Conversation interaction method and device, equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a dialogue interaction method and device, equipment and a medium. The method comprises the steps of obtaining a to-be-replied dialogue text; based on a pre-trained classification model, classifying the dialogue text, and obtaining emotional polarities and theme tags corresponding to the dialogue text; determining a corresponding dialogue strategy based on the emotion polarity and the theme label; and based on a pre-trained generative pre-training GPT model, determining a target reply text corresponding to the dialogue strategy. The emotion and theme element controllable mode is introduced in the reply generation process of the dialogue system, the dialogue of the user is divided into more accurate emotion and theme classification, the corresponding dialogue strategy can be determined according to the emotion and theme classification, the reply generation direction of the dialogue system is flexibly adjusted, and the interaction experience quality of the dialogue system is improved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, and in particular, to a dialog interaction method, device, device and medium. Background technique [0002] With the continuous development of artificial intelligence technology, dialogue systems have developed rapidly in the commercial field, and many products have entered the stage of mass production. Among the dialogue systems in limited fields, Apple Siri and Microsoft Cortana have been used by people through mobile phones, operating systems and other media. Widely used, in addition, smart speakers equipped with virtual assistants such as Amazon Echo, Baidu Duer, Google Home and Tmall Genie have also entered thousands of households, and the human-machine dialogue system has gradually changed from the role of service to the role of emotional partner. [0003] In recent years, with the rapid growth of social data on the Internet, data-driven open-domain dialogue systems ...

AI Technical Summary

Problems solved by technology

[0003] In recent years, with the rapid growth of social data on the Internet, data-driven open domain dialogue systems have gradually become a hot topic in industry and academia. The application of deep learning technology has made breakthroughs in open domain dialogue systems. The end-to-end neural network model has been able to generate fluent and in line with the above responses in open-domain dialogue scenarios. However, the responses generated by training the neural network model only by learning a large amount of social data have general and uninteresting characteristics, and even generate responses to these questions. The reply is out of date. In the process of human-computer interaction, in the open-domain dialogue scenario, the dialogue system cannot pay attention to the theme and emotion-related interaction in the dialogue interaction, and it is even more unable to control the theme and emotion generated by the reply, which is automatically generated. Replies cannot achieve accurate sentiment and topic classification in the dialogue with the user, making the user's interaction experience poor

Images