Cultural Adaptation of Menus: A Fine-Grained Approach
This work addresses the problem of cultural adaptation in machine translation for menus, offering incremental improvements by integrating human translation theories into LLM processes.
The paper tackles the challenge of translating culture-specific items (CSIs) in menus by introducing the ChineseMenuCSI dataset and a novel automatic identification method, resulting in up to a 7-point increase in COMET scores for translation accuracy.
Machine Translation of Culture-Specific Items (CSIs) poses significant challenges. Recent work on CSI translation has shown some success using Large Language Models (LLMs) to adapt to different languages and cultures; however, a deeper analysis is needed to examine the benefits and pitfalls of each method. In this paper, we introduce the ChineseMenuCSI dataset, the largest for Chinese-English menu corpora, annotated with CSI vs Non-CSI labels and a fine-grained test set. We define three levels of CSI figurativeness for a more nuanced analysis and develop a novel methodology for automatic CSI identification, which outperforms GPT-based prompts in most categories. Importantly, we are the first to integrate human translation theories into LLM-driven translation processes, significantly improving translation accuracy, with COMET scores increasing by up to 7 points.