@inproceedings{jana-etal-2026-indicmt,
title = "{I}ndic{MT} at {WMT} 2026: Full Fine-Tuning of {NLLB}-200-600{M} with Round-Trip Filtering and Forward Self-Training for {A}rabic-{A}sian Translation",
author = "Jana, Pramit and
Adhikari, Arindam and
Acharya, Priyobroto and
Das, Dipankar",
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.172/",
pages = "2311--2317",
ISBN = "979-8-89176-474-3",
abstract = "We describe the IndicMT submission to the WMT 2026 Shared Task on Low-Resource Arabic-Asian Language Translation. Our system covers all ten translation directions between Arabic and Bangla, English, Hindi, Indonesian, and Urdu. For each direction, we fully fine-tune a separate NLLB-200-distilled-600M model using two NVIDIA T4 graphics processing units. The training pipeline combines Unicode normalization and length filtering with direction-specific round-trip consistency filtering of the available parallel data. We additionally augment each direction with 500 source-side sentences whose target translations are generated by the pretrained NLLB model, a procedure corresponding to forward self-training rather than conventional back-translation. On the official test set, our strongest relative results are obtained for Arabic{\$}{\textbackslash}rightarrow{\$}Indonesian and Indonesian{\$}{\textbackslash}rightarrow{\$}Arabic. Across all five language pairs, translation from Arabic obtains higher BLEU scores than translation into Arabic. Our results demonstrate that full adaptation of a compact multilingual translation model, combined with model-based data selection and synthetic-target augmentation, provides a practical baseline under limited computational resources."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jana-etal-2026-indicmt">
<titleInfo>
<title>IndicMT at WMT 2026: Full Fine-Tuning of NLLB-200-600M with Round-Trip Filtering and Forward Self-Training for Arabic-Asian Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Pramit</namePart>
<namePart type="family">Jana</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arindam</namePart>
<namePart type="family">Adhikari</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Priyobroto</namePart>
<namePart type="family">Acharya</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dipankar</namePart>
<namePart type="family">Das</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 the IndicMT submission to the WMT 2026 Shared Task on Low-Resource Arabic-Asian Language Translation. Our system covers all ten translation directions between Arabic and Bangla, English, Hindi, Indonesian, and Urdu. For each direction, we fully fine-tune a separate NLLB-200-distilled-600M model using two NVIDIA T4 graphics processing units. The training pipeline combines Unicode normalization and length filtering with direction-specific round-trip consistency filtering of the available parallel data. We additionally augment each direction with 500 source-side sentences whose target translations are generated by the pretrained NLLB model, a procedure corresponding to forward self-training rather than conventional back-translation. On the official test set, our strongest relative results are obtained for Arabic$\textbackslashrightarrow$Indonesian and Indonesian$\textbackslashrightarrow$Arabic. Across all five language pairs, translation from Arabic obtains higher BLEU scores than translation into Arabic. Our results demonstrate that full adaptation of a compact multilingual translation model, combined with model-based data selection and synthetic-target augmentation, provides a practical baseline under limited computational resources.</abstract>
<identifier type="citekey">jana-etal-2026-indicmt</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.172/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>2311</start>
<end>2317</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T IndicMT at WMT 2026: Full Fine-Tuning of NLLB-200-600M with Round-Trip Filtering and Forward Self-Training for Arabic-Asian Translation
%A Jana, Pramit
%A Adhikari, Arindam
%A Acharya, Priyobroto
%A Das, Dipankar
%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 jana-etal-2026-indicmt
%X We describe the IndicMT submission to the WMT 2026 Shared Task on Low-Resource Arabic-Asian Language Translation. Our system covers all ten translation directions between Arabic and Bangla, English, Hindi, Indonesian, and Urdu. For each direction, we fully fine-tune a separate NLLB-200-distilled-600M model using two NVIDIA T4 graphics processing units. The training pipeline combines Unicode normalization and length filtering with direction-specific round-trip consistency filtering of the available parallel data. We additionally augment each direction with 500 source-side sentences whose target translations are generated by the pretrained NLLB model, a procedure corresponding to forward self-training rather than conventional back-translation. On the official test set, our strongest relative results are obtained for Arabic$\textbackslashrightarrow$Indonesian and Indonesian$\textbackslashrightarrow$Arabic. Across all five language pairs, translation from Arabic obtains higher BLEU scores than translation into Arabic. Our results demonstrate that full adaptation of a compact multilingual translation model, combined with model-based data selection and synthetic-target augmentation, provides a practical baseline under limited computational resources.
%U https://aclanthology.org/2026.wmt-main.172/
%P 2311-2317
Markdown (Informal)
[IndicMT at WMT 2026: Full Fine-Tuning of NLLB-200-600M with Round-Trip Filtering and Forward Self-Training for Arabic-Asian Translation](https://aclanthology.org/2026.wmt-main.172/) (Jana et al., WMT 2026)
ACL