@inproceedings{maiti-etal-2026-sce-kiit,
title = "{SCE}-{KIIT}: {K}ok{LL}a{MA}: Cross-Task Adaptation for Low-Resource {E}nglish{--}Kokborok Translation",
author = "Maiti, Agniva and
Muley, Aarsh and
Sahoo, Sovan",
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.157/",
pages = "2183--2192",
ISBN = "979-8-89176-474-3",
abstract = "Kokborok is a critically under-resourced Tibeto-Burman language spoken by over one million people primarily in Tripura, India, and is absent from modern massively multilingual machine translation (MT) systems. We describe the SCE-KIIT submission to the WMT 2026 Shared Task on Low-Resource Indic Language Translation (Category 2: English-Kokborok). We ask whether a large language model can be adapted to translate a language using no parallel supervision at all. Our system, KokLLaMA-3.2-3B-Instruct, fine-tunes Llama-3.2-3B-Instruct via QLoRA purely on Kokborok conversational instruction data, and recovers translation behaviour at inference time through structured prompting and a rule-based post-processing pipeline. On the official WMT 2026 test set it obtains 5.11 BLEU / 27.09 chrF++ (EN-TRP) and 3.58 BLEU / 26.14 chrF++ (TRP-EN), placing third of four and third of five submitted primary systems, and coming within 1.9 BLEU of the best WMT 2025 system for this pair despite using no parallel data. The approach is substantially more effective at generating Kokborok than at generating English, and we analyse why. We further show that our in-domain validation split under-estimated official performance by more than an order of magnitude, a cautionary result for participants who tune on held-out slices of provided training corpora, and we characterise a pathological Translation Edit Rate (TER) failure mode in which an untrained baseline scores well by silently under-generating."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="maiti-etal-2026-sce-kiit">
<titleInfo>
<title>SCE-KIIT: KokLLaMA: Cross-Task Adaptation for Low-Resource English–Kokborok Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Agniva</namePart>
<namePart type="family">Maiti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aarsh</namePart>
<namePart type="family">Muley</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sovan</namePart>
<namePart type="family">Sahoo</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>Kokborok is a critically under-resourced Tibeto-Burman language spoken by over one million people primarily in Tripura, India, and is absent from modern massively multilingual machine translation (MT) systems. We describe the SCE-KIIT submission to the WMT 2026 Shared Task on Low-Resource Indic Language Translation (Category 2: English-Kokborok). We ask whether a large language model can be adapted to translate a language using no parallel supervision at all. Our system, KokLLaMA-3.2-3B-Instruct, fine-tunes Llama-3.2-3B-Instruct via QLoRA purely on Kokborok conversational instruction data, and recovers translation behaviour at inference time through structured prompting and a rule-based post-processing pipeline. On the official WMT 2026 test set it obtains 5.11 BLEU / 27.09 chrF++ (EN-TRP) and 3.58 BLEU / 26.14 chrF++ (TRP-EN), placing third of four and third of five submitted primary systems, and coming within 1.9 BLEU of the best WMT 2025 system for this pair despite using no parallel data. The approach is substantially more effective at generating Kokborok than at generating English, and we analyse why. We further show that our in-domain validation split under-estimated official performance by more than an order of magnitude, a cautionary result for participants who tune on held-out slices of provided training corpora, and we characterise a pathological Translation Edit Rate (TER) failure mode in which an untrained baseline scores well by silently under-generating.</abstract>
<identifier type="citekey">maiti-etal-2026-sce-kiit</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.157/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>2183</start>
<end>2192</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SCE-KIIT: KokLLaMA: Cross-Task Adaptation for Low-Resource English–Kokborok Translation
%A Maiti, Agniva
%A Muley, Aarsh
%A Sahoo, Sovan
%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 maiti-etal-2026-sce-kiit
%X Kokborok is a critically under-resourced Tibeto-Burman language spoken by over one million people primarily in Tripura, India, and is absent from modern massively multilingual machine translation (MT) systems. We describe the SCE-KIIT submission to the WMT 2026 Shared Task on Low-Resource Indic Language Translation (Category 2: English-Kokborok). We ask whether a large language model can be adapted to translate a language using no parallel supervision at all. Our system, KokLLaMA-3.2-3B-Instruct, fine-tunes Llama-3.2-3B-Instruct via QLoRA purely on Kokborok conversational instruction data, and recovers translation behaviour at inference time through structured prompting and a rule-based post-processing pipeline. On the official WMT 2026 test set it obtains 5.11 BLEU / 27.09 chrF++ (EN-TRP) and 3.58 BLEU / 26.14 chrF++ (TRP-EN), placing third of four and third of five submitted primary systems, and coming within 1.9 BLEU of the best WMT 2025 system for this pair despite using no parallel data. The approach is substantially more effective at generating Kokborok than at generating English, and we analyse why. We further show that our in-domain validation split under-estimated official performance by more than an order of magnitude, a cautionary result for participants who tune on held-out slices of provided training corpora, and we characterise a pathological Translation Edit Rate (TER) failure mode in which an untrained baseline scores well by silently under-generating.
%U https://aclanthology.org/2026.wmt-main.157/
%P 2183-2192
Markdown (Informal)
[SCE-KIIT: KokLLaMA: Cross-Task Adaptation for Low-Resource English–Kokborok Translation](https://aclanthology.org/2026.wmt-main.157/) (Maiti et al., WMT 2026)
ACL