Leveraging Low-resource Parallel Data for Text Style Transfer

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F23%3A10476023" target="_blank" >RIV/00216208:11320/23:10476023 - isvavai.cz</a>

Výsledek na webu

<a href="https://aclanthology.org/2023.inlg-main.27" target="_blank" >https://aclanthology.org/2023.inlg-main.27</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Leveraging Low-resource Parallel Data for Text Style Transfer

Popis výsledku v původním jazyce

Text style transfer (TST) involves transforming a text into a desired style while approximately preserving its content. The biggest challenge in TST in the general lack of parallel data. Many existing approaches rely on complex models using substantial non-parallel data, with mixed results. In this paper, we leverage a pretrained BART language model with minimal parallel data and incorporate low-resource methods such as hyperparameter tuning, data augmentation, and self-training, which have not been explored in TST. We further include novel style-based rewards in the training loss. Through extensive experiments in sentiment transfer, a sub-task of TST, we demonstrate that our simple yet effective approaches achieve well-balanced results, surpassing non-parallel approaches and highlighting the usefulness of parallel data even in small amounts.

Název v anglickém jazyce

Leveraging Low-resource Parallel Data for Text Style Transfer

Popis výsledku anglicky

Text style transfer (TST) involves transforming a text into a desired style while approximately preserving its content. The biggest challenge in TST in the general lack of parallel data. Many existing approaches rely on complex models using substantial non-parallel data, with mixed results. In this paper, we leverage a pretrained BART language model with minimal parallel data and incorporate low-resource methods such as hyperparameter tuning, data augmentation, and self-training, which have not been explored in TST. We further include novel style-based rewards in the training loss. Through extensive experiments in sentiment transfer, a sub-task of TST, we demonstrate that our simple yet effective approaches achieve well-balanced results, surpassing non-parallel approaches and highlighting the usefulness of parallel data even in small amounts.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the 16th International Natural Language Generation Conference

ISBN

979-8-89176-001-1

ISSN

—

e-ISSN

—

Počet stran výsledku

8

Strana od-do

388-395

Název nakladatele

Association for Computational Linguistics

Místo vydání

Stroudsburg, PA, USA

Místo konání akce

Praha, Czechia

Datum konání akce

13. 9. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—