Verification method and device of intelligent question-answering system, computer equipment and storage medium

Abstract

The invention discloses a verification method and device for an intelligent question-answering system, computer equipment and a storage medium. The method comprises the steps: generating a triad according to a preset intention set and a knowledge graph, and the triad comprises an initial entity and an entity relation; generating a to-be-replied question through a preset question template accordingto the entity relationship and the initial entity; sending the to-be-replied question to a preset question-answer simulator, and receiving reply information of the question-answer simulator to the to-be-replied question; and verifying the triple, the to-be-replied question and the replied information by adopting a deep learning method to obtain a classification result. According to the embodiment, the to-be-replied problem is generated based on knowledge graph simulation; according to the method, the diversity and comprehensiveness of the to-be-replied questions are improved, the knowledge graph is combined with the deep learning method, replies of the preset question-answer simulator are verified, the problem that a return function cannot be effectively measured and quantified in deep learning is solved, and the reply accuracy of the preset question-answer simulator is improved.

Description

technical field [0001] The present invention relates to the field of artificial intelligence, in particular to a verification method, device, computer equipment and storage medium for an intelligent question answering system Background technique [0002] With the development of computer technology, human-computer intelligent question answering system has received extensive attention in recent years, and it is applied to intelligent customer service, chat robots and smart speakers. [0003] At present, most question answering systems for task types use reinforcement learning to realize their functions, but the question answering systems under reinforcement learning need to interact with users through a large amount of information, and the existing corpus cannot be directly used in training. Increased time cost. There is also a way of using a user simulator to implement a question answering system, but this question answering system lacks a suitable reward function, and the r...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a verification method, device, computer equipment, and storage medium for a

Images