Chasing COMET: Leveraging Minimum Bayes Risk Decoding for Self-Improving Machine Translation
This work addresses efficient self-improvement in machine translation for domain adaptation and low-resource languages, representing an incremental advancement.
The paper tackles the problem of improving machine translation quality by using Minimum Bayes Risk decoding with COMET as a utility metric for self-improvement, resulting in significant enhancements for all examined language pairs, including domain-adapted models and low-resource settings.
This paper explores Minimum Bayes Risk (MBR) decoding for self-improvement in machine translation (MT), particularly for domain adaptation and low-resource languages. We implement the self-improvement process by fine-tuning the model on its MBR-decoded forward translations. By employing COMET as the MBR utility metric, we aim to achieve the reranking of translations that better aligns with human preferences. The paper explores the iterative application of this approach and the potential need for language-specific MBR utility metrics. The results demonstrate significant enhancements in translation quality for all examined language pairs, including successful application to domain-adapted models and generalisation to low-resource settings. This highlights the potential of COMET-guided MBR for efficient MT self-improvement in various scenarios.