My research project
Sentiment preservation in neural machine translation
Despite the progress of neural machine translation in recent years, the translation of emotion-intensive texts like microblogs on social media is still not satisfactory. While current studies in natural language processing focus on developing larger general-purpose language models and using fewer parallel data, this project investigates the errors made by current neural machine translation systems and its linguistic background, and then develops a neural machine translation system specifically targeted to emotion-intensive texts to better convey human emotions.
- Machine Translation;
- Sentiment Analysis;
- Topic Modeling;
- Statistical Analysis of Translation Process;
- Topics Related to Language Technology