. We introduce INQUISITIVE, a dataset of ~19K questions that are elicited while a person is reading through a document. Compared to existing datasets, INQUISITIVE questions target more towards high-level (semantic and discourse) comprehension of text. We show that readers engage in a series of pragmatic strategies to seek information. Finally, we evaluate question generation models based on GPT-2 and show that our model is able to generate reasonable questions although the task is challenging, and highlight the importance of context to generate INQUISITIVE questions.