Method, device and equipment for constructing speech recognition model and storage medium

Abstract

The present invention relates to the field of artificial intelligence, and provides a method, a device and equipment for constructing a speech recognition model and a storage medium. The method comprises the following steps: acquiring a plurality of training speech samples; constructing the speech recognition model by using an independent convolution layer, a convolution residual layer, a full connection layer and an output layer; inputting speech training information to the speech recognition model, updating a weight value of neurons in the speech recognition model with the speech informationand a text label corresponding to the speech information through a natural language processing NLP technology, and then obtaining a target model; evaluating an error of the target model by L(S) = -ln[Pi]<(h(x),z) being an element of a set S> p(z|h(x))= -sigma<(h(x),z) being an element of a set S> ln p(z|h(x)); adjusting the weight value of the neurons in the target model until the error is less than a threshold value; setting the weight value of the neurons with the error less than the threshold value as an ideal weight value; deploying the target model and the ideal weight value on a client.The method of the present invention reduces influence of tone in the speech information on a predicted text and computation burden during recognition process in the speech recognition model.

Description

technical field [0001] The present application relates to the field of intelligent decision-making, and in particular to a method, device, device and storage medium for constructing a speech recognition model. Background technique [0002] Speech recognition is used to convert speech to text. With the continuous development of deep learning technology, the application range of speech recognition is becoming wider and wider. [0003] At present, deep neural networks (DNN) have become a research hotspot in the field of automatic speech recognition. Convolutional neural networks (CNN) and recurrent neural networks (RNN) have achieved good results in the creation of speech recognition models, and deep learning has become the mainstream solution for speech recognition. [0004] In a deep neural network, the depth of the network is often closely related to the correct rate of recognition, because the traditional deep neural network can extract low-level, middle-level and high-le...

AI Technical Summary

Problems solved by technology

However, the alignment process of speech training samples is cumbersome and complicated, and takes a lot of time

Images