Unlock instant, AI-driven research and patent intelligence for your innovation.

Speech processing device, equipment and training method thereof

A technology of speech processing and training methods, which is applied in the computer field and can solve problems such as failure to achieve optimal performance, target object speech interference, and inconsistent loss functions.

Active Publication Date: 2020-06-09
TENCENT TECH (SHENZHEN) CO LTD
View PDF3 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, there are many problems in this separated system, for example: 1) the speech of the target object after the front-end enhancement will interfere, and such interference will greatly reduce the performance of the back-end recognition; 2) the loss function of the front-end enhancement system and the back-end The loss function of the recognition system is inconsistent, so that the front-end enhancement does not completely serve the back-end recognition, resulting in the failure of the entire multi-channel-based target object speech recognition system to achieve optimal performance

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Speech processing device, equipment and training method thereof
  • Speech processing device, equipment and training method thereof
  • Speech processing device, equipment and training method thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0030]In order to make the objects, technical solutions, and advantages of the present disclosure more apparent, exemplary embodiments according to the present disclosure will be described in detail below with reference to the accompanying drawings. Apparently, the described embodiments are only some of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure, and it should be understood that the present disclosure is not limited by the exemplary embodiments described here.

[0031] In this specification and the drawings, substantially the same or similar steps and elements are denoted by the same or similar reference numerals, and repeated descriptions of these steps and elements will be omitted. Meanwhile, in the description of the present disclosure, the terms "first", "second" and the like are only used to distinguish descriptions, and cannot be understood as indicating or implying relative importance or ranking.

[0032] For the...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a speech processing device, equipment and a training method thereof. The method comprises the steps: carrying out speech extraction of training speech input through a to-be-trained speech extraction model, so as to obtain a target object speech; performing speech recognition on the extracted target object speech by using a to-be-trained speech recognition model to obtain aspeech recognition result; and performing joint training on the to-be-trained speech extraction model and the to-be-trained speech recognition model based on the speech recognition result. The speechprocessing device is simple in structure and easy to expand, and the recognition performance and robustness of the speech processing model are greatly improved.

Description

technical field [0001] The present disclosure relates to the field of computer technology, in particular to an artificial intelligence-based speech processing device, equipment and training method thereof. Background technique [0002] With the rapid development of artificial intelligence technology, voice processing systems based on artificial intelligence have emerged. For example, based on the multi-channel target object speech recognition system, the whole system consists of two independent subsystems: 1) the front-end multi-channel target object speech enhancement system, which enhances the multi-channel noisy speech into a single-channel target object speech ; 2) the voice recognition system at the back end, which recognizes the enhanced single-channel target object voice into text. However, there are many problems in this separated system, for example: 1) the speech of the target object after the front-end enhancement will interfere, and such interference will greatl...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G10L15/06G10L15/02G10L15/16G10L15/20G10L21/0208
CPCG10L15/063G10L15/02G10L15/16G10L15/20G10L21/0208G10L2015/025
Inventor 吴渤于蒙陈联武温超苏丹俞栋
Owner TENCENT TECH (SHENZHEN) CO LTD