@inproceedings{he-etal-2026-hw-tscs,
title = "{HW}-{TSC}{'}s Submission to the {WMT} 2026 General Machine Translation Track",
author = "He, Yu and
Lan, Xiaoqing and
Wei, Daimeng and
Guo, Jiaxin and
Luo, Yuanchang and
Shang, Hengchao and
Li, Zongyao and
Yang, Jinlong and
Wu, Zhanglin and
Huang, Boqi",
editor = "Haddow, Barry and
Kocmi, Tom and
Koehn, Philipp and
Monz, Christof",
booktitle = "Proceedings of the Eleventh Conference on Machine Translation",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.wmt-main.66/",
pages = "1258--1263",
ISBN = "979-8-89176-474-3",
abstract = "Machine translation based on large language models (LLMs) has achieved remarkable progress in recent years. However, relying on a single translation model and fixed translation instructions often limits translation quality, especially for multilingual and multi-domain scenarios. To address this issue, we propose HW-TSC-Agent, a multi-stage translation agent framework that integrates translation model fine-tuning, dual-model candidate generation, and instruction-aware translation refinement into a unified pipeline. First, we perform Supervised Fine-Tuning (SFT) and Contrastive Preference Optimization (CPO) on openPangu-Embedded-7B using high-quality bilingual data to enhance its multilingual translation capabilities. The fine-tuned openPangu-Embedded-7B and HY-MT2-7B are then used to generate candidate translations. Finally, source-language analysis, candidate translations, and task-specific instructions are incorporated into a dynamic prompt, which guides DeepSeek-V4-Flash to compare, refine, and fuse the candidate translations to produce the final output. Our system is submitted to the WMT2026 General Machine Translation Shared Task, covering 13 translation directions. On the WMT2025 English-to-Chinese development test set, enriching the dynamic prompt generally improves translation quality over an instruction-only DeepSeek-V4-Flash baseline."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="he-etal-2026-hw-tscs">
<titleInfo>
<title>HW-TSC’s Submission to the WMT 2026 General Machine Translation Track</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaoqing</namePart>
<namePart type="family">Lan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daimeng</namePart>
<namePart type="family">Wei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiaxin</namePart>
<namePart type="family">Guo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuanchang</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hengchao</namePart>
<namePart type="family">Shang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zongyao</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jinlong</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhanglin</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Boqi</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Eleventh 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">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-474-3</identifier>
</relatedItem>
<abstract>Machine translation based on large language models (LLMs) has achieved remarkable progress in recent years. However, relying on a single translation model and fixed translation instructions often limits translation quality, especially for multilingual and multi-domain scenarios. To address this issue, we propose HW-TSC-Agent, a multi-stage translation agent framework that integrates translation model fine-tuning, dual-model candidate generation, and instruction-aware translation refinement into a unified pipeline. First, we perform Supervised Fine-Tuning (SFT) and Contrastive Preference Optimization (CPO) on openPangu-Embedded-7B using high-quality bilingual data to enhance its multilingual translation capabilities. The fine-tuned openPangu-Embedded-7B and HY-MT2-7B are then used to generate candidate translations. Finally, source-language analysis, candidate translations, and task-specific instructions are incorporated into a dynamic prompt, which guides DeepSeek-V4-Flash to compare, refine, and fuse the candidate translations to produce the final output. Our system is submitted to the WMT2026 General Machine Translation Shared Task, covering 13 translation directions. On the WMT2025 English-to-Chinese development test set, enriching the dynamic prompt generally improves translation quality over an instruction-only DeepSeek-V4-Flash baseline.</abstract>
<identifier type="citekey">he-etal-2026-hw-tscs</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.66/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1258</start>
<end>1263</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T HW-TSC’s Submission to the WMT 2026 General Machine Translation Track
%A He, Yu
%A Lan, Xiaoqing
%A Wei, Daimeng
%A Guo, Jiaxin
%A Luo, Yuanchang
%A Shang, Hengchao
%A Li, Zongyao
%A Yang, Jinlong
%A Wu, Zhanglin
%A Huang, Boqi
%Y Haddow, Barry
%Y Kocmi, Tom
%Y Koehn, Philipp
%Y Monz, Christof
%S Proceedings of the Eleventh Conference on Machine Translation
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-474-3
%F he-etal-2026-hw-tscs
%X Machine translation based on large language models (LLMs) has achieved remarkable progress in recent years. However, relying on a single translation model and fixed translation instructions often limits translation quality, especially for multilingual and multi-domain scenarios. To address this issue, we propose HW-TSC-Agent, a multi-stage translation agent framework that integrates translation model fine-tuning, dual-model candidate generation, and instruction-aware translation refinement into a unified pipeline. First, we perform Supervised Fine-Tuning (SFT) and Contrastive Preference Optimization (CPO) on openPangu-Embedded-7B using high-quality bilingual data to enhance its multilingual translation capabilities. The fine-tuned openPangu-Embedded-7B and HY-MT2-7B are then used to generate candidate translations. Finally, source-language analysis, candidate translations, and task-specific instructions are incorporated into a dynamic prompt, which guides DeepSeek-V4-Flash to compare, refine, and fuse the candidate translations to produce the final output. Our system is submitted to the WMT2026 General Machine Translation Shared Task, covering 13 translation directions. On the WMT2025 English-to-Chinese development test set, enriching the dynamic prompt generally improves translation quality over an instruction-only DeepSeek-V4-Flash baseline.
%U https://aclanthology.org/2026.wmt-main.66/
%P 1258-1263
Markdown (Informal)
[HW-TSC’s Submission to the WMT 2026 General Machine Translation Track](https://aclanthology.org/2026.wmt-main.66/) (He et al., WMT 2026)
ACL
- Yu He, Xiaoqing Lan, Daimeng Wei, Jiaxin Guo, Yuanchang Luo, Hengchao Shang, Zongyao Li, Jinlong Yang, Zhanglin Wu, and Boqi Huang. 2026. HW-TSC’s Submission to the WMT 2026 General Machine Translation Track. In Proceedings of the Eleventh Conference on Machine Translation, pages 1258–1263, Budapest, Hungary. Association for Computational Linguistics.