@inproceedings{liang-etal-2026-grad,
title = "{GRAD}: {GRAD}ient Rectification based on Authentic Data for Low-Resource Neural Machine Translation with Back Translation",
author = "Liang, Zhuofei and
Lyu, Xinglin and
Mu, Lingling and
Xu, Hongfei",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.768/",
pages = "16189--16200",
ISBN = "979-8-89176-462-0",
abstract = "Synthetic data augmentation, especially back-translation with Large Language Models (LLMs), has become a standard strategy for low-resource Neural Machine Translation (NMT). However, jointly training on synthetic and authentic data may suffer from conflicts. We propose GRADient Rectification based on Authentic Data (GRAD), an asymmetric gradient-level framework that treats authentic parallel data as an Authentic Anchor, and rectifies conflicting synthetic updates through orthogonal projection while preserving non-conflicting components. Furthermore, we leverage gradients of multiple authentic batch for comparison and use the most similar one for gradient rectification to alleviate information loss with the synthetic data. Experiments on six low-resource pairs across three LLM families (Qwen, Gemma, Llama) and different scales (4B/8B) show that GRAD consistently outperforms supervised fine-tuning, back-translation, and multi-stage baselines. Notably, on the lowest-resource En-Kk pair, GRAD achieves the highest COMET (83.65) among all evaluated systems. Further analysis shows that synthetic-induced gradient conflicts are localized, and parameter-wise rectification substantially outperforms global-wise rectification."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liang-etal-2026-grad">
<titleInfo>
<title>GRAD: GRADient Rectification based on Authentic Data for Low-Resource Neural Machine Translation with Back Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zhuofei</namePart>
<namePart type="family">Liang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinglin</namePart>
<namePart type="family">Lyu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lingling</namePart>
<namePart type="family">Mu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongfei</namePart>
<namePart type="family">Xu</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Synthetic data augmentation, especially back-translation with Large Language Models (LLMs), has become a standard strategy for low-resource Neural Machine Translation (NMT). However, jointly training on synthetic and authentic data may suffer from conflicts. We propose GRADient Rectification based on Authentic Data (GRAD), an asymmetric gradient-level framework that treats authentic parallel data as an Authentic Anchor, and rectifies conflicting synthetic updates through orthogonal projection while preserving non-conflicting components. Furthermore, we leverage gradients of multiple authentic batch for comparison and use the most similar one for gradient rectification to alleviate information loss with the synthetic data. Experiments on six low-resource pairs across three LLM families (Qwen, Gemma, Llama) and different scales (4B/8B) show that GRAD consistently outperforms supervised fine-tuning, back-translation, and multi-stage baselines. Notably, on the lowest-resource En-Kk pair, GRAD achieves the highest COMET (83.65) among all evaluated systems. Further analysis shows that synthetic-induced gradient conflicts are localized, and parameter-wise rectification substantially outperforms global-wise rectification.</abstract>
<identifier type="citekey">liang-etal-2026-grad</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.768/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>16189</start>
<end>16200</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GRAD: GRADient Rectification based on Authentic Data for Low-Resource Neural Machine Translation with Back Translation
%A Liang, Zhuofei
%A Lyu, Xinglin
%A Mu, Lingling
%A Xu, Hongfei
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F liang-etal-2026-grad
%X Synthetic data augmentation, especially back-translation with Large Language Models (LLMs), has become a standard strategy for low-resource Neural Machine Translation (NMT). However, jointly training on synthetic and authentic data may suffer from conflicts. We propose GRADient Rectification based on Authentic Data (GRAD), an asymmetric gradient-level framework that treats authentic parallel data as an Authentic Anchor, and rectifies conflicting synthetic updates through orthogonal projection while preserving non-conflicting components. Furthermore, we leverage gradients of multiple authentic batch for comparison and use the most similar one for gradient rectification to alleviate information loss with the synthetic data. Experiments on six low-resource pairs across three LLM families (Qwen, Gemma, Llama) and different scales (4B/8B) show that GRAD consistently outperforms supervised fine-tuning, back-translation, and multi-stage baselines. Notably, on the lowest-resource En-Kk pair, GRAD achieves the highest COMET (83.65) among all evaluated systems. Further analysis shows that synthetic-induced gradient conflicts are localized, and parameter-wise rectification substantially outperforms global-wise rectification.
%U https://aclanthology.org/2026.findings-emnlp.768/
%P 16189-16200
Markdown (Informal)
[GRAD: GRADient Rectification based on Authentic Data for Low-Resource Neural Machine Translation with Back Translation](https://aclanthology.org/2026.findings-emnlp.768/) (Liang et al., Findings 2026)
ACL