HeyBusan at WMT26: A System for the Sorbian Track

Mikhail Khasykov

HeyBusan at WMT26: A System for the Sorbian Track

Correct Metadata for

Use this form to create a GitHub issue with structured data describing the correction. You will need a GitHub account. Once you create that issue, the correction will be reviewed by a staff member.

⚠️ Mobile Users: Submitting this form to create a new issue will only work with github.com, not the GitHub Mobile app.

Important: The Anthology treat PDFs as authoritative. Please use this form only to correct data that is out of line with the PDF. See our corrections guidelines if you need to change the PDF.

Title Adjust the title. Retain tags such as <fixed-case>.

Authors Adjust author names and order to match the PDF.

Abstract Correct abstract if needed. Retain XML formatting tags such as <tex-math>. You may use ... for bold, ... for italic, ... for underline, <sc>...</sc> for small-caps, <tt>...<tt> for typewriter text, <url>...</url> for URLs, <a href=...> for hyperlinks, and <par/> for paragraph breaks.

Verification against PDF Ensure that the new title/authors match the snapshot below. (If there is no snapshot or it is too small, consult the PDF.)

Authors concatenated from the text boxes above:

ALL author names match the snapshot above—including middle initials, hyphens, and accents.

Abstract

We describe our submission (team HeyBusan) to the Sorbian track of the WMT26 Multitask LLMs with Limited Resources shared task. A single Qwen3.5-2B model handles machine translation, question answering, spell checking, grammar checking, and mathematical reasoning in Upper Sorbian, Lower Sorbian, and German. We fine-tuned all model parameters on public data (the official task data plus external corpora) for approximately 55 GPU-hours on one 24 GB consumer GPU. Task-specific decoding then improved the results further without changing the model’s weights. Team HeyBusan placed first in machine translation, spell checking, and grammar checking, and tied with team LT3 for the overall track win.

Anthology ID:: 2026.wmt-main.210
Volume:: Proceedings of the Eleventh Conference on Machine Translation
Month:: October
Year:: 2026
Address:: Budapest, Hungary
Editors:: Barry Haddow, Tom Kocmi, Philipp Koehn, Christof Monz
Venues:: WMT | WS
SIG:: SIGMT
Publisher:: Association for Computational Linguistics
Note:
Pages:: 2662–2667
Language:
URL:: https://aclanthology.org/2026.wmt-main.210/
DOI:
Bibkey:
Cite (ACL):: Mikhail Khasykov. 2026. HeyBusan at WMT26: A System for the Sorbian Track. In Proceedings of the Eleventh Conference on Machine Translation, pages 2662–2667, Budapest, Hungary. Association for Computational Linguistics.
Cite (Informal):: HeyBusan at WMT26: A System for the Sorbian Track (Khasykov, WMT 2026)
Copy Citation:
PDF:: https://aclanthology.org/2026.wmt-main.210.pdf

PDF Cite Search Fix data

Export citation

BibTeX
MODS XML
Endnote
Preformatted

@inproceedings{khasykov-2026-heybusan,
    title = "{H}ey{B}usan at {WMT}26: A System for the Sorbian Track",
    author = "Khasykov, Mikhail",
    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.210/",
    pages = "2662--2667",
    ISBN = "979-8-89176-474-3",
    abstract = "We describe our submission (team HeyBusan) to the Sorbian track of the WMT26 Multitask LLMs with Limited Resources shared task. A single Qwen3.5-2B model handles machine translation, question answering, spell checking, grammar checking, and mathematical reasoning in Upper Sorbian, Lower Sorbian, and German. We fine-tuned all model parameters on public data (the official task data plus external corpora) for approximately 55 GPU-hours on one 24 GB consumer GPU. Task-specific decoding then improved the results further without changing the model{'}s weights. Team HeyBusan placed first in machine translation, spell checking, and grammar checking, and tied with team LT3 for the overall track win."
}

Download as File

<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="khasykov-2026-heybusan">
    <titleInfo>
        <title>HeyBusan at WMT26: A System for the Sorbian Track</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Mikhail</namePart>
        <namePart type="family">Khasykov</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>We describe our submission (team HeyBusan) to the Sorbian track of the WMT26 Multitask LLMs with Limited Resources shared task. A single Qwen3.5-2B model handles machine translation, question answering, spell checking, grammar checking, and mathematical reasoning in Upper Sorbian, Lower Sorbian, and German. We fine-tuned all model parameters on public data (the official task data plus external corpora) for approximately 55 GPU-hours on one 24 GB consumer GPU. Task-specific decoding then improved the results further without changing the model’s weights. Team HeyBusan placed first in machine translation, spell checking, and grammar checking, and tied with team LT3 for the overall track win.</abstract>
    <identifier type="citekey">khasykov-2026-heybusan</identifier>
    <location>
        <url>https://aclanthology.org/2026.wmt-main.210/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>2662</start>
            <end>2667</end>
        </extent>
    </part>
</mods>
</modsCollection>

Download as File

%0 Conference Proceedings
%T HeyBusan at WMT26: A System for the Sorbian Track
%A Khasykov, Mikhail
%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 khasykov-2026-heybusan
%X We describe our submission (team HeyBusan) to the Sorbian track of the WMT26 Multitask LLMs with Limited Resources shared task. A single Qwen3.5-2B model handles machine translation, question answering, spell checking, grammar checking, and mathematical reasoning in Upper Sorbian, Lower Sorbian, and German. We fine-tuned all model parameters on public data (the official task data plus external corpora) for approximately 55 GPU-hours on one 24 GB consumer GPU. Task-specific decoding then improved the results further without changing the model’s weights. Team HeyBusan placed first in machine translation, spell checking, and grammar checking, and tied with team LT3 for the overall track win.
%U https://aclanthology.org/2026.wmt-main.210/
%P 2662-2667

Download as File

Markdown (Informal)

[HeyBusan at WMT26: A System for the Sorbian Track](https://aclanthology.org/2026.wmt-main.210/) (Khasykov, WMT 2026)

HeyBusan at WMT26: A System for the Sorbian Track (Khasykov, WMT 2026)

ACL

Mikhail Khasykov. 2026. HeyBusan at WMT26: A System for the Sorbian Track. In Proceedings of the Eleventh Conference on Machine Translation, pages 2662–2667, Budapest, Hungary. Association for Computational Linguistics.