Automatic question generation (QG) is a useful yet challenging task in NLP.
Recent neural network-based approaches represent the state-of-the-art in this
task. In this work, we attempt to strengthen them significantly by adopting a
holistic and novel generator-evaluator framework that directly optimizes
objectives that reward semantics and structure. The {\it generator} is a
sequence-to-sequence model that incorporates the {\it structure} and {\it
semantics} of the question being generated. The generator predicts an answer in
the passage that the question can pivot on. Employing the copy and coverage
mechanisms, it also acknowledges other contextually important (and possibly
rare) keywords in the passage that the question needs to conform to, while not
redundantly repeating words. The {\it evaluator} model evaluates and assigns a
reward to each predicted question based on its conformity to the {\it
structure} of ground-truth questions. We propose two novel QG-specific reward
functions for text conformity and answer conformity of the generated question.
The evaluator also employs structure-sensitive rewards based on evaluation
measures such as BLEU, GLEU, and ROUGE-L, which are suitable for QG. In
contrast, most of the previous works only optimize the cross-entropy loss,
which can induce inconsistencies between training (objective) and testing
(evaluation) measures. Our evaluation shows that our approach significantly
outperforms state-of-the-art systems on the widely-used SQuAD benchmark as per
both automatic and human evaluation.

采用新颖的生成 - 评估框架，结合结构和语义优化自动生成问题，采用生成器和评估器模型，引入复制和覆盖机制，提出两个新的针对 QG 特定的奖励函数，综合利用 BLEU，GLEU 和 ROUGE-L 等评价指标，在 SQuAD 基准测试上，相较于最先进的系统，自动和人工评估都表现出更好的性能。