TL;DR本文旨在通过运用多阶段关注机制和序列到序列模型,在长文档的背景下产生更真实的问题,结果表明在 SQuAD、MS MARCO 和 NewsQA 三个问答数据集上,该方法优于现有方法。
Abstract
automatic question generation can benefit many applications ranging from
dialogue systems to reading comprehension. While questions are often asked with
respect to long documents, there are many challenges with m