@inproceedings{bar-etal-2024-robustness,
title = "Robustness of Fine-Tuned {LLM}s for Machine Translation with Varying Noise Levels: Insights for {A}sturian, {A}ragonese and Aranese",
author = {B{\"a}r, Martin and
Forcada Rodr{\'\i}guez, Elisa and
Garcia-Abadillo, Maria},
editor = "Haddow, Barry and
Kocmi, Tom and
Koehn, Philipp and
Monz, Christof",
booktitle = "Proceedings of the Ninth Conference on Machine Translation",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.wmt-1.89",
pages = "918--924",
abstract = "We present the LCT-LAP proposal for the shared task on Translation into Low-Resource Languages of Spain at WMT24 within the constrained submission category. Our work harnesses encoder-decoder models pretrained on higher-resource Iberian languages to facilitate MT model training for Asturian, Aranese and Aragonese. Furthermore, we explore the robustness of these models when fine-tuned on datasets with varying levels of alignment noise. We fine-tuned a Spanish-Galician model using Asturian data filtered by BLEU score thresholds of 5, 15, 30 and 60, identifying BLEU 15 as the most effective. This threshold was then applied to the Aranese and Aragonese datasets. Our findings indicate that filtering the corpora reduces computational costs and improves performance compared to using nearly raw data or data filtered with language identification. However, it still falls short of the performance achieved by the rule-based system Apertium in Aranese and Aragonese.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="bar-etal-2024-robustness">
<titleInfo>
<title>Robustness of Fine-Tuned LLMs for Machine Translation with Varying Noise Levels: Insights for Asturian, Aragonese and Aranese</title>
</titleInfo>
<name type="personal">
<namePart type="given">Martin</namePart>
<namePart type="family">Bär</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Elisa</namePart>
<namePart type="family">Forcada Rodríguez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Garcia-Abadillo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Ninth Conference on Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Barry</namePart>
<namePart type="family">Haddow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Kocmi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philipp</namePart>
<namePart type="family">Koehn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Christof</namePart>
<namePart type="family">Monz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Miami, Florida, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We present the LCT-LAP proposal for the shared task on Translation into Low-Resource Languages of Spain at WMT24 within the constrained submission category. Our work harnesses encoder-decoder models pretrained on higher-resource Iberian languages to facilitate MT model training for Asturian, Aranese and Aragonese. Furthermore, we explore the robustness of these models when fine-tuned on datasets with varying levels of alignment noise. We fine-tuned a Spanish-Galician model using Asturian data filtered by BLEU score thresholds of 5, 15, 30 and 60, identifying BLEU 15 as the most effective. This threshold was then applied to the Aranese and Aragonese datasets. Our findings indicate that filtering the corpora reduces computational costs and improves performance compared to using nearly raw data or data filtered with language identification. However, it still falls short of the performance achieved by the rule-based system Apertium in Aranese and Aragonese.</abstract>
<identifier type="citekey">bar-etal-2024-robustness</identifier>
<location>
<url>https://aclanthology.org/2024.wmt-1.89</url>
</location>
<part>
<date>2024-11</date>
<extent unit="page">
<start>918</start>
<end>924</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Robustness of Fine-Tuned LLMs for Machine Translation with Varying Noise Levels: Insights for Asturian, Aragonese and Aranese
%A Bär, Martin
%A Forcada Rodríguez, Elisa
%A Garcia-Abadillo, Maria
%Y Haddow, Barry
%Y Kocmi, Tom
%Y Koehn, Philipp
%Y Monz, Christof
%S Proceedings of the Ninth Conference on Machine Translation
%D 2024
%8 November
%I Association for Computational Linguistics
%C Miami, Florida, USA
%F bar-etal-2024-robustness
%X We present the LCT-LAP proposal for the shared task on Translation into Low-Resource Languages of Spain at WMT24 within the constrained submission category. Our work harnesses encoder-decoder models pretrained on higher-resource Iberian languages to facilitate MT model training for Asturian, Aranese and Aragonese. Furthermore, we explore the robustness of these models when fine-tuned on datasets with varying levels of alignment noise. We fine-tuned a Spanish-Galician model using Asturian data filtered by BLEU score thresholds of 5, 15, 30 and 60, identifying BLEU 15 as the most effective. This threshold was then applied to the Aranese and Aragonese datasets. Our findings indicate that filtering the corpora reduces computational costs and improves performance compared to using nearly raw data or data filtered with language identification. However, it still falls short of the performance achieved by the rule-based system Apertium in Aranese and Aragonese.
%U https://aclanthology.org/2024.wmt-1.89
%P 918-924
Markdown (Informal)
[Robustness of Fine-Tuned LLMs for Machine Translation with Varying Noise Levels: Insights for Asturian, Aragonese and Aranese](https://aclanthology.org/2024.wmt-1.89) (Bär et al., WMT 2024)
ACL