这篇综述对比了大型语言模型（LLM）与模型驱动逆向工程（MDRE）在从Java/Python代码抽象OCL（Object Constraint Language）规范中的表现，提出结合两者的LLM4Models方法，显著提升了规范生成的准确性（F1分数提高30%）和显式风格占比（达100%），为软件现代化提供了高效、低门槛的逆向工程解决方案。

This Diploma programme is specially designed for those who wish to be trained as a professional Chinese language teacher, qualify for a teaching career in international schools, language centres or ...