Punish the Pun-ish: Enhancing Text-to-Pun Generation with Synthetic Data from Supervised Fine-tuned Models

Tomohito Minami, Ryohei Orihara, Yasuyuki Tahara, Akihiko Ohsuga, Yuichi Sei

2025

Abstract

Puns are clever wordplays that exploit sound similarities while contrasting different meanings. Such complex puns remain challenging to create, even with today’s advanced large language models. This study focuses on generating Japanese juxtaposed puns while preserving the original meaning of input sentences. We propose a novel approach, applying Direct Preference Optimization (DPO) after supervised fine-tuning (SFT) of a pre-trained language model, utilizing synthetic data generated from the SFT model to refine pun generation. Experimental results indicate that our approach yields a marked improvement, evaluated using neural network-based and rule-based metrics designed to measure pun-ness, with a 2.3-point increase and a 7.9-point increase, respectively, over the baseline SFT model. These findings suggest that integrating SFT with DPO enhances the model’s ability to capture phonetic nuances essential for generating juxtaposed puns.

Paper Citation

in Harvard Style

Minami T., Orihara R., Tahara Y., Ohsuga A. and Sei Y. (2025). Punish the Pun-ish: Enhancing Text-to-Pun Generation with Synthetic Data from Supervised Fine-tuned Models. In Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART; ISBN 978-989-758-737-5, SciTePress, pages 1093-1100. DOI: 10.5220/0013262900003890

in Bibtex Style

@conference{icaart25,
author={Tomohito Minami and Ryohei Orihara and Yasuyuki Tahara and Akihiko Ohsuga and Yuichi Sei},
title={Punish the Pun-ish: Enhancing Text-to-Pun Generation with Synthetic Data from Supervised Fine-tuned Models},
booktitle={Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART},
year={2025},
pages={1093-1100},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013262900003890},
isbn={978-989-758-737-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART
TI - Punish the Pun-ish: Enhancing Text-to-Pun Generation with Synthetic Data from Supervised Fine-tuned Models
SN - 978-989-758-737-5
AU - Minami T.
AU - Orihara R.
AU - Tahara Y.
AU - Ohsuga A.
AU - Sei Y.
PY - 2025
SP - 1093
EP - 1100
DO - 10.5220/0013262900003890
PB - SciTePress