Improving Question Generation with Multi-level Content Planning

TL;DR

This paper introduces MultiFactor, a multi-level content planning framework for question generation that improves the connection between key phrases and answers, leading to better performance on QG datasets.

Contribution

The paper proposes a novel multi-level content planning framework with phrase-enhanced transformers for improved question generation from extended contexts.

Findings

Outperforms strong baselines on two QG datasets

Uses a joint model for phrase selection and answer generation

Introduces answer-aware summaries to enhance question generation

Abstract

This paper addresses the problem of generating questions from a given context and an answer, specifically focusing on questions that require multi-hop reasoning across an extended context. Previous studies have suggested that key phrase selection is essential for question generation (QG), yet it is still challenging to connect such disjointed phrases into meaningful questions, particularly for long context. To mitigate this issue, we propose MultiFactor, a novel QG framework based on multi-level content planning. Specifically, MultiFactor includes two components: FA-model, which simultaneously selects key phrases and generates full answers, and Q-model which takes the generated full answer as an additional input to generate questions. Here, full answer generation is introduced to connect the short answer with the selected key phrases, thus forming an answer-aware summary to facilitate QG. Both FA-model and Q-model are formalized as simple-yet-effective Phrase-Enhanced Transformers, our joint model for phrase selection and text generation. Experimental results show that our method outperforms strong baselines on two popular QG datasets. Our code is available at https://github.com/zeaver/MultiFactor.