Exploring Question-Specific Rewards for Generating Deep Questions

TL;DR

This paper proposes using reinforcement learning with question-specific rewards to improve the quality of generated questions, focusing on fluency, relevance, and answerability, and analyzes their impact through experiments.

Contribution

It introduces a reinforcement learning framework with tailored rewards for question generation, highlighting the importance of reward relevance to human judgment.

Findings

Relevance reward improves question quality according to human judgment.

Answerability reward can bias the model, reducing question quality.

Optimizing for relevant rewards enhances automatic evaluation metrics.

Abstract

Recent question generation (QG) approaches often utilize the sequence-to-sequence framework (Seq2Seq) to optimize the log-likelihood of ground-truth questions using teacher forcing. However, this training objective is inconsistent with actual question quality, which is often reflected by certain global properties such as whether the question can be answered by the document. As such, we directly optimize for QG-specific objectives via reinforcement learning to improve question quality. We design three different rewards that target to improve the fluency, relevance, and answerability of generated questions. We conduct both automatic and human evaluations in addition to a thorough analysis to explore the effect of each QG-specific reward. We find that optimizing question-specific rewards generally leads to better performance in automatic evaluation metrics. However, only the rewards that correlate well with human judgement (e.g., relevance) lead to real improvement in question quality. Optimizing for the others, especially answerability, introduces incorrect bias to the model, resulting in poor question quality. Our code is publicly available at https://github.com/YuxiXie/RL-for-Question-Generation.