Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method and device for dynamically loading language model based on user input scene

A language model and input scene technology, applied in natural language data processing, special data processing applications, instruments, etc., can solve the problem that the Ngram model is difficult to handle multiple typical input scene input requirements at the same time, and the Ngram model does not have automatic learning and cannot satisfy users And other issues

Active Publication Date: 2014-02-12
TENCENT TECH (SHENZHEN) CO LTD
View PDF0 Cites 22 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, a single Ngram model is difficult to handle the input requirements of multiple typical input scenarios at the same time, and cannot meet the needs of all users
On the other hand, the Ngram model itself has no automatic learning mechanism, and cannot learn and adapt to each user's input habits

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 dynamically loading language model based on user input scene
  • Method and device for dynamically loading language model based on user input scene
  • Method and device for dynamically loading language model based on user input scene

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] The technical solutions of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments.

[0025] Aiming at the shortcomings of the commonly used language models at present, the present invention proposes a scene-based language modeling technology. In the language model training phase, in addition to the normal training of a single general language model, scene language models are also trained for several typical input scenarios. In the actual use of the input method, the input method software dynamically judges the user's current input scene, dynamically loads the corresponding scene language model, and combines with the general language model to jointly process the user's input. This method can effectively improve the processing ability of human-computer interaction software such as voice recognition and handwritten character recognition, especially input method software, for user input requirements in var...

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 relates to a method and a device for dynamically loading a language model based on a user input scene. The method comprises the following steps: judging a current user input scene; selecting a corresponding scene language model according to the user input scene; fusing the scene language model with a universal language model into a hybrid language model; loading the hybrid language model to process the user input, wherein the scene language model and the universal language model are obtained through training. The user input scene is dynamically judged in a user input process by training the scene language model; the user input is treated together by loading the corresponding scene language model and the universal language model; human-computer interaction software is improved, particularly the processing capacity of input method software on user input requirements under a typical scene is improved.

Description

technical field [0001] The invention relates to a method and device for dynamically loading a language model based on a user input scene. Background technique [0002] With the continuous improvement of computer hardware performance and software intelligence, people increasingly expect computers to provide a more natural way of human-computer interaction, which is reflected in: (1) providing more intelligent Chinese input methods; (2) Provide more accurate speech recognition function; (3) Provide continuous handwritten character recognition function. The realization of these three interaction methods requires the support of language modeling technology at the bottom layer. The pros and cons of language modeling technology directly determines the performance of the language model, and also determines the quality of the above-mentioned human-computer interaction software. [0003] The statistical language model is based on probability theory and mathematical statistics theor...

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
IPC IPC(8): G06F17/20G06F40/00
Inventor 肖镜辉李鑫刘廷超汤利华
Owner TENCENT TECH (SHENZHEN) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap 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