关键词: 自然语言处理, 命名实体识别, 迁移学习, 跨语言, 注意力机制

Abstract:

With the development of deep learning, great success has been achieved in the field of natural language processing. In order to improve the performance of Chinese Named Entity Recognition（NER）, this paper proposes a novel method to transfer the information extracted from the English NER model. This method uses a translation model to translate Chinese into English, and applies the English NER model to extract features. After that, the method utilizes the attention weight of the translation model to transfer information. The features extracted from the pre-trained English NER model are used to improve the performance of Chinese NER. This method can transfer the task-specific information obtained by the pre-trained English NER model and enrich the representation of Chinese sentences. Experiments on two Chinese NER datasets show that the proposed method outperforms state-of-the-art methods.

Key words: natural language processing, named entity recognition, transfer learning, cross-lingual, attention mechanism