Multi-round dialogue reply selection method and system based on QA knowledge base reasoning

Abstract

The invention belongs to the field of natural language processing, particularly relates to a multi-round dialogue reply selection method and system based on QA knowledge base reasoning which aims to solve the problem of reply selection in a man-machine multi-round dialogue interaction system. The method comprises the steps of retrieving a candidate question set from a QA knowledge base by taking the keywords extracted from the current multi-round dialogues and the current input as questions, acquiring the corresponding contexts, and constructing a candidate multi-round dialogue set; calculating the semantic similarity between the current input and the candidate problem as a first similarity; calculating the semantic similarity between the currently input context and the context of each candidate problem as a second similarity; calculating the similarity of the abstract information of the current multi-round dialogue and each candidate multi-round dialogue as a third similarity; and carrying out weighted summation on the three similarities to obtain the similarity between each candidate question and the current input, and taking the reply corresponding to the candidate question withthe maximum similarity as an output reply, so that the quality of the reply statement can be effectively improved, and the user experience is enhanced.

Description

technical field [0001] The invention belongs to the technical field of natural language processing, and in particular relates to a multi-round dialogue reply selection method and system based on QA knowledge base reasoning. Background technique [0002] With the rapid development of artificial intelligence, especially in recent years, with the expansion of deep learning technology, natural language processing technology and artificially constructed knowledge base, the human-computer dialogue system has made great progress in the fields of smart home, smart assistant and smart customer service. development, and users have higher and higher requirements for the quality of chat content. [0003] The current chatbots can be divided into generative methods and retrieval methods according to different reply sentence generation strategies. Compared with the generative method, the retrieval-based method has the following advantages: 1) The model implementation is relatively simple;...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above problems in the prior art, that is, in order to solve the problem of insufficient robustness of the existing reply sentence production method, the first aspect of the present invention provides a multi-round dialog reply selection method based on QA knowledge base reasoning, Include the following steps:

Images