Method and system for optimizing compressed speech recognition model

Abstract

According to the embodiment, the invention provides a compressed speech recognition model optimizing method. The method comprises the following steps: based on a speech recognition model before compressing, confirming a teacher model, and generating a student model based on the compressed speech recognition model and un-annotated speech data in a speech database; extracting annotated speech data sequences from the speech database, so that a training data set is obtained, implementing neural network forward propagation on the student model via the training data set, and conforming a first posterior probability of the student model; implementing forward-backward computing on the teacher model via the training data set, and confirming a second posterior probability of the teacher model; comparing the first posterior probability and the second posterior probability, and conforming errors between the student model and the teacher model; and implementing neural network backward propagation on the student model in accordance with the errors when the errors are not convergent, so that the student model is optimized. According to the embodiment, the invention also provides a compressed speech recognition model optimizing system. In the embodiment, the compressed model is optimized in accordance with a source model.

Description

technical field [0001] The invention relates to the field of speech recognition, in particular to a method and system for optimizing a compressed speech recognition model. Background technique [0002] Speech recognition is an artificial intelligence technology that allows devices to convert voice signals into corresponding text or commands. Through the deep learning of the speech recognition model, the accuracy of speech recognition has been greatly improved. [0003] Although deep learning guarantees the accuracy of speech recognition performance, the large number of parameters in its models takes up a lot of storage space. On the one hand, the speech recognition model of the large parameter type requires a lot of calculation and occupies a lot of memory resources, and on the other hand, the speech recognition model of the large parameter type runs slowly. These factors hinder the deployment of this large parameter type of speech recognition models on resource-constraine...

AI Technical Summary

Problems solved by technology

[0007] In order to at least solve the problem that the compressed speech recognition model in the prior art usually cannot better retain the generalization ability and model accuracy of the pre-compression model

Images