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

Apparatus and method for verbal model switching and self-adapting

By selecting an appropriate language model when the application state changes and modifying the model using feedback text flow results, the problem of language model performance degradation when the input range changes is solved, enabling flexible switching and adaptation between multiple input ranges. , improving the performance and scalability of language models.

Inactive Publication Date: 2007-09-12
NEC (CHINA) CO LTD
View PDF0 Cites 47 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, the problem of relying on the input range when switching between different language models is solved, and the adaptation to a specific language model improves the performance of the language model

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
  • Apparatus and method for verbal model switching and self-adapting
  • Apparatus and method for verbal model switching and self-adapting
  • Apparatus and method for verbal model switching and self-adapting

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035]Fig. 1 shows a structural diagram of an apparatus for language model switching and adaptation according to the present invention. The device includes a CPU 01 for providing computing functions to all applications; an internal bus 02 through which the device exchanges data between internal memory 07 and permanent storage 04 (which can be hard disk and flash memory); general application programs 09, and the user interface 03, for example, may be a keyboard for key input and a microphone for voice input, and so on. The persistent storage 04 stores a language model set 05 and some other data 06 . The device also includes a target application program 14, a language model adaptation part 12, a language model switching part 11, a notification manager 13 and a language model related (LMB) engine 10.

[0036] The LMB engine 10 generally uses language models and provides language model-related services for applications. For example, the LMB engine 10 is a text input engine or a ...

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 is a device for the switching and the automatic adaptation of language model, including: Notification manager, when the application state changes, it will notify the language model switching part of the current state information of the purpose application or the request to a language model; The language model switching part, is used to choose one or more of the language mode switched from the language model set on the basis of the information of the current state or a request described; The engine based on the language model, for the use of decoding the user input with one or more selected language model; The auto-adaptive part of language model, for receiving the decoding results and amending one or more selected language model on the basis of the results in decoding. An approach is used for the switching and auto-adaptive operations of the language model. Thus increasing the input of language model, and improving the accuracy of user input.

Description

technical field [0001] The present invention generally relates to the field of natural language processing, and more specifically, the present invention relates to a device and method for language model switching and self-adaptation. Background technique [0002] Language model technology is a key component in natural language processing, which is widely used in many fields, such as speech recognition (SR), optical character recognition (OCR), predictive text input (PTE), usually for non-English languages ​​and full text Sentence text input, generally referred to as a sentence-level input method), etc. Typically, language models are used to estimate the probabilities of sentences. For example, in speech recognition, a voice recognizer is given a sequence of acoustic hypotheses that can generate different candidate sentences. Afterwards, each candidate sentence is scored by the language model, and the candidate sentence with the highest score will be considered the best can...

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): G06F17/20G06F40/00G10L15/065G10L15/18
CPCG06F17/275G10L15/183G10L2015/228G06F40/263
Inventor 吴根清许荔秦
Owner NEC (CHINA) 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