Educational Question Generation of Children Storybooks via Question Type Distribution Learning and Event-Centric Summarization
Zhenjie Zhao, Yufang Hou, Dakuo Wang, Mo Yu, Chengzhong Liu, Xiaojuan Ma
TL;DR
The paper tackles automatic generation of high-cognitive-demand educational questions for fairy tales. It introduces a three-stage framework: (1) learning the question-type distribution with a BERT-based predictor, (2) event-centric summarization conditioned on type and order, and (3) educational question generation from the summaries, trained with silver QA-derived data. The training objective combines KL-divergence over predicted type distributions and cross-entropy for multi-label prediction, via $L = \gamma L_{KL} + (1-\gamma) L_{CE}$. On the FairytaleQA dataset, the method shows superior automatic metrics and favorable human evaluations, highlighting the value of decomposing type distribution learning and event-centric summarization for educational QG.
Abstract
Generating educational questions of fairytales or storybooks is vital for improving children's literacy ability. However, it is challenging to generate questions that capture the interesting aspects of a fairytale story with educational meaningfulness. In this paper, we propose a novel question generation method that first learns the question type distribution of an input story paragraph, and then summarizes salient events which can be used to generate high-cognitive-demand questions. To train the event-centric summarizer, we finetune a pre-trained transformer-based sequence-to-sequence model using silver samples composed by educational question-answer pairs. On a newly proposed educational question answering dataset FairytaleQA, we show good performance of our method on both automatic and human evaluation metrics. Our work indicates the necessity of decomposing question type distribution learning and event-centric summary generation for educational question generation.