Construction and Application of Ambiguity Resolution Translation System for English Legal Text Based on BERT Model

Abstract

 Legal English text is difficult to understand and translate due to its strong professionalism, rich terminology and multiple semantic interpretations. In order to promote China's rule of law culture and promote the construction of China's legal discourse system, this study proposes a BERT-based model to construct an ambiguity resolution translation system for English legal texts, hoping to accurately deal with long and difficult sentences in legal English, ambiguity resolution of polysemous words and terminology in different jurisdictions. Through targeted pre-training and fine-tuning, the BERT model enables it to learn specific grammatical rules, terminologies and context dependencies of legal English. This study explains the types and causes of ambiguity in legal translation, introduces the basic principles and advantages of BERT model in resolving ambiguity in legal translation, and discusses the architecture design, system optimization and evaluation effect of legal ambiguity resolution based on BERT model. The system is expected to provide a high-precision ambiguity processing framework for legal machine translation and lay a solid foundation for the exchange and research of cross-language legal information.