Voice identification method using long-short term memory model recurrent neural network

Abstract

The invention discloses a voice identification method using a long-short term memory model recurrent neural network. The voice identification method comprises training and identification. The training process comprises steps of introducing voice data and text data to generate a commonly-trained acoustic and language mode, and using an RNN sensor to perform decoding to form a model parameter. The identification process comprises steps of converting voice input to a frequency spectrum graph through Fourier conversion, using the recursion neural network of the long-short term memory model to perform orientational searching decoding and finally generating an identification result. The voice identification method adopts the recursion neural network (RNNs) and adopts connection time classification (CTC) to train RNNs through an end-to-end training method. These LSTM units combining with the long-short term memory have good effects and combines with multi-level expression to prove effective in a deep network; only one neural network model (end-to-end model) exits from a voice characteristic (an input end) to a character string (an output end) and the neural network can be directly trained by a target function which is a some kind of a proxy of WER, which avoids to cost useless work to optimize an individual target function.

Description

technical field [0001] The invention relates to the field of speech recognition, in particular to a speech recognition method using a long-short-term memory model recursive neural network. Background technique [0002] Speech recognition is often used in various smart devices, smart homes and other fields. So far, the error of speech recognition is relatively large, and the effect is not ideal. Recurrent neural network (RNNs) models can be used to model the relationship between two sequences. However, in traditional RNNs, there is a one-to-one correspondence between the labeled sequence and the input sequence. Not suitable for sequence modeling in speech recognition: the length of the recognized character sequence or phoneme sequence is much smaller than the input feature frame sequence. So it cannot be modeled directly with RNN. [0003] The present invention adopts recursive neural network (RNNs), and adopts connection time classification (CTC) to train RNNs through an ...

AI Technical Summary

Problems solved by technology

[0004] For the problem of network performance in speech recognition and the training target of each part of the traditional speech recognition system is inconsistent with the training target of the whole system, the object of the present invention is to provide a speech recognition method utilizing the long-short-term memory model recursive neural network, Model parameters can be obtained through training, and then used for speech and text data recognition

Images