Iterative Translation Refinement with Large Language Models
This addresses translation quality for users of language models, but it is incremental as it builds on existing LLM capabilities.
The paper tackles the problem of improving translation quality by iteratively prompting a large language model to self-correct translations, resulting in better fluency and naturalness compared to initial translations and human references, as indicated by human evaluations.
We propose iteratively prompting a large language model to self-correct a translation, with inspiration from their strong language understanding and translation capability as well as a human-like translation approach. Interestingly, multi-turn querying reduces the output's string-based metric scores, but neural metrics suggest comparable or improved quality. Human evaluations indicate better fluency and naturalness compared to initial translations and even human references, all while maintaining quality. Ablation studies underscore the importance of anchoring the refinement to the source and a reasonable seed translation for quality considerations. We also discuss the challenges in evaluation and relation to human performance and translationese.