Method and device for domain adaptive speech recognition
A speech recognition and self-adaptive technology, applied in speech recognition, speech analysis, instruments, etc., can solve the problems of low efficiency and accuracy
Active Publication Date: 2020-06-16
北京如布科技有限公司
View PDF6 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0003] The purpose of the present invention is to provide a field adaptive speech recognition method, which can overcome the problems of low efficiency and accuracy when the existing speech recognition methods are applied to different fields
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0110] Preferred embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although preferred embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
[0111] figure 1 A flow chart showing a domain adaptive speech recognition method according to an embodiment of the present invention, such as figure 1 As shown, the method specifically includes the following steps:
[0112] Step 1: Train based on the initial audio corpus, establish an initial acoustic model, and train based on the initial text corpus, establish an initial language model
[0113] When implementing the domain-adaptive speec...
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
Login to View More Abstract
The invention relates to a field adaptive voice recognition method and device. The method comprises steps that training is carried out based on initial audio corpus data, an initial acoustics model is established, training is carried out based on initial text corpus data, and an initial language model is established; industry field corpus data collection and processing are carried out, training is carried out based on the industry field corpus data after processing, and an industry field language model is established; based on the initial language model and the industry field language model, matching calculation is carried out, and a hybrid language model is established; phoneme clustering is carried out based on the industry field corpus data after processing, training is carried out according to the phoneme clustering result, and a clustering acoustics model is established; the initial acoustics model and the clustering acoustics model are fused, and a fusion acoustics model is established; an inputted voice signal is received, and a word sequence corresponding to the voice signal is determined based on the fusion acoustics model and the hybrid language model.
Description
technical field [0001] The present disclosure relates to the field of speech recognition, in particular to a method and device for field-adaptive speech recognition. Background technique [0002] Speech recognition is the basis for realizing artificial intelligence. Many robots, Internet of Things, and mobile devices use voice as an interactive portal. Due to the diversification of industries and spoken languages, the industry recordings and annotation data required for speech recognition are relatively scarce, resulting in low recognition accuracy when speech recognition systems are applied in different fields. Existing speech recognition systems are usually only applicable to a specific field or application scenario, such as only in the entertainment field, or only in general small talk scenarios. When changing fields or application scenarios, it is necessary to reselect audio data and retrain the speech recognition model based on the reselected audio data. Such a speech...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G10L15/06G10L15/183
CPCG10L15/063G10L15/183G10L2015/0631
Inventor 郭祥郭瑞雷宇
Owner 北京如布科技有限公司