Skip to content
Open access

Every Word Presented in Context: Syntactic Coverage as Objective for Low-Resource Machine Translation with Large Language Models

2026 · International Conference on Language Resources and Evaluation · pp. 8824-8837 · 1 citation · 52 references
Computer Science

TL;DR

Fragment-Shot Prompting is introduced, a novel few-shot prompting method that aims to retrieve examples for every word occurring in the sentence to be translated, illustrating their use and meaning in context.

Abstract

Large Language Models (LLMs) have demonstrated strong capabilities in multilingual machine translation. However, they underperform for low-resource languages, indicating the need for more explicit instructional guidance. In this work, we introduce Fragment-Shot Prompting, a novel few-shot prompting method that aims to retrieve examples for every word occurring in the sentence to be translated, illustrating their use and meaning in context. We evaluate our method on translation between Italian, Ladin (Val Badia) and Ladin (Gherdëina) and compare its performance with zero-shot prompting, random few-shot prompting, as well as established lexical and semantic retrieval strategies. We conduct these experiments using state-of-the-art LLMs, including GPT-3.5, GPT-4o, o1-mini, LlaMA-3.3, and DeepSeek-R1. Our results demonstrate that LLMs can extract substantial value from limited data when translating from a low-to the high-resource language. However, this does not apply to translations into the low-resource languages, where the prompting method plays a much more important role. In particular, our method consistently delivers the best results and enables significant gains. Even though translation performance into Ladin remains limited with the available resources, our results highlight the importance of syntactic coverage for improving translation accuracy and variant-specific adaptation in low-resource scenarios.

Read PDF

Similar papers

#artificial intelligence Preprint Oct 2026

Back-Translation Still Matters: Leveraging Monolingual Data for Low-Resource Machine Translation with LLMs

Pre-trained Large Language Model (LLMs) are effective on many tasks in zero-shot and few-shot settings. For machine translation of low-resource languages, the obvious options are few-shot prompting and fine-tuning on a small parallel set. We ask: what if moderate amounts of low-resource monolingual data are also availa...

Divyanshu Bathla, Parag Singla, Jitin Singla · 0 citations
Open access Sep 2026

Lexical diversity and translation literalness in machine translation: a comparative study of pretrained, fine-tuned, and few-shot approaches

This paper presents a comparative study of neural machine translation (MT) models and large language models, emphasizing translation literalness, lexical diversity, and translation quality. We evaluate open-source neural machine translation (NMT) models, LLaMa 3.1 Instruct, and GPT-4o across nine language pairs in pr...

Zeynep Yirmibeşoğlu Balal, Tunga Güngör · 0 citations
Open access Oct 2026

Resource Sufficient Data Generation Technique for Machine Translation Models Using Named Entity Transliteration and Cross- Lingual Word Embeddings

Machine translation offers speed and cost advantages, but human translation remains superior in quality and contextual understanding. For low-resource languages like Hausa, the severe scarcity of parallel data leads to poor model performance, overfitting, and inadequate handling of named entities. While existing synthe...

Rabiu Ibrahim Abdullahi, G. Obunadike, A. Abdullahi et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.