A Multi-round Emotional Dialogue Method Based on Deep Learning

Abstract

The invention discloses a multi-round emotional dialogue method based on deep learning, which performs word segmentation on the text information input by the user, and vectorizes the text through a pre-trained word vector model; uses the deep learning model to perform emotional processing on the text input by the user Analyze and analyze the dialogue topic and background; retrieve the most likely dialogue reply from the emotional corpus based on retrieval; based on the emotional category of the user dialogue, as well as the chat topic and background, use the generated confrontation network to generate a natural dialogue reply; according to Two different dialogue generation methods, select a dialogue with the most relevant dialogue emotion and theme background and user input to send to the user. The present invention uses a combination of retrieval-based and confrontation-generated dialogs to generate replies that are consistent with the user's dialog topics and emotions. It is beneficial to improve the quality of dialogue generated by the system, thereby increasing the number of dialogue rounds, and the dialogue reply of the chat robot is emotional, making the human-machine dialogue more harmonious.

Description

technical field [0001] The present invention relates to the technical field of human-computer interaction, in particular to a multi-round emotional dialogue method based on deep learning. Background technique [0002] Dialogue system is the core technology in the field of human-computer interaction and an important way to realize harmonious human-computer interaction, which has great research significance and application value. At present, dialogue systems are attracting more and more attention in various fields, and the continuous progress of deep learning technology has greatly promoted the development of dialogue systems. For dialogue systems, deep learning techniques can utilize large amounts of data to learn feature representation and response generation strategies, which only require a small amount of manual operations. Now that we have easy access to "big data" of conversations on the web, we may be able to learn how to respond, and how to respond to almost any input...

AI Technical Summary

Problems solved by technology

[0005] Current dialogue systems tend to only focus on a single round of dialogue, or at most two, because it is difficult to give the system the ability to plan long-term and conduct fluent, coherent, and meaningful multi-turn dialogues

At the same time, the existing dialogue system tends to generate blunt and cold replies, and cannot generate emotional dialogue replies, which makes users want to end the dialogue as soon as possible, and cannot achieve multiple rounds of dialogue

Images