City Research Online

Syllable Neural Language Models for English Poem Generation

Lewis, D., Zugarini, A. & Alonso, E. ORCID: 0000-0002-3306-695X (2021). Syllable Neural Language Models for English Poem Generation. In: Proceedings of the Twelfth International Conference on Computational Creativity. 12th International Conference on Computational Creativity (ICCC'21), 14-18 Sep 2021, Mexico City, Mexico.

Abstract

Automatic Poem Generation is an ambitious Natural Language Generation (NLG) problem. Indeed, models have to replicate the precise structure of poems, rhymes, meters, while producing creative and emotional verses. Furthermore, the lack of abundant poetic corpora, especially for ancient poetry, is a serious limitation for the development of strong poem generators. In this paper, we propose a syllable neural language model to the case of English language, focusing on the generation of verses with the style of a target author: William Wordsworth. To alleviate the problem of limited available data, we exploit transfer learning. Furthermore, we bias the generation of verses according to a combination of different scoring functions based on meter, style and gram-mar in order to select lines more compliant with the author’s characteristics. The results of both quantitative and human evaluations shows the effectiveness of our approach. In particular, human judges struggle to recognize real verses from the generated ones.

Publication Type: Conference or Workshop Item (Paper)
Subjects: P Language and Literature > PE English
Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments: School of Science & Technology
[thumbnail of ICCC_2021.pdf]
Preview
Text - Accepted Version
Available under License Creative Commons: Attribution International Public License 4.0.

Download (286kB) | Preview

Export

Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login