Implementation method for fusing network question and answer system based on multi-attention mechanism

Abstract

The invention discloses an implementation method of a fusion network question and answer system based on a multi-attention mechanism, which comprises the following steps of constructing a question andanswer system network model, preprocessing an original data set to obtain a standby data set, and performing text length distribution analysis; subjecting text in standby data set to one-hot vector representation, using a CBOW model to train one-hot word vector and forming a word2vec word list; adjusting the sequence length of each sentence in the text, and adding a sentence end mark; training the word2vec vector by using an ELMO language model to obtain an ELMO word vector; encoding the ELMO vector to obtain a sentence vector; performing coarse-fine granularity attention on the sentence vectors respectively to obtain memory vectors and attention vectors based on each word; carrying out vector splicing to obtain expression vectors based on words and sentences; and decoding an answer representing the vector generation question sentence. According to the method, the representation ability of sentences is improved through an ELMO language model; and various attention mechanisms are fused, so that the decision making accuracy of the system is improved, and the interpretability of the system is enhanced.

Description

technical field [0001] The present invention relates to the field of question answering systems, and more specifically, to a method for realizing a network question answering system based on a multi-attention mechanism fusion. Background technique [0002] Question and answer is one of the main ways of human communication. With the continuous growth of data scale, more and more people are paying attention to how to quickly obtain the answers they want from massive amounts of information, while traditional search engine systems can no longer keep up. With the pace of the times, the accuracy and diversity of its search results need to make huge changes to meet the needs of users. Foreign researchers believe that an automatic question answering system that answers users' natural language questions in a direct and accurate manner will constitute the basic form of the next generation of search engines. Different from the traditional search engine system, the question answering s...

AI Technical Summary

Problems solved by technology

[0003] The present invention provides an implementation of a network question answering system based on a multi-attention mechanism fusion in order to overcome the defects of insufficie

Images