Polyphone pronunciation prediction method and device and electronic equipment

Abstract

The invention discloses a polyphone pronunciation prediction method and device and electronic equipment, and relates to the technical field of x voice. According to the specific implementation scheme,semantic analysis is conducted on an input text through a multi-head self-attention language model to obtain text semantic information of the input text; and pronunciation of the polyphone in the input text is predicted according to the text semantic information. The accuracy of predicting the pronunciation of the polyphone can be improved.

Description

technical field [0001] The present application relates to the field of speech technology in computer technology, and in particular to a polyphone pronunciation prediction method, device and electronic equipment. Background technique [0002] At present, the pronunciation of multiple words is predicted mainly through the speech recurrent neural network (Recurrent Neural Network, RNN) or convolutional neural network (Convolutional Neural Networks, CNN). However, due to the lack of semantic information of many words used by RNN and CNN, the ability to represent longer contexts is weak, and RNN and CNN require feature engineering assistance and thus introduce error accumulation. It can be seen that the accuracy of predicting the pronunciation of polyphonic characters is relatively poor at present. Contents of the invention [0003] The present application provides a polyphonic pronunciation prediction method, device and electronic equipment to solve the problem of poor accura...

AI Technical Summary

Problems solved by technology

[0003] This application provides a method, device and electronic equipment for predicting the pronunciation of polyphonic characters to solve the problem of poor accuracy in predicting the pronunciation of polyphonic characters

Images