Large Language Models as Annotators for Machine Translation Quality Estimation
This work addresses the practical deployment challenge of LLMs in MTQE for translation systems, though it is incremental as it builds on existing annotation and training frameworks.
The authors tackled the high inference cost of using large language models (LLMs) for machine translation quality estimation (MTQE) by proposing a method to generate MQM-style annotations with LLMs to train a COMET model, achieving competitive performance on segment-level QE for Chinese-English and English-German.
Large Language Models (LLMs) have demonstrated excellent performance on Machine Translation Quality Estimation (MTQE), yet their high inference costs make them impractical for direct application. In this work, we propose applying LLMs to generate MQM-style annotations for training a COMET model: following Fernandes et al. (2023), we reckon that segment-level annotations provide a strong rationale for LLMs and are key to good segment-level QE. We propose a simplified MQM scheme, mostly restricted to top-level categories, to guide LLM selection. We present a systematic approach for the development of a GPT-4o-based prompt, called PPbMQM (Prompt-Pattern-based-MQM). We show that the resulting annotations correlate well with human annotations and that training COMET on them leads to competitive performance on segment-level QE for Chinese-English and English-German.