Reinforced Multi-task Approach for Multi-hop Question Generation

Proceedings of the 28th International Conference on Computational Linguistics, pages 2760–2775 Barcelona, Spain (Online), December 8-13, 2020 2760 Reinforced Multi-task Approach for Multi-hop Question Generation Deepak Gupta, Hardik Chauhan∗, Akella Ravi Tej∗, Asif Ekbal, and Pushpak Bhattacharyya Indian Institute of Technology Patna, India {deepak.pcs16,asif,pb}@iitp.ac.in {chauhanhardik23,ravitej.akella}@gmail.com Abstract Question generation (QG) attempts to solve the inverse of question answering (QA) problem by generating a natural language question given a document and an answer. While sequence to se- quence neural models surpass rule-based systems for QG, they are limited in their capacity to focus on more than one supporting fact. For QG, we often require multiple supporting facts to generate high-quality questions. Inspired by recent works on multi-hop reasoning in QA, we take up Multi-hop question generation, which aims at generating relevant questions based on supporting facts in the context. We employ multitask learning with the auxiliary task of answer- aware supporting fact prediction to guide the question generator. In addition, we also proposed a question-aware reward function in a Reinforcement Learning (RL) framework to maximize the utilization of the supporting facts. We demonstrate the effectiveness of our approach through ex- periments on the multi-hop question answering dataset, HotPotQA. Empirical evaluation shows our model to outperform the single-hop neural question generation models on both automatic evaluation metrics such as BLEU, METEOR, and ROUGE, and human evaluation metrics for quality and coverage of the generated questions. 1 Introduction In natural language processing (NLP), question generation is considered to be an important yet challeng- ing problem. Given a passage and answer as inputs to the model, the task is to generate a semantically coherent question for the given answer. In the past, question generation has been tackled using rule-based approaches such as question tem- plates (Lindberg et al., 2013) or utilizing named entity information and predictive argument structures of sentences (Chali and Hasan, 2015). Recently, neural-based approaches have accomplished impressive results (Du et al., 2017; Sun et al., 2018; Kim et al., 2018; Gupta et al., 2019b) for the task of ques- tion generation. The availability of large-scale machine reading comprehension datasets such as SQuAD (Rajpurkar et al., 2016), NewsQA (Trischler et al., 2017), MSMARCO (Nguyen et al., 2016) etc. have facilitated research in question answering task. These large-scale datasets have also been used to create the resources (Singh et al., 2019; Gupta et al., 2018a) and evaluate the performance (Asai et al., 2018; Gupta et al., 2018b; Gupta et al., 2019a) of the system in low-resource reading comprehension task. SQuAD dataset itself has been the de facto choice for most of the previous works in question genera- tion. However, 90% of the questions in SQuAD can be answered from a single sentence (Min et al., 2018), hence former QG systems trained on SQuAD are not capable of distilling and utilizing informa- tion from multiple sentences. Recently released multi-hop datasets such as QAngaroo (Welbl et al., 2018), ComplexWebQuestions (Talmor and Berant, 2018) and HotPotQA (Yang et al., 2018) are more suitable for building QG systems …

Content truncated to 500 words