Problem
Research questions and friction points this paper is trying to address.
low-resource machine translation
synthetic data generation
endangered languages
grammar books
parallel corpora
A Factorial Study of Synthetic Data Generation for Low-Resource Machine Translation using Grammar Books
🤖 AI Summary
Endangered languages often lack parallel corpora for machine translation despite having detailed descriptive grammars. This work proposes the first systematic approach to automatically extract grammatical rules, example sentences, and lexical entries from static grammar books using large language models, generating synthetic parallel data for fine-tuning rather than relying on in-context prompting during inference. Through a comprehensive factorial experiment across 96 configurations on three endangered languages—Kalamang, Tuatschin, and Mandan—the method demonstrates significant effectiveness: fine-tuned models achieve ChrF++ improvements of up to +8.8, +5.3, and +3.3, respectively, with 75% of configurations substantially outperforming the baseline. The study further identifies key effective combinations and failure boundaries in synthetic data construction for low-resource machine translation.
📝 Abstract
Most endangered languages lack the parallel data required for machine translation, despite the existence of descriptive grammar books. We introduce a pipeline that uses large language models to extract grammatical rules, example sentences, and lexicons from grammar books and generate synthetic parallel corpora for fine-tuning-rather than feeding grammar content into prompts at inference time, as in prior work. Validated on three typologically diverse low-resource languages-Kalamang (Papuan), Tuatschin (Romance), and Mandan (Siouan)-we show that fine-tuning on synthetic data improves over seed-data baselines in 75% of configurations for Kalamang and 59% for Tuatschin, with best-case ChrF++ gains of +8.8, +5.3, and +3.3 respectively. Through a systematic factorial study across 96 configurations varying target part-of-speech, retrieval granularity, and sample volume, we identify which factor combinations drive gains and where they break down. Our results demonstrate that static linguistic documentation can be repurposed for machine translation fine-tuning, offering a practical path towards translation tools for severely under-resourced languages.
Innovation
Methods, ideas, or system contributions that make the work stand out.
synthetic data generation
low-resource machine translation
grammar book mining
factorial study
fine-tuning
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
