Voice service method and system based on adversarial artificial intelligence network

Abstract

Based on a BP neural network and a GAN network, the invention provides a method for enabling computer voice to generate corresponding emotional changes according to emotional changes of users. A BP neural network is trained in two processes of information forward propagation and error back propagation. According to the voice service system based on the adversarial artificial intelligence network,the type of the voice emotional state input by the user is accurately output. The corresponding emotional voice is output through the generative adversarial process of the GAN network with the emotional state type as the reference. The voice service system based on the adversarial artificial intelligence network is designed through the method.

Description

technical field [0001] The present invention relates to the field of artificial intelligence, and more specifically relates to a voice service method and system based on an adversarial artificial intelligence network. Background technique [0002] At present, human-machine dialogue is becoming more and more common. For example, in the voice interaction function between various smart devices and humans or in automatic telephone customer service, smart devices can interact with users through voices automatically generated by computers. In the process, The computer generates semantic content according to the needs of the dialogue with the user. Through the semantic content, the voice segment corresponding to the semantic content is found from the voice library that stores the pre-recorded voice segments and combined to play, thereby realizing the man-machine dialogue. [0003] However, the technical means of automatically generating voices by computers cannot achieve the purpos...

AI Technical Summary

Problems solved by technology

[0003] However, the technical means of automatically generating voices by computers cannot achieve the purpose of emotional communication through man-machine dialogue. The voices automatically generated by computers are always characterized by uniform speech speed, gentle tone, and small volume fluctuations. The change of the computer will produce a corresponding emotional change, so that the user cannot experience the real and natural emotion of the computer voice in the communication, and the communication process does not feel natural enough.

Images