Jul 2026· ACM Transactions on Asian and Low-Resource Language Information Processing· 0 citations· 6 references
Abstract
Large language models continue to face challenges in translating low-resource languages with scarce parallel data. This study investigates how to fine-tune them effectively using target-side monolingual data. Existing approaches—dominated by back-translation and recent LLM-based rewriting—remain limited by noisy synthetic sources, unguided simplification, and the absence of a principled mechanism for integrating monolingual sentences into the training objective. To address this, we developed a semi-supervised framework that integrates marginal distribution estimation and curriculum-guided rewriting to exploit monolingual data for low-resource translation. Experiments in four low-resource directions demonstrated substantial gains, averaging +8 spBLEU and +10 COMET over strong baselines, while three additional mid-resource directions showed stable improvements and consistent trends. Reference-free metrics further validated robust gains in fluency and adequacy. The findings establish a scalable paradigm for low-resource translation, revealing that the principled integration of marginal likelihood estimation and generative rewriting enables large language models to achieve superior performance under extreme data scarcity.
This work proposes augmenting existing benchmarks to increase translation difficulty by combining adversarial optimization with a differentiable translation difficulty estimator, and uses gradients from a combined difficulty and fluency objective to iteratively replace tokens in Adversarial Translation Optimization (ATO).
William Kalikman, Šimon Sukup, Michal Tesnar et al.· 0 citations
The Onramp-Sequence Cross-Distillation (OSCD) is introduced, a post-training algorithm that projects high-resource reasoning trajectories into low-resource vocabulary subspaces during generative training rollouts via an integrated translator agentic loop, ensuring the stable and efficient translation of dynamically generated reference samples for fine-tuning.
Sean Gip Lim, William-Chandra Tjhi, Hai Leong Chieu· 0 citations
We study reference-free post-training for multilingual machine translation with open large language models. Starting from the supervised-finetuned MiLMMT-46-v0.1 models, we apply Group Relative Policy Optimization (GRPO) with a reward that averages two reference-free quality estimation models and is gated by language identification. We then linearly interpolate the supervised fine-tuning (SFT) and reinforcement learning (RL) model checkpoints to obtain MiLMMT-46-v1.0. Across 46 languages, the resulting models consistently improve translation quality over their SFT counterparts, outperform strong recent open baselines, including Seed-X, HY-MT2, and TranslateGemma, and achieve leading reference-free scores against evaluated proprietary systems such as Google Translate, Gemini 3 Pro, and GPT-5. We further investigate on-policy distillation (OPD) and find that it reaches, but does not surpass, the quality frontier achieved by RL with checkpoint interpolation. We release the models and code to facilitate future research.
Chris Han, Pengzhi Gao, Pei Fu et al.· 0 citations
Empirical analyses show that MoSE improves high-, medium-, and low-resource languages simultaneously, with the largest gains on low-resource speech, thereby breaking the curse of multilinguality without compromising high-resource performance.
Yexing Du, Kaiyuan Liu, Youcheng Pan et al.· 0 citations
Machine translation for low-resource languages suffers from domain-imbalanced corpora, causing quality degradation on technical text. However, in-context learning opens the possibility to rely on limited in-domain corpora to inform translation. We present lessons learned from Tulun, a retrieval-augmented system combining neural MT with LLM post-editing, guided by user-configurable translation memories and glossaries. Deployed for medical translation in Timor-Leste (Tetun) and disaster relief translation in Vanuatu (Bislama), the system achieves accuracy improvements over baseline MT by 16.90-22.41 ChrF++ points, while offering rapid adaptability and transparency to end-users. Key recommendations include: domain granularity matters more than broad categories; translation target audience should inform retrieval; and RAG-augmented MT is most effective for languages that lack domain corpora but remain within LLM pretraining distributions.
Raphael Merx, Ekaterina Vylomova· Annual International ACM SIG...· 0 citations
Low-resource languages face significant challenges in Machine Translation (MT) due to the scarcity of high-quality parallel datasets. This study presents a low-resource language translation system using a multilingual Neural Machine Translation (NMT) approach. The proposed system employs the Multilingual Bidirectional and Auto-Regressive Transformer-50 (mBART-50) model. mBART-50 is a transformer-based multilingual architecture, fine-tuned on the Samanantar dataset to enhance translation quality for Indic languages. English–Tamil and English–Telugu translations are the primary experimental focus, in which cross-lingual transfer learning is leveraged to improve performance despite data scarcity. Model performance is evaluated using the Bilingual Evaluation Understudy (BLEU) score, which demonstrates an improvement over the pretrained baseline model. The mBART-50 is designed to be multilingual to enable its easy generalization to other low-resource Indic languages. This study extends previous research by assessing the effectiveness of multilingual transfer under the low-resource conditions of Indic languages. The fine-tuned model achieves BLEU scores of 45.91 for English–Tamil and 31.23 for English–Telugu, significantly outperforming the pretrained baseline. These results demonstrate the effectiveness of multilingual transfer learning in low-resource settings.
G. Harshitha, Vasudeva, Nisha P. Poojary et al.· Engineering, Technology &...· 0 citations