Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Method and device for processing voice texts

A technology of speech text and rules, applied in the field of information processing, can solve problems such as inapplicable language environment, inability to fully play its role, insufficient optimization of speech and text processing methods, etc., to achieve the effect of expanding the scope of application and optimizing the processing method

Active Publication Date: 2015-05-20
TENCENT TECH (SHENZHEN) CO LTD
View PDF8 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Regular rules are not flexible enough to apply to actual locales
Named entity rules cannot fully play a role in relatively fixed language environments and scenarios where it is not suitable to establish an entity naming library
Therefore, the two methods of processing speech and text provided by the prior art have certain limitations, resulting in insufficient optimization of the processing of speech and text

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
  • Method and device for processing voice texts
  • Method and device for processing voice texts
  • Method and device for processing voice texts

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0027] The embodiment of the present invention provides a method for processing voice text, see figure 1 , the method flow includes the following steps:

[0028] 101: Perform named entity mapping on the speech text to obtain a first mapping result;

[0029] 102: Perform vocabulary mapping on the first mapping result to obtain a second mapping result;

[0030] Further, before the vocabulary mapping is performed on the first mapping result, it includes:

[0031] Expanding one or more named entities in the first mapping result into corresponding phonetic texts before mapping to obtain at least two third mapping results;

[0032] Vocabulary mapping is performed on the first mapping result, including:

[0033] Vocabulary mapping is performed on the speech text that is not mapped as a named entity in each third mapping result to obtain a second mapping result.

[0034] 103: Match the second mapping result with preset rules including regular rules, and if a matching rule is obtai...

Embodiment 2

[0056] The embodiment of the present invention provides a method for processing speech text, combined with the content of the first embodiment above, see figure 2 , the method flow includes:

[0057] 201: Perform named entity mapping on the speech text to obtain the first mapping result;

[0058]Specifically, performing named entity mapping on the speech text, including but not limited to: establishing a named entity library; searching the phonetic text that can be recognized as a named entity in the named entity library in the speech text, and replacing the found phonetic text with the named entity . It should be noted that named entities are collected from a large amount of information on the Internet, and the identification of named entities is realized by using independent dictionary trees in various fields, that is, it can support finding all named entities when all or part of the named entities overlap. named entity.

[0059] For ease of understanding, take the voice...

Embodiment 3

[0100] see image 3 , an embodiment of the present invention provides a device for processing speech and text, the device comprising:

[0101] The first mapping module 301 is configured to perform named entity mapping on the speech text to obtain a first mapping result;

[0102] The second mapping module 302 is configured to perform vocabulary mapping on the first mapping result to obtain a second mapping result;

[0103] A matching module 303, configured to match the second mapping result with preset rules including regular rules;

[0104] The first processing module 304 is configured to, when a matching rule is obtained, process the speech text according to the obtained matching rule.

[0105] As a preferred embodiment, see Figure 4 , the device also includes:

[0106] An expansion module 305, configured to sequentially expand one or more named entities in the first mapping result into corresponding phonetic texts before mapping to obtain at least two third mapping resu...

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

A method and apparatus for processing speech texts belong to the technical field of information processing. The method comprises: performing named entity mapping on a speech text to obtain a first mapping result (101); performing vocabulary mapping on the first mapping result to obtain a second mapping result (102); matching the second mapping result with a preset rule, and processing the speech text according to a obtained matching rule, wherein the preset rule comprises a canonical rule (103). The configuration formats of the canonical rule and a named entity rule are unified, so that the scope of application of speech text processing technologies is extended, and therefore, the processing modes of speech texts are optimized.

Description

technical field [0001] The invention relates to the technical field of information processing, in particular to a method and device for processing voice text. Background technique [0002] With the continuous development of information processing technology, human-computer interaction in natural language has become a reality. The key to realizing human-computer interaction is to accurately understand the natural language instructions issued by users and perform corresponding operations. After the user sends a natural language instruction, the instruction is converted into a voice text, and how to process the voice text has become a concern of people. [0003] There are two ways to process speech text in the prior art. The first method: perform vocabulary mapping on the speech text to obtain the mapping result; extract the position parameters in the mapping result through the rule card position to obtain the card position extraction result; extract the card position The res...

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): G06F17/27
CPCG06F16/685
Inventor 王飞徐浩褚攀韩贵平廖玲
Owner TENCENT TECH (SHENZHEN) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com