Improving Question Answering with Generation of NQ-like Questions

TL;DR

This paper presents an algorithm that automatically generates NQ-like questions from longer trivia questions, enhancing data availability and improving QA system performance, especially in low-resource settings.

Contribution

The paper introduces a novel method to automatically convert trivia questions into NQ-like questions, increasing data for QA systems while maintaining quality.

Findings

Generated data improves QA performance in low-resource settings

Neural classifier effectively filters ill-formed questions

Method enhances scalability of training data for QA systems

Abstract

Question Answering (QA) systems require a large amount of annotated data which is costly and time-consuming to gather. Converting datasets of existing QA benchmarks are challenging due to different formats and complexities. To address these issues, we propose an algorithm to automatically generate shorter questions resembling day-to-day human communication in the Natural Questions (NQ) dataset from longer trivia questions in Quizbowl (QB) dataset by leveraging conversion in style among the datasets. This provides an automated way to generate more data for our QA systems. To ensure quality as well as quantity of data, we detect and remove ill-formed questions using a neural classifier. We demonstrate that in a low resource setting, using the generated data improves the QA performance over the baseline system on both NQ and QB data. Our algorithm improves the scalability of training data while maintaining quality of data for QA systems.