Text abstract generation method, system and device and storage medium

Abstract

The invention discloses a text abstract generation method, which comprises the steps of performing word embedding processing, local attention processing and convolution operation on an input text sequence to obtain a feature expression matrix of the text sequence; and inputting the feature expression matrix of the text sequence into the trained abstract generation model to obtain an output result,and taking the output result as an abstract of the text sequence. According to the invention, the obtained feature expression matrix of the text sequence can better express deeper hidden features ofthe input text, and features of input text are further extracted; compared with the prior art, the method focuses on the text information generated in the previous decoding step, so that the semantemeof the finally generated abstract text is coherent, and the problems that the semanteme of the generated abstract is irrelevant and the statement is not smooth in the prior art are solved. The invention furthermore provides a text abstract generation system and device, and a computer readable storage medium, which have the above beneficial effects.

Description

technical field [0001] The present application relates to the field of text summary generation, and in particular to a method, system, device and computer-readable storage medium for text summary generation. Background technique [0002] With the development of information technology, most of the current popular automatic text summarization methods are based on deep learning. Among them, the sequence-to-sequence (Seq2Sseq) model based on the attention mechanism (sometimes called the encoder -decoder model, encoder-decoder model) generative summarization methods are particularly common. Since the model combines Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN) and attention mechanism, it can better extract hidden feature information contained in text, compared with traditional natural language processing method has more obvious advantages. [0003] However, the existing neural network-based text summarization methods ignore the text information generated ...

AI Technical Summary

Problems solved by technology

[0003] However, the existing neural network-based text summarization methods ignore the text information generated in the previous decoding step, which leads to the problems of irrelevant semantics and unclear sentences in the generated summaries.

[0004] Therefore, how to avoid semantically irrelevant and incoherent sentences in the generated text summary is a technical problem that those skilled in the art need to solve at present.

Images