Automated poetry scoring using BERT with multi-scale poetry representation
Automated poetry scoring using BERT with multi-scale poetry representation
Automated poetry scoring is an emerging task in automated text scoring, which is receiving increasing attention in AI for education. Poetry is distinct from other text in its complexity and specialty in language feature moreover, poems are usually rated from multiple criteria besides the overall impression. However, few existing methods to the best of our knowledge have considered a tailored text representation model for encoding poetry. Moreover, the lack of large poetry corpus and extensive labelled data is another major constraint to construct an effective poetry scoring model. To address such problems, we proposed BERT-based models with multi-scale poetry representation. In addition, we employ multiple losses and R-Drop strategy to align the distribution of manual and model scoring and mitigate the tendency of consistent score in poems. Experiment results demonstrate that our model with multi-scale poetry representation stands out when comparing with single-scale representation model.
automated poetry scoring, multi-scale text representation, pre-trained language model
250-261
Gao, Mingzhi
6d5418dd-4e15-4676-a662-85c1c7fd31b9
Ahipasaoglu, Selin
d69f1b80-5c05-4d50-82df-c13b87b02687
Schuster, Kristin
ad891fc3-0a92-49ad-8f9e-b9d8ef573ddf
4 September 2023
Gao, Mingzhi
6d5418dd-4e15-4676-a662-85c1c7fd31b9
Ahipasaoglu, Selin
d69f1b80-5c05-4d50-82df-c13b87b02687
Schuster, Kristin
ad891fc3-0a92-49ad-8f9e-b9d8ef573ddf
Gao, Mingzhi, Ahipasaoglu, Selin and Schuster, Kristin
(2023)
Automated poetry scoring using BERT with multi-scale poetry representation.
International Journal of Intelligent Systems Technologies and Applications, 21 (3), .
(doi:10.1504/IJISTA.2023.133694).
Abstract
Automated poetry scoring is an emerging task in automated text scoring, which is receiving increasing attention in AI for education. Poetry is distinct from other text in its complexity and specialty in language feature moreover, poems are usually rated from multiple criteria besides the overall impression. However, few existing methods to the best of our knowledge have considered a tailored text representation model for encoding poetry. Moreover, the lack of large poetry corpus and extensive labelled data is another major constraint to construct an effective poetry scoring model. To address such problems, we proposed BERT-based models with multi-scale poetry representation. In addition, we employ multiple losses and R-Drop strategy to align the distribution of manual and model scoring and mitigate the tendency of consistent score in poems. Experiment results demonstrate that our model with multi-scale poetry representation stands out when comparing with single-scale representation model.
Text
poetry scoring
- Accepted Manuscript
Restricted to Repository staff only until 4 September 2024.
Request a copy
More information
e-pub ahead of print date: 4 September 2023
Published date: 4 September 2023
Additional Information:
Funding Information: we thank the Faculty of Social Sciences at University of Southampton for providing funding that was spent for the NGO organising poetry writing workshop, collecting and scoring poems.
Keywords:
automated poetry scoring, multi-scale text representation, pre-trained language model
Identifiers
Local EPrints ID: 485423
URI: http://eprints.soton.ac.uk/id/eprint/485423
ISSN: 1740-8865
PURE UUID: b1bfd0ee-ca97-42d2-91a7-2c58d43406e2
Catalogue record
Date deposited: 06 Dec 2023 17:38
Last modified: 18 Mar 2024 03:58
Export record
Altmetrics
Contributors
Author:
Mingzhi Gao
Author:
Kristin Schuster
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
