Generation method and device of intelligent question and answer model, computing equipment and storage medium

Abstract

The invention provides a method for generating an intelligent question and answer model, which comprises the following steps of: expanding an original sample data set to enable positive samples corresponding to questions in the original sample data set to exist, and generating a new sample data set; the generated new sample data set is input into an existing question and answer model, the existing question and answer model obtains feature codes related to questions from a knowledge base and documents in the new sample data set, and the loss Lqa of the existing question and answer model is calculated; adding the obtained feature codes into an improved comparative learning model, and respectively calculating the similarity between difficult positive and negative samples in a real sample data set and a new sample data set to obtain a comparative loss Lcl; and combining the loss Lqa of the existing question and answer model with the comparison loss Lcl to obtain the loss L of a final question and answer model, and training the final question and answer model. The comparative learning model is added to the existing question-answering model, the generalization ability of the model is higher, and meanwhile, the effect obtained by question-answering in a knowledge base is better than that obtained by a traditional method.

Description

technical field [0001] The present invention relates to the field of natural language processing, in particular to a method, device, computing device and storage medium for generating an intelligent question answering model. Background technique [0002] In the question answering system based on natural language, the knowledge base is a necessary resource for answering factual questions. The knowledge base can also be called a knowledge graph, which is widely used in major browsers and recommendation algorithms. For example, Google Chrome has GoogleKnowledge Graph, Microsoft's Bing search has Bing Satori, Baidu's Baidu Knowledge Graph, and Facebook Social Graph in the recommendation system Facebook and Alicoco in Taobao, etc. [0003] The knowledge base often contains a lot of triple information, each triple information describes a fact, the general form of triple includes two forms: (head entity, relationship, tail entity) and (entity, attribute) ,value). Taking the first...

AI Technical Summary

Problems solved by technology

[0005] However, the current question-answering model based on the knowledge base does not have a deep understanding of the information in the knowledge base, that is, the knowledge base may lack critical knowledge that can be used to answer questions correctly, resulting in a low answer accuracy rate of the question-answering model. to further enhance

Images