Research on Translation Corpus Building with the Assistance of AI

The quality of machine translation depends on the corpus available for it to learn from. With the development of AI, the translation corpus is now the key to realize Intelligent translation and scenario-based translation. The data on the Internet, data of the enterprise, user-generated data, machine-generated data, all of which can be the sources of AI-Assisted translation corpus data. Types of these data are diversified. There are term data, text data, unstructured data, and knowledge graph. AI-Assisted translation corpus construction can collect data in the following four modes: open-source data based on the third-party, corpus-sharing mode based on crowdsourcing translation, self-learning mode based on machine, collaborative mode based on human-machine cooperation. In the future, corpus-based translation teaching will be an important part of intelligent translation.