Evaluating Sentence Alignment Methods in a Low-Resource Setting: An English-YorùBá Study Case

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216224%3A14330%2F23%3A00138715" target="_blank" >RIV/00216224:14330/23:00138715 - isvavai.cz</a>

Výsledek na webu

<a href="https://aclanthology.org/2023.loresmt-1.10.pdf" target="_blank" >https://aclanthology.org/2023.loresmt-1.10.pdf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.18653/v1/2023.loresmt-1.10" target="_blank" >10.18653/v1/2023.loresmt-1.10</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Evaluating Sentence Alignment Methods in a Low-Resource Setting: An English-YorùBá Study Case

Popis výsledku v původním jazyce

Parallel corpora are still crucial to train effective Machine Translation systems. This is even more true for low-resource language pairs, for which Neural Machine Translation has been shown to be less robust to domain mismatch and noise. Due to time and resource constraints, parallel corpora are mostly created with sentence alignment methods which automatically infer alignments. Recent work focused on state-of-the-art pre-trained sentence embeddings-based methods which are available only for a tiny fraction of the world’s languages. In this paper, we evaluate the performance of four widely used algorithms on the low-resource English-Yorùbá language pair against a multidomain benchmark parallel corpus on two experiments involving 1-to-1 alignments with and without reordering. We find that, at least for this language pair, earlier and simpler methods are more suited to the task, all the while not requiring additional data or resources. We also report that the methods we evaluated perform differently across distinct domains, thus indicating that some approach may be better for a specific domain or textual structure.

Název v anglickém jazyce

Evaluating Sentence Alignment Methods in a Low-Resource Setting: An English-YorùBá Study Case

Popis výsledku anglicky

Parallel corpora are still crucial to train effective Machine Translation systems. This is even more true for low-resource language pairs, for which Neural Machine Translation has been shown to be less robust to domain mismatch and noise. Due to time and resource constraints, parallel corpora are mostly created with sentence alignment methods which automatically infer alignments. Recent work focused on state-of-the-art pre-trained sentence embeddings-based methods which are available only for a tiny fraction of the world’s languages. In this paper, we evaluate the performance of four widely used algorithms on the low-resource English-Yorùbá language pair against a multidomain benchmark parallel corpus on two experiments involving 1-to-1 alignments with and without reordering. We find that, at least for this language pair, earlier and simpler methods are more suited to the task, all the while not requiring additional data or resources. We also report that the methods we evaluated perform differently across distinct domains, thus indicating that some approach may be better for a specific domain or textual structure.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10200 - Computer and information sciences

Projekt

<a href="/cs/project/LM2023062" target="_blank" >LM2023062: Digitální výzkumná infrastruktura pro jazykové technologie, umění a humanitní vědy</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

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 Sixth Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT 2023)

ISBN

9781959429555

ISSN

—

e-ISSN

—

Počet stran výsledku

7

Strana od-do

123-129

Název nakladatele

Association for Computational Linguistics

Místo vydání

Stroudsburg, PA 18360

Místo konání akce

Dubrovnik

Datum konání akce

6. 5. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—