Improving paragraph-level question generation with extended answer network and uncertainty-aware beam search

作     者：Zeng, Hongwei Zhi, Zhuo Liu, Jun Wei, Bifan 

作者机构：Xi An Jiao Tong Univ Sch Comp Sci & Technol Xian 710049 Shaanxi Peoples R China Xi An Jiao Tong Univ Natl Engn Lab Big Data Analyt Xian 710049 Shaanxi Peoples R China Xi An Jiao Tong Univ Sch Elect Engn Xian 710049 Shaanxi Peoples R China Xi An Jiao Tong Univ Sch Continuing Educ Xian 710049 Shaanxi Peoples R China 

出 版 物：《INFORMATION SCIENCES》 (信息科学)

年 卷 期：2021年第571卷

页      面：50-64页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：National Key Research and Development Program of China [2020AAA0108800] National Natural Science Foundation of China [61877050, 61937001, 62020194] Innovative Research Group of the National Natural Science Foundation of China Innovation Research Team of Ministry of Education [IRT_17R86] Project of China Knowledge Centre for Engineering Science and Technology, MoE-CMCC "Artifical Intelligence" Project [MCM20190701] National Social Science Fund of China [18XXW005] National Statistical Science Research Project [2020LY103] Ministry of Education Humanities and Social Sciences Fund [17YJA860028] 

主　　题：Question generation Attention network Coverage mechanism Beam search 

摘      要：Question Generation (QG), which aims to generate a question given the relevant context, is essential to build conversational and question-answering systems. Existing neural question generation models suffer from the inadequate representation of the target answer and inappropriate techniques to reduce repetition. To address these issues, we propose an Extended Answer-aware Network (EAN) which is trained with Word-based Coverage Mechanism (WCM) and decoded with Uncertainty-aware Beam Search (UBS). The EAN represents the target answer by its surrounding sentence with an encoder and incorporates the extended answer to paragraph representation using gated paragraph-to-answer attention to tackle the problem of the inadequate representation of the target answer. To reduce undesirable repetition, the WCM penalizes repeatedly attending to the same words of different time-steps in the training stage. The UBS incorporates an uncertainty score into beam search to alleviate text degeneration and reduce repeated copying words of the paragraph. Experiments on two benchmark datasets demonstrate the effectiveness of our methods of paragraph-level question generation. Specifically, our model has achieved 4.2% and 27.2% improvement over BLEU-4 compared to the best paragraph-level QG baseline in SQuAD and NewsQA datasets respectively. (c) 2021 Elsevier Inc. All rights reserved.