Generating Self-Contained and Summary-Centric Question Answer Pairs via Differentiable Reward Imitation Learning
This work addresses a domain-specific problem for conversational news recommendation systems, offering an incremental improvement by mitigating exposure bias with a differentiable reward function.
The paper tackles generating question-answer pairs from news articles, focusing on self-contained, summary-centric questions and length-constrained answers, and achieves high answer accuracy as shown by automatic metrics and human evaluation.
Motivated by suggested question generation in conversational news recommendation systems, we propose a model for generating question-answer pairs (QA pairs) with self-contained, summary-centric questions and length-constrained, article-summarizing answers. We begin by collecting a new dataset of news articles with questions as titles and pairing them with summaries of varying length. This dataset is used to learn a QA pair generation model producing summaries as answers that balance brevity with sufficiency jointly with their corresponding questions. We then reinforce the QA pair generation process with a differentiable reward function to mitigate exposure bias, a common problem in natural language generation. Both automatic metrics and human evaluation demonstrate these QA pairs successfully capture the central gists of the articles and achieve high answer accuracy.