Question and answer method and device based on neural network, storage medium and terminal

Abstract

The invention discloses a question and answer method and device based on a neural network, a storage medium and a terminal. The method comprises the steps of determining a document associated with a question input by a user; respectively preprocessing the question and the document to obtain a vector representation of the question and a vector representation of the document; inputting the vector representation of the problem and the vector representation of the document into a preset neural network model, carrying out semantic analysis and matching on the document, and obtaining at least one candidate answer associated with the question and the degree of association of each candidate answer from the document, wherein the preset neural network model is obtained by performing semantic analysis and matching on the document on the basis of history; and determining the candidate answer with the highest correlation degree as the answer of the question and outputting the answer. Through the scheme of the invention, a question and answer system with higher universality and higher accuracy can be provided, and the intelligent answering of questions asked by a user is realized.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a neural network-based question answering method and device, a storage medium, and a terminal. Background technique [0002] The intelligent question answering system is a typical application of human-computer interaction. When the user asks a question, the intelligent question answering system gives the answer to the question. [0003] In the existing question answering system, widely used technologies include information retrieval, rule matching, and end-to-end generative technology. Among them, the question answering system based on information retrieval and rule matching technology needs to define a large number of question answering knowledge bases in advance, which is difficult to construct, takes a long time to project, and is not easy to promote the system and expand the field; based on the end-to-end generative question answering system, the gener...

AI Technical Summary

Problems solved by technology

Among them, the question answering system based on information retrieval and rule matching technology needs to pre-define a large number of question answering knowledge bases, which is difficult to construct, takes a long time for the project, and is not easy to promote the system and expand the field; based on the end-to-end generative question answering system, the generated answers It is highly dependent on training data and has certain errors, so it is difficult to apply to vertical business fields with low error tolerance

Images