| Large document writing is a challenging issue for businesses and users.With the advent of the information age,the channels for users to search for materials in writing are becoming more and more complicated.So that browsing,screening,understanding,and preserving the information become more and more difficult.Automatic summarization is a technique for extracting its central ideas from one or more documents.It helps people quickly read documents and extract important information from them.In academia,sequence-to-sequence architecture with attention mechanism is widely used in abstractive summarization,and many related works have achieved a series of remarkable results.However,this method may suffer from error accumulation.That is to say,at the testing stage,the input of decoder is the word generated at the previous time,so that decoder-side error will be continuously amplifed.This paper proposes a Summarization model using a Bidirectional decoder(BiSum),in which the backward decoder provides a reference for the forward decoder.We use attention mechanism at both encoder and backward decoder sides to ensure that the summary generated by backward decoder can be understood.Also,pointer mechanism is added in both the backward decoder and the forward decoder to solve the out-of-vocabulary problem.We remove the word segmentation step in regular Chinese preprocessing,which greatly improves the quality of summary.Experimental results show that our work can produce higher-quality summary on Chinese datasets TTNews and English datasets CNN/Daily Mail.At the same time,this paper implements the auxiliary document writing system based on BiSum.The system also assists in writing through tag extraction,extractive automatic summarization,web page noise reduction and other technologies,which help users quickly filter and read massive amounts of documents.The system provides the function of document import,document and material management,structured label management,and document template definition.At the same time,it supports multiple users to complete one large document together.The system fully integrates the above models and functions with the Django framework and React component development technology. |
