<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="luo-etal-2026-rasst">
    <titleInfo>
        <title>RASST: Retrieval-Augmented Simultaneous Speech Translation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Jiaxuan</namePart>
        <namePart type="family">Luo</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">Jiaxing</namePart>
        <namePart type="family">Xu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Lei</namePart>
        <namePart type="family">Li</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 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</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">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Simultaneous speech translation produces target text incrementally from partial speech input. Recent speech large language models have markedly improved SST quality but still struggle with rare and domain-specific terminology. Retrieval augmentation has helped in automatic speech recognition and neural machine translation, but extending it to SST is non-trivial: retrieval must be fast and accurate under partial speech, and the model must decide whether and when to apply retrieved terms during incremental generation. We propose Retrieval-Augmented Simultaneous Speech Translation (RASST), which addresses both challenges. For accurate cross-modal retrieval under partial input, RASST trains a lightweight speech-text retriever that produces chunkwise terminology hints for the Speech LLM via multi-scale retrieval. To use these hints correctly, we synthesize training data that teaches the Speech LLM to decide whether and when to apply each retrieved term. Experiments on the ACL 60/60 development set and the ESO oncology test set show that RASST improves terminology accuracy over InfiniSST by 12.3-31.22% and BLEU by up to 3 points across all tested languages and latency settings. Furthermore, the retriever incurs negligible computational overhead and remains robust as glossary size increases to 10K entries.</abstract>
    <identifier type="citekey">luo-etal-2026-rasst</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.1680/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>38241</start>
            <end>38260</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T RASST: Retrieval-Augmented Simultaneous Speech Translation
%A Luo, Jiaxuan
%A Ouyang, Siqi
%A Xu, Jiaxing
%A Li, Lei
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F luo-etal-2026-rasst
%X Simultaneous speech translation produces target text incrementally from partial speech input. Recent speech large language models have markedly improved SST quality but still struggle with rare and domain-specific terminology. Retrieval augmentation has helped in automatic speech recognition and neural machine translation, but extending it to SST is non-trivial: retrieval must be fast and accurate under partial speech, and the model must decide whether and when to apply retrieved terms during incremental generation. We propose Retrieval-Augmented Simultaneous Speech Translation (RASST), which addresses both challenges. For accurate cross-modal retrieval under partial input, RASST trains a lightweight speech-text retriever that produces chunkwise terminology hints for the Speech LLM via multi-scale retrieval. To use these hints correctly, we synthesize training data that teaches the Speech LLM to decide whether and when to apply each retrieved term. Experiments on the ACL 60/60 development set and the ESO oncology test set show that RASST improves terminology accuracy over InfiniSST by 12.3-31.22% and BLEU by up to 3 points across all tested languages and latency settings. Furthermore, the retriever incurs negligible computational overhead and remains robust as glossary size increases to 10K entries.
%U https://aclanthology.org/2026.emnlp-main.1680/
%P 38241-38260

Markdown (Informal)

[RASST: Retrieval-Augmented Simultaneous Speech Translation](https://aclanthology.org/2026.emnlp-main.1680/) (Luo et al., EMNLP 2026)

• RASST: Retrieval-Augmented Simultaneous Speech Translation (Luo et al., EMNLP 2026)

ACL

• Jiaxuan Luo, Siqi Ouyang, Jiaxing Xu, and Lei Li. 2026. RASST: Retrieval-Augmented Simultaneous Speech Translation. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 38241–38260, Budapest, Hungary. Association for Computational Linguistics.