Diversify Question Generation with Retrieval-Augmented Style Transfer
This work addresses the challenge of expression diversity in question generation systems, which is incremental by leveraging external templates for improved performance.
The paper tackles the problem of generating diverse question expressions from a given passage and answer by proposing RAST, a retrieval-augmented style transfer framework that uses external knowledge from diverse templates, resulting in outperforming previous baselines on diversity metrics while maintaining comparable consistency scores.
Given a textual passage and an answer, humans are able to ask questions with various expressions, but this ability is still challenging for most question generation (QG) systems. Existing solutions mainly focus on the internal knowledge within the given passage or the semantic word space for diverse content planning. These methods, however, have not considered the potential of external knowledge for expression diversity. To bridge this gap, we propose RAST, a framework for Retrieval-Augmented Style Transfer, where the objective is to utilize the style of diverse templates for question generation. For training RAST, we develop a novel Reinforcement Learning (RL) based approach that maximizes a weighted combination of diversity reward and consistency reward. Here, the consistency reward is computed by a Question-Answering (QA) model, whereas the diversity reward measures how much the final output mimics the retrieved template. Experimental results show that our method outperforms previous diversity-driven baselines on diversity while being comparable in terms of consistency scores. Our code is available at https://github.com/gouqi666/RAST.