Document Summarization using Word and Part-of-speech based on Attention Mechanism

Automatic text summarization is one of the research hotspots in natural language processing (NLP). Recently, there has been a considerable research interest in summarization. While most of the current researches about automatic text summarization are only based on word embedding features, which has the problem of too few text features and can't effectively utilize some other features of the text. In this paper, we propose a novel deep learning model called Document Summarization using Word and Part-of-speech based on Attention Mechanism (WPABS). We employ the word embedding and part-of-speech embedding to make full use of the features of the text and evaluate our model on English datasets Gigaword and DUC-2004. The experimental results show that our model is better than most methods. Also, compared with WABS(Word Attention-Based Summarization), our proposed model performs better. It is proved that it is useful to combine the word and part-of-speech.