Saved in:
Bibliographic Details
Main Authors: Leite, Bernardo, Cardoso, Henrique Lopes
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.02800
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908397999226880
author Leite, Bernardo
Cardoso, Henrique Lopes
author_facet Leite, Bernardo
Cardoso, Henrique Lopes
contents Question Generation aims to automatically generate questions based on a given input provided as context. A controllable question generation scheme focuses on generating questions with specific attributes, allowing better control. In this study, we propose a few-shot prompting strategy for controlling the generation of question-answer pairs from children's narrative texts. We aim to control two attributes: the question's explicitness and underlying narrative elements. With empirical evaluation, we show the effectiveness of controlling the generation process by employing few-shot prompting side by side with a reference model. Our experiments highlight instances where the few-shot strategy surpasses the reference model, particularly in scenarios such as semantic closeness evaluation and the diversity and coherency of question-answer pairs. However, these improvements are not always statistically significant. The code is publicly available at github.com/bernardoleite/few-shot-prompting-qg-control.
format Preprint
id arxiv_https___arxiv_org_abs_2404_02800
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle On Few-Shot Prompting for Controllable Question-Answer Generation in Narrative Comprehension
Leite, Bernardo
Cardoso, Henrique Lopes
Computation and Language
Artificial Intelligence
Question Generation aims to automatically generate questions based on a given input provided as context. A controllable question generation scheme focuses on generating questions with specific attributes, allowing better control. In this study, we propose a few-shot prompting strategy for controlling the generation of question-answer pairs from children's narrative texts. We aim to control two attributes: the question's explicitness and underlying narrative elements. With empirical evaluation, we show the effectiveness of controlling the generation process by employing few-shot prompting side by side with a reference model. Our experiments highlight instances where the few-shot strategy surpasses the reference model, particularly in scenarios such as semantic closeness evaluation and the diversity and coherency of question-answer pairs. However, these improvements are not always statistically significant. The code is publicly available at github.com/bernardoleite/few-shot-prompting-qg-control.
title On Few-Shot Prompting for Controllable Question-Answer Generation in Narrative Comprehension
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2404.02800