@inproceedings{arora-etal-2026-bvslp,
title = "{BVSLP}: Enhancing Machine Translation through {NMT} Fine-Tuning, Back-Translation, and {LLM}-Assisted Synthetic Data Generation",
author = "Arora, Palak and
Sharma, Adhikarimayum and
Indoria, Mrityunjaya and
Lhoungu, Keneiwenuo and
Lalthafamkimi, Lalthafamkimi and
Nathani, Bharti and
Joshi, Nisheeth",
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.147/",
pages = "2086--2093",
ISBN = "979-8-89176-474-3",
abstract = "This paper presents a two-stage neural machine translation framework for English and selected low-resource Northeastern Indian languages. The approach incorporated transfer learning, bidirectional fine-tuning, back-translation, and LLM-assisted synthetic data augmentation. In Stage 1, the NMT model was fine-tuned on cleaned gold-standard parallel data. In Stage 2, additional synthetic sentence pairs were generated using back-translation and Qwen2.5-Instruct, followed by filtering based on language identification, length ratio, duplicate removal, named-entity preservation, semantic similarity, and round-trip consistency. The framework was evaluated on English{--}Assamese, English{--}Mizo, English{--}Manipuri, English{--}Bodo, English{--}Khasi, and Nagamese{--}English translation directions. Results show stronger performance for several Northeastern-language-to-English directions, while the contrastive system notably improved Manipuri (Bengali){--}English translation from 26.56 to 30.59 BLEU. The findings indicate that controlled synthetic augmentation can improve low-resource MT, although its effectiveness remains language- and direction-dependent."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="arora-etal-2026-bvslp">
<titleInfo>
<title>BVSLP: Enhancing Machine Translation through NMT Fine-Tuning, Back-Translation, and LLM-Assisted Synthetic Data Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Palak</namePart>
<namePart type="family">Arora</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Adhikarimayum</namePart>
<namePart type="family">Sharma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mrityunjaya</namePart>
<namePart type="family">Indoria</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Keneiwenuo</namePart>
<namePart type="family">Lhoungu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lalthafamkimi</namePart>
<namePart type="family">Lalthafamkimi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bharti</namePart>
<namePart type="family">Nathani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nisheeth</namePart>
<namePart type="family">Joshi</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>This paper presents a two-stage neural machine translation framework for English and selected low-resource Northeastern Indian languages. The approach incorporated transfer learning, bidirectional fine-tuning, back-translation, and LLM-assisted synthetic data augmentation. In Stage 1, the NMT model was fine-tuned on cleaned gold-standard parallel data. In Stage 2, additional synthetic sentence pairs were generated using back-translation and Qwen2.5-Instruct, followed by filtering based on language identification, length ratio, duplicate removal, named-entity preservation, semantic similarity, and round-trip consistency. The framework was evaluated on English–Assamese, English–Mizo, English–Manipuri, English–Bodo, English–Khasi, and Nagamese–English translation directions. Results show stronger performance for several Northeastern-language-to-English directions, while the contrastive system notably improved Manipuri (Bengali)–English translation from 26.56 to 30.59 BLEU. The findings indicate that controlled synthetic augmentation can improve low-resource MT, although its effectiveness remains language- and direction-dependent.</abstract>
<identifier type="citekey">arora-etal-2026-bvslp</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.147/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>2086</start>
<end>2093</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T BVSLP: Enhancing Machine Translation through NMT Fine-Tuning, Back-Translation, and LLM-Assisted Synthetic Data Generation
%A Arora, Palak
%A Sharma, Adhikarimayum
%A Indoria, Mrityunjaya
%A Lhoungu, Keneiwenuo
%A Lalthafamkimi, Lalthafamkimi
%A Nathani, Bharti
%A Joshi, Nisheeth
%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 arora-etal-2026-bvslp
%X This paper presents a two-stage neural machine translation framework for English and selected low-resource Northeastern Indian languages. The approach incorporated transfer learning, bidirectional fine-tuning, back-translation, and LLM-assisted synthetic data augmentation. In Stage 1, the NMT model was fine-tuned on cleaned gold-standard parallel data. In Stage 2, additional synthetic sentence pairs were generated using back-translation and Qwen2.5-Instruct, followed by filtering based on language identification, length ratio, duplicate removal, named-entity preservation, semantic similarity, and round-trip consistency. The framework was evaluated on English–Assamese, English–Mizo, English–Manipuri, English–Bodo, English–Khasi, and Nagamese–English translation directions. Results show stronger performance for several Northeastern-language-to-English directions, while the contrastive system notably improved Manipuri (Bengali)–English translation from 26.56 to 30.59 BLEU. The findings indicate that controlled synthetic augmentation can improve low-resource MT, although its effectiveness remains language- and direction-dependent.
%U https://aclanthology.org/2026.wmt-main.147/
%P 2086-2093
Markdown (Informal)
[BVSLP: Enhancing Machine Translation through NMT Fine-Tuning, Back-Translation, and LLM-Assisted Synthetic Data Generation](https://aclanthology.org/2026.wmt-main.147/) (Arora et al., WMT 2026)
ACL
- Palak Arora, Adhikarimayum Sharma, Mrityunjaya Indoria, Keneiwenuo Lhoungu, Lalthafamkimi Lalthafamkimi, Bharti Nathani, and Nisheeth Joshi. 2026. BVSLP: Enhancing Machine Translation through NMT Fine-Tuning, Back-Translation, and LLM-Assisted Synthetic Data Generation. In Proceedings of the Eleventh Conference on Machine Translation, pages 2086–2093, Budapest, Hungary. Association for Computational Linguistics.