FIELD: computer technology for speech recognition.
SUBSTANCE: technical result consists in increasing the accuracy of recognizing the terms of a specific area. The technical result is achieved by forming a network for decoding a specific area based on a language model of a specific area and a universal language model; and combining the area-specific decoding network with the common decoding network to obtain the target decoding network, wherein the formation of the area-specific decoding network based on the area-specific language model and the universal language model includes: performing interpolation on the area-specific language model and the area-specific language model, wherein part interpolated on includes all parts in the domain language model and a part in the universal language model that also occurs in the domain language model; and generating a region-specific decoding network based on the part on which the interpolation has been performed.
EFFECT: accuracy of recognizing the terms of a specific area.
13 cl, 9 dwg
Title | Year | Author | Number |
---|---|---|---|
UNIVERSAL ORTHOGRAPHIC SYMBOLIC CIRCUITS | 2005 |
|
RU2441287C2 |
METHOD AND DEVICE FOR PROVIDING A TEXT MESSAGE | 2004 |
|
RU2320082C2 |
SYNCHRONOUS UNDERSTANDING OF SEMANTIC OBJECTS REALISED BY MEANS OF TAGS OF SPEECH APPLICATION | 2004 |
|
RU2349969C2 |
SYNCHRONOUS COMPREHENSION OF SEMANTIC OBJECTS FOR HIGHLY ACTIVE INTERFACE | 2004 |
|
RU2352979C2 |
ANAPHORA RESOLUTION BASED ON A DEEP ANALYSIS TECHNOLOGY | 2015 |
|
RU2601166C2 |
METHOD AND SYSTEM FOR GENERATING TEXT | 2023 |
|
RU2817524C1 |
METHOD AND SYSTEM FOR DIGITAL ASSISTANT TEXT GENERATION | 2022 |
|
RU2796208C1 |
FLEXIBLE CIRCUIT FOR ADJUSTING LANGUAGE MODEL | 2015 |
|
RU2689203C2 |
LANGUAGE-DEPENDENT POSITIONING AND SIGNALLING | 2011 |
|
RU2587990C2 |
OPTICAL CHARACTER RECOGNITION BY MEANS OF COMBINATION OF NEURAL NETWORK MODELS | 2020 |
|
RU2768211C1 |
Authors
Dates
2023-05-16—Published
2019-12-12—Filed