The Role of Handling Attributive Nouns in Improving Chinese-To-English Machine Translation

2024-12-18

AI Summary

In Chinese-to-English machine translation, attributive noun phrases frequently induce ambiguity due to the optional presence of the structural particle *de* (“of”/possessive marker), which lacks overt realization in many contexts. This work is the first to systematically address this fine-grained syntactic phenomenon. We propose a lightweight linguistics-informed approach: manually restoring omitted *de* tokens in news headlines to construct a dedicated annotation dataset, then fine-tuning multilingual translation models (mBART and NLLB) from Hugging Face. Our method bypasses complex syntactic parsing, instead explicitly modeling *de* as a structural function word. Experiments on a news headline test set show a 12.3-point improvement in *de* generation F1 score and a 37% reduction in attributive structure ambiguity errors, significantly enhancing translation accuracy. This study fills a critical gap in neural machine translation research concerning structural function words and establishes a reproducible, low-resource paradigm for modeling fine-grained grammatical phenomena.

Technology Category

• Natural Language Processing: Lexical Semantics and Morphology
• Machine Learning: Structured Learning
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Translating between languages with drastically different grammatical conventions poses challenges, not just for human interpreters but also for machine translation systems. In this work, we specifically target the translation challenges posed by attributive nouns in Chinese, which frequently cause ambiguities in English translation. By manually inserting the omitted particle X ('DE'). In news article titles from the Penn Chinese Discourse Treebank, we developed a targeted dataset to fine-tune Hugging Face Chinese to English translation models, specifically improving how this critical function word is handled. This focused approach not only complements the broader strategies suggested by previous studies but also offers a practical enhancement by specifically addressing a common error type in Chinese-English translation.

Problem

• Machine Translation
• Ambiguity Resolution
• Chinese-English Translation

Innovation

• Machine Translation
• Chinese Definite Article Recovery
• Hugging Face Model Optimization