04:00
2026-07-27
machinebrief.com
artificial-intelligence
A Factorial Study of Synthetic Data Generation for Low-Resource Machine Translation using Grammar Books
A new pipeline using large language models to extract grammatical rules and lexicons from grammar books generates synthetic parallel corpora for fine-tuning machine translation models, improving over โฆ