machine-translation
Filtering by topic machine-translation(3)Clear all filters
- PaperLanguage Teaching Research30 Jun 2026
Strategically Informed Machine Translation Post-Editing: Enhancing Translation Performance among Intermediate Chinese EFL Learners
Zhiying Li
Intermediate Chinese EFL learners completed human translation and machine translation post-editing (MTPE) tasks guided by strategies. Post-edited texts showed higher lexical and syntactic complexity than human translations, but learners struggled with error detection and maintained neutral-to-negative attitudes toward PE. The study proposes refined PE guidelines and a multi-task design to enhance translation performance.
Original abstract
Machine translation (MT) errors may be seen as a limitation, but it is precisely these imperfections that provide learners with valuable opportunities for error-driven practice through post-editing (PE), thus developing their language proficiency and, in turn, supporting translation performance. However, little is known about how machine translation post-editing (MTPE) benefits intermediate learners of English as a foreign language (EFL). This empirical study explores the role of strategically informed MTPE practices in supporting intermediate Chinese EFL learners’ linguistic and translation development. Participants completed human translation and MTPE tasks guided by PE strategies, followed by reflection on their edits. Text complexity analysis using Eng-Editor revealed that post-edited texts exhibited higher overall textual, lexical, and syntactic complexity than human translations, with correct revisions marginally exceeding errors, suggesting that intermediate learners still struggle with error detection and correction. These findings underscore the need for clearer strategic support in PE tasks. Sentiment analysis of learner reflections based on SnowNLP with contextual adjustments showed a neutral-to-negative attitude toward PE, highlighting challenges related to syntactic proficiency and technical support. This study contributes empirical evidence on pedagogical MTPE use with intermediate learners, leading to refined PE guidelines and a multi-task design that emphasizes both selection competence and linguistic development to enhance translation performance.
- PaperarXiv — Language & NLP (cs.CL)4 Jun 2026
Reinforcement Learning Elicits Contextual Learning of Unseen Language Translation
Hanxu Hu, Zdeněk Šnajdr, Pinzhen Chen, Jannis Vamvas et al.
Reinforcement learning with a surface-level translation metric (chrF) as reward enables LLMs to acquire the meta-skill of using in-context linguistic knowledge for translating completely unseen languages, outperforming both in-context learning and supervised fine-tuning. This approach goes beyond memorization and suggests a general recipe for language learning from context.
Original abstract
Prior work has shown that large language models (LLMs) can translate unseen or low-resource languages by undergoing continued training or even by encoding a grammar book in their context. However, both methods typically overfit specific languages, with limited zero-shot transfer at test time. To translate extremely low-resource languages at scale, we argue that LLMs must acquire the meta-skill of utilizing in-context linguistic knowledge rather than memorizing specific languages. In this paper, we propose a reinforcement learning (RL) approach to unseen language translation given rich linguistic context, using a surface-level translation metric (chrF) as the reward. Empirically, despite the lightweight reward, our RL-trained models effectively extract and apply relevant linguistic information from the provided context, leading to better translations on completely unseen languages than in-context learning or supervised fine-tuning. Our analyses suggest that outcome-based RL can extend beyond conventional reasoning tasks like math and coding to serve as a recipe for language learning from context.
- PaperDOAJ — ELT & TESOL1 May 2026
Machine translation errors as learning resources: Developing an English-Indonesian dataset for English language teaching
Firqo Amelia, Aenor Rofek, Evynurul Laily Zen
Developed an Indonesian Grammatical Error Correction (GEC) dataset from machine translation errors, using Statistical Machine Translation and Sederet.com as primary error sources. Identified semantic, diction, synonymy, conjunction, and preposition errors as most prevalent, offering authentic materials for translation-based English teaching activities.
Original abstract
This study aims to develop an Indonesian Grammatical Error Correction (GEC) dataset using Statistical Machine Translation (SMT) and Sederet.com as the primary source of grammatical errors. Additionally, this research seeks to identify the types of errors in Indonesian translations of English sentences produced by Machine Translation (MT) that can serve as learning resources in English language teaching. This study extracts data from social media (X) as texts and processes them across three main stages to construct a GEC dataset. The first stage involves data collection, comprising the source, target, and control texts. The second stage consists of translation error analysis, which is conducted using Nord’s (2005) Translation Problems Theory. The third stage involves data annotation, which is performed using the UAM CorpusTool software. To identify translation errors, this study compared MT translations with those of professional translators, which were then validated by professional editors. The findings revealed that linguistic errors, particularly those related to semantics, diction, synonymy, conjunctions, and prepositions, were the most prevalent and relevant categories for inclusion in the dataset. These errors were identified through comparisons between the target and control texts and were subsequently annotated. Through the stages of schema creation, data input, error labelling, and correction insertion, a fully annotated corpus was produced. The implications of this study extend to both research and pedagogy. The dataset model supports the advancement of Indonesian GEC systems and offers teachers authentic materials to engage learners in translation-based activities.