@inproceedings{yan-etal-2025-nvidia,
title = "Nvidia-Nemo{'}s {WMT} 2025 Metrics Shared Task Submission",
author = "Yan, Brian and
Ding, Shuoyang and
Wang, Kuang-Da and
Ouyang, Siqi and
Hrinchuk, Oleksii and
Lavrukhin, Vitaly and
Ginsburg, Boris",
editor = "Haddow, Barry and
Kocmi, Tom and
Koehn, Philipp and
Monz, Christof",
booktitle = "Proceedings of the Tenth Conference on Machine Translation",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.wmt-1.66/",
pages = "920--925",
ISBN = "979-8-89176-341-8",
abstract = "This paper describes Nvidia-Nemo{'}s WMT 2025 Metrics Shared Task submission. We investigated two strategies for extending Machine Translation (MT) evaluation to unsegmented documents: 1) first segmenting into sentences and then applying regression-based metrics and 2) directly utilizing the long-context capabilities of LLMs. The base comparison of the segmentation-based and LLM-based metrics on the WMT 2023-24 evaluation sets indicated that the former performs more robustly across language pairs.Thus we sought to improve the LLM-based approach by incorporating relative evaluation - this setting jointly evaluates all candidate translations at once and relative to each other, rather than evaluating each separately. Our experiments using the open-source Qwen3 LLM show that relative evaluation improves score correlations with human judgment, but only if the task is structured as a 2-stage evaluate-then-refine problem."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yan-etal-2025-nvidia">
<titleInfo>
<title>Nvidia-Nemo’s WMT 2025 Metrics Shared Task Submission</title>
</titleInfo>
<name type="personal">
<namePart type="given">Brian</namePart>
<namePart type="family">Yan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuoyang</namePart>
<namePart type="family">Ding</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kuang-Da</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Siqi</namePart>
<namePart type="family">Ouyang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Oleksii</namePart>
<namePart type="family">Hrinchuk</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vitaly</namePart>
<namePart type="family">Lavrukhin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Boris</namePart>
<namePart type="family">Ginsburg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Tenth 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">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-341-8</identifier>
</relatedItem>
<abstract>This paper describes Nvidia-Nemo’s WMT 2025 Metrics Shared Task submission. We investigated two strategies for extending Machine Translation (MT) evaluation to unsegmented documents: 1) first segmenting into sentences and then applying regression-based metrics and 2) directly utilizing the long-context capabilities of LLMs. The base comparison of the segmentation-based and LLM-based metrics on the WMT 2023-24 evaluation sets indicated that the former performs more robustly across language pairs.Thus we sought to improve the LLM-based approach by incorporating relative evaluation - this setting jointly evaluates all candidate translations at once and relative to each other, rather than evaluating each separately. Our experiments using the open-source Qwen3 LLM show that relative evaluation improves score correlations with human judgment, but only if the task is structured as a 2-stage evaluate-then-refine problem.</abstract>
<identifier type="citekey">yan-etal-2025-nvidia</identifier>
<location>
<url>https://aclanthology.org/2025.wmt-1.66/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>920</start>
<end>925</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Nvidia-Nemo’s WMT 2025 Metrics Shared Task Submission
%A Yan, Brian
%A Ding, Shuoyang
%A Wang, Kuang-Da
%A Ouyang, Siqi
%A Hrinchuk, Oleksii
%A Lavrukhin, Vitaly
%A Ginsburg, Boris
%Y Haddow, Barry
%Y Kocmi, Tom
%Y Koehn, Philipp
%Y Monz, Christof
%S Proceedings of the Tenth Conference on Machine Translation
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-341-8
%F yan-etal-2025-nvidia
%X This paper describes Nvidia-Nemo’s WMT 2025 Metrics Shared Task submission. We investigated two strategies for extending Machine Translation (MT) evaluation to unsegmented documents: 1) first segmenting into sentences and then applying regression-based metrics and 2) directly utilizing the long-context capabilities of LLMs. The base comparison of the segmentation-based and LLM-based metrics on the WMT 2023-24 evaluation sets indicated that the former performs more robustly across language pairs.Thus we sought to improve the LLM-based approach by incorporating relative evaluation - this setting jointly evaluates all candidate translations at once and relative to each other, rather than evaluating each separately. Our experiments using the open-source Qwen3 LLM show that relative evaluation improves score correlations with human judgment, but only if the task is structured as a 2-stage evaluate-then-refine problem.
%U https://aclanthology.org/2025.wmt-1.66/
%P 920-925
Markdown (Informal)
[Nvidia-Nemo’s WMT 2025 Metrics Shared Task Submission](https://aclanthology.org/2025.wmt-1.66/) (Yan et al., WMT 2025)
ACL
- Brian Yan, Shuoyang Ding, Kuang-Da Wang, Siqi Ouyang, Oleksii Hrinchuk, Vitaly Lavrukhin, and Boris Ginsburg. 2025. Nvidia-Nemo’s WMT 2025 Metrics Shared Task Submission. In Proceedings of the Tenth Conference on Machine Translation, pages 920–925, Suzhou, China. Association for Computational Linguistics.