@inproceedings{iranzo-sanchez-etal-2025-mllp,
title = "{MLLP}-{VRAIN} {UPV} system for the {IWSLT} 2025 Simultaneous Speech Translation Translation task",
author = "Iranzo-S{\'a}nchez, Jorge and
Iranzo-Sanchez, Javier and
Gim{\'e}nez Pastor, Adri{\`a} and
Civera Saiz, Jorge and
Juan, Alfons",
editor = "Salesky, Elizabeth and
Federico, Marcello and
Anastasopoulos, Antonis",
booktitle = "Proceedings of the 22nd International Conference on Spoken Language Translation (IWSLT 2025)",
month = jul,
year = "2025",
address = "Vienna, Austria (in-person and online)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.iwslt-1.35/",
doi = "10.18653/v1/2025.iwslt-1.35",
pages = "340--346",
ISBN = "979-8-89176-272-5",
abstract = "This work describes the participation of the MLLP-VRAIN research group in the shared task of the IWSLT 2025 Simultaneous Speech Translation track. Our submission addresses the unique challenges of real-time translation of long-form speech by developing a modular cascade system that adapts strong pre-trained models to streaming scenarios. We combine Whisper Large-V3-Turbo for ASR with the multilingual NLLB-3.3B model for MT, implementing lightweight adaptation techniques rather than training new end-to-end models from scratch. Our approach employs document-level adaptation with prefix training to enhance the MT model{'}s ability to handle incomplete inputs, while incorporating adaptive emission policies including a wait-k strategy and RALCP for managing the translation stream. Specialized buffer management techniques and segmentation strategies ensure coherent translations across long audio sequences. Experimental results on the ACL60/60 dataset demonstrate that our system achieves a favorable balance between translation quality and latency, with a BLEU score of 31.96 and non-computational-aware StreamLAAL latency of 2.94 seconds. Our final model achieves a preliminary score on the official test set (IWSLT25Instruct) of 29.8 BLEU. Our work demonstrates that carefully adapted pre-trained components can create effective simultaneous translation systems for long-form content without requiring extensive in-domain parallel data or specialized end-to-end training."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="iranzo-sanchez-etal-2025-mllp">
<titleInfo>
<title>MLLP-VRAIN UPV system for the IWSLT 2025 Simultaneous Speech Translation Translation task</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jorge</namePart>
<namePart type="family">Iranzo-Sánchez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Javier</namePart>
<namePart type="family">Iranzo-Sanchez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Adrià</namePart>
<namePart type="family">Giménez Pastor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jorge</namePart>
<namePart type="family">Civera Saiz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alfons</namePart>
<namePart type="family">Juan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 22nd International Conference on Spoken Language Translation (IWSLT 2025)</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">Marcello</namePart>
<namePart type="family">Federico</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonis</namePart>
<namePart type="family">Anastasopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria (in-person and online)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-272-5</identifier>
</relatedItem>
<abstract>This work describes the participation of the MLLP-VRAIN research group in the shared task of the IWSLT 2025 Simultaneous Speech Translation track. Our submission addresses the unique challenges of real-time translation of long-form speech by developing a modular cascade system that adapts strong pre-trained models to streaming scenarios. We combine Whisper Large-V3-Turbo for ASR with the multilingual NLLB-3.3B model for MT, implementing lightweight adaptation techniques rather than training new end-to-end models from scratch. Our approach employs document-level adaptation with prefix training to enhance the MT model’s ability to handle incomplete inputs, while incorporating adaptive emission policies including a wait-k strategy and RALCP for managing the translation stream. Specialized buffer management techniques and segmentation strategies ensure coherent translations across long audio sequences. Experimental results on the ACL60/60 dataset demonstrate that our system achieves a favorable balance between translation quality and latency, with a BLEU score of 31.96 and non-computational-aware StreamLAAL latency of 2.94 seconds. Our final model achieves a preliminary score on the official test set (IWSLT25Instruct) of 29.8 BLEU. Our work demonstrates that carefully adapted pre-trained components can create effective simultaneous translation systems for long-form content without requiring extensive in-domain parallel data or specialized end-to-end training.</abstract>
<identifier type="citekey">iranzo-sanchez-etal-2025-mllp</identifier>
<identifier type="doi">10.18653/v1/2025.iwslt-1.35</identifier>
<location>
<url>https://aclanthology.org/2025.iwslt-1.35/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>340</start>
<end>346</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MLLP-VRAIN UPV system for the IWSLT 2025 Simultaneous Speech Translation Translation task
%A Iranzo-Sánchez, Jorge
%A Iranzo-Sanchez, Javier
%A Giménez Pastor, Adrià
%A Civera Saiz, Jorge
%A Juan, Alfons
%Y Salesky, Elizabeth
%Y Federico, Marcello
%Y Anastasopoulos, Antonis
%S Proceedings of the 22nd International Conference on Spoken Language Translation (IWSLT 2025)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria (in-person and online)
%@ 979-8-89176-272-5
%F iranzo-sanchez-etal-2025-mllp
%X This work describes the participation of the MLLP-VRAIN research group in the shared task of the IWSLT 2025 Simultaneous Speech Translation track. Our submission addresses the unique challenges of real-time translation of long-form speech by developing a modular cascade system that adapts strong pre-trained models to streaming scenarios. We combine Whisper Large-V3-Turbo for ASR with the multilingual NLLB-3.3B model for MT, implementing lightweight adaptation techniques rather than training new end-to-end models from scratch. Our approach employs document-level adaptation with prefix training to enhance the MT model’s ability to handle incomplete inputs, while incorporating adaptive emission policies including a wait-k strategy and RALCP for managing the translation stream. Specialized buffer management techniques and segmentation strategies ensure coherent translations across long audio sequences. Experimental results on the ACL60/60 dataset demonstrate that our system achieves a favorable balance between translation quality and latency, with a BLEU score of 31.96 and non-computational-aware StreamLAAL latency of 2.94 seconds. Our final model achieves a preliminary score on the official test set (IWSLT25Instruct) of 29.8 BLEU. Our work demonstrates that carefully adapted pre-trained components can create effective simultaneous translation systems for long-form content without requiring extensive in-domain parallel data or specialized end-to-end training.
%R 10.18653/v1/2025.iwslt-1.35
%U https://aclanthology.org/2025.iwslt-1.35/
%U https://doi.org/10.18653/v1/2025.iwslt-1.35
%P 340-346
Markdown (Informal)
[MLLP-VRAIN UPV system for the IWSLT 2025 Simultaneous Speech Translation Translation task](https://aclanthology.org/2025.iwslt-1.35/) (Iranzo-Sánchez et al., IWSLT 2025)
ACL