Topic prediction and emotional reasoning integrated emotional dialogue generation method

Abstract

The invention relates to the field of natural language processing, deep learning, text sentiment analysis and dialogue generation, in particular to an emotional dialogue generation method fusing topic prediction and sentiment reasoning, which comprises the following steps: constructing an emotional dialogue generation model fusing topic prediction and sentiment reasoning; the model comprises a topic prediction module, an emotion reasoning module and a dialogue generation module, and the topic prediction module is used for performing context-controlled dialogue topic prediction to obtain a predicted topic; the emotion reasoning module is used for predicting context emotion labels and identifying emotion reason words to obtain corresponding emotion information; inputting the obtained theme information and emotion information into a dialogue generation module to obtain a common-situation reply which resonates with the emotion of the user; according to the method, the prediction of the context control theme is effectively utilized, the context emotion label is accurately predicted, and the words related to the emotion reason are accurately identified, so that the emotional co-emotional property, theme consistency and content correlation of the reply generated by the model are improved.

Description

technical field [0001] The invention relates to the fields of natural language processing, deep learning, text sentiment analysis, and dialogue generation, and in particular to an empathy dialogue generation method that integrates topic prediction and emotional reasoning. Background technique [0002] The scenes of human-computer dialogue are full of people's lives. As the core field of artificial intelligence, human-computer dialogue has always been widely concerned by all walks of life. People have always hoped to create robots that can communicate smoothly with humans. As early as 1950, Turing clearly proposed a standard for judging whether a robot is intelligent, that is, the Turing test. Inspired by the Turing test, human-computer dialogue has attracted the attention of scholars. With the development of deep learning technology, artificial intelligence has ushered in the third wave. A large amount of dialogue data on the Internet has become the foundation for human-comp...

AI Technical Summary

Problems solved by technology

In terms of dialogue topic prediction, previous studies only predict dialogue topics based on dialogue history, only considering the individual semantics of each topic, ignoring its specific dialogue context, which may lead to inaccurate topic representation and affect response coherence ; In terms of emotion recognition, the current methods mainly focus on learning a model to predict the emotional label, but ignore the detection of the reasons behind the emotion, which will cause the model to only use the surface emotional information

Images