论文标题
通过利用个人域专业知识来完善机器翻译中的最新技术，为JA <-> EN语言对优化NMT
Refining the state-of-the-art in Machine Translation, optimizing NMT for the JA <-> EN language pair by leveraging personal domain expertise
论文作者
论文摘要
基于变压器体系结构利用OpenNMT框架，记录用于EN/JA的NMT（神经机转换）系统的构建。对Corpora预处理，高参数调整和模型架构进行系统的探索以获得最佳性能。使用标准自动评估指标（例如BLEU）以及我作为日本语言学家的主观观点对该系统进行评估。
Documenting the construction of an NMT (Neural Machine Translation) system for En/Ja based on the Transformer architecture leveraging the OpenNMT framework. A systematic exploration of corpora pre-processing, hyperparameter tuning and model architecture is carried out to obtain optimal performance. The system is evaluated using standard auto-evaluation metrics such as BLEU, and my subjective opinion as a Japanese linguist.
