Below are some simple introductions about my published papers as the first author, if you are interested, you can click on their corresponding web links to read the whole papers.
Master Thesis: 《Research on Machine Translation for the Titles of Scientific Papers with Contextual Information》
This paper proposes a method for improving Russian-to-Chinese machine translation of scientific paper titles by integrating context. Leveraging large language models, the study acquires specialized context from abstracts when traditional retrieval mechanisms fail. A neural machine translation model is then developed based on pivot word embeddings, addressing word mismatch challenges. Finally, a multi-encoder integration translation model is introduced to incorporate contextual information, significantly enhancing title translation quality. Experimental results demonstrate notable improvements in Bleu4 scores. The paper advocates the effective combination of context and English as a pivot language for enhanced Russian-to-Chinese machine translation of scientific paper titles, showcasing advancements with 26 images, 12 tables, and 70 references.
《ISTIC’s Thai-to-Chinese Neural Machine Translation System for CCMT’ 2022》
This paper introduces technical details of Thai-to-Chinese neural machine translation system of Institute of Scientific and Technical Information of China (ISTIC) for the 18th China Conference on Machine Translation (CCMT’ 2022). ISTIC participated in a low resource evaluation task: Thai-to-Chinese MT task. The paper mainly illuminates its system framework based on Transformer, data preprocessing methods and some strategies adopted in this system. In addition, the paper evaluates the system performance under different methods.
This paper describes the technical details of ISTIC’s neural machine translation systems for the 19th China Conference on Machine Translation (CCMT’ 2023). ISTIC participated in two evaluation tasks of machine translation (MT) The team participated in two machine translation(MT) evaluation tasks: Low resource MT task(Vietnamese↔Chinese, Czech↔Chinese, Lao↔Chinese, Mongolian↔Chinese) and Chinese-Centric Multilingual MT task(Vietnamese↔Chinese, Thailand↔Chinese, Kazakh↔Chinese, Hindi↔Chinese, Uyghur↔Chinese). Context-aware systems and a multilingual system are built for two tasks respectively. The paper mainly illuminates its system framework based on Transformer, data preprocessing methods and some strategies adopted in this system. In addition, the paper evaluates the system performance under different methods.