<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sun-etal-2026-cuhksz">
    <titleInfo>
        <title>The CUHKSZ System for the IWSLT 2026 Low-Resource Speech-to-Text Task</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">ruiyan</namePart>
        <namePart type="family">SUN</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Qingming</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Satoshi</namePart>
        <namePart type="family">Nakamura</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 23rd International Conference on Spoken Language Translation (IWSLT 2026)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Elizabeth</namePart>
            <namePart type="family">Salesky</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Antonios</namePart>
            <namePart type="family">Anastasopoulos</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Matteo</namePart>
            <namePart type="family">Negri</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Marcello</namePart>
            <namePart type="family">Federico</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">San Diego, USA (in-person and online)</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-411-8</identifier>
    </relatedItem>
    <abstract>This paper describes the CUHKSZ system for the IWSLT 2026 Low-Resource Speech-to-Text task. We propose Gradient-Driven Parameter Sharing (GDPS), a framework that analyzes inter-language gradient behaviors to automatically determine optimal language groupings and shared-private parameter ratios. Built upon SeamlessM4T-Medium, GDPS reduces negative transfer by specializing Layer 11 FFN2 while maintaining shared encoder representations across languages. Additionally, we incorporate curriculum distillation with progressive pseudo-label mixing and test-time reranking combining prior-BLEU weighting and self-consistency scoring. Evaluation on eight low-resource languages (bem, ckb, gle, hau, ibo, yor, aeb, est) demonstrates strongest gains on bem (+2.07 BLEU), hau (+1.50), and ibo (+0.38) compared to unified fine-tuning, while ckb and yor benefit more from prior-based reranking at inference.</abstract>
    <identifier type="citekey">sun-etal-2026-cuhksz</identifier>
    <location>
        <url>https://aclanthology.org/2026.iwslt-1.33/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>296</start>
            <end>304</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T The CUHKSZ System for the IWSLT 2026 Low-Resource Speech-to-Text Task
%A SUN, ruiyan
%A Li, Qingming
%A Nakamura, Satoshi
%Y Salesky, Elizabeth
%Y Anastasopoulos, Antonios
%Y Negri, Matteo
%Y Federico, Marcello
%S Proceedings of the 23rd International Conference on Spoken Language Translation (IWSLT 2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, USA (in-person and online)
%@ 979-8-89176-411-8
%F sun-etal-2026-cuhksz
%X This paper describes the CUHKSZ system for the IWSLT 2026 Low-Resource Speech-to-Text task. We propose Gradient-Driven Parameter Sharing (GDPS), a framework that analyzes inter-language gradient behaviors to automatically determine optimal language groupings and shared-private parameter ratios. Built upon SeamlessM4T-Medium, GDPS reduces negative transfer by specializing Layer 11 FFN2 while maintaining shared encoder representations across languages. Additionally, we incorporate curriculum distillation with progressive pseudo-label mixing and test-time reranking combining prior-BLEU weighting and self-consistency scoring. Evaluation on eight low-resource languages (bem, ckb, gle, hau, ibo, yor, aeb, est) demonstrates strongest gains on bem (+2.07 BLEU), hau (+1.50), and ibo (+0.38) compared to unified fine-tuning, while ckb and yor benefit more from prior-based reranking at inference.
%U https://aclanthology.org/2026.iwslt-1.33/
%P 296-304

Markdown (Informal)

[The CUHKSZ System for the IWSLT 2026 Low-Resource Speech-to-Text Task](https://aclanthology.org/2026.iwslt-1.33/) (SUN et al., IWSLT 2026)

• The CUHKSZ System for the IWSLT 2026 Low-Resource Speech-to-Text Task (SUN et al., IWSLT 2026)

ACL

• ruiyan SUN, Qingming Li, and Satoshi Nakamura. 2026. The CUHKSZ System for the IWSLT 2026 Low-Resource Speech-to-Text Task. In Proceedings of the 23rd International Conference on Spoken Language Translation (IWSLT 2026), pages 296–304, San Diego, USA (in-person and online). Association for Computational Linguistics.