Method and system for dynamic creation of contexts
AssigneeNuance Communications, Inc.
InventorsGerhard Grobauer, Miklos Papai
Life of this patent
Estimating the term…
Abstract
A method and a system for a speech recognition system, comprising an electronic speech-based document is associated with a document template and comprises one or more sections of text recognized or transcribed from sections of speech. The sections of speech are transcribed by the speech recognition system into corresponding sections of text of the electronic speech based document. The method includes the steps of dynamically creating sub contexts and associating the sub context to sections of text of the document template.
Claims15
A method for operating on an electronic speech-based document with a speech recognition system, wherein the electronic speech-based document is associated with a document template and comprises at least one section of text recognized or transcribed from at least one section of speech, wherein the at least one section of text comprises a first section of text associated with a first template section of the document template, the method comprising acts of:
- dynamically creating and/or adapting, by the speech recognition system, at least one sub context from a basic context, wherein the at least one sub context comprises a first sub context comprising at least a first set of words and a first language model, and wherein the basic context comprises at least a second set of words and a second language model;
- associating the first sub context to the first template section of the document template;
- identifying a basic topic related to the electronic speech-based document;
- associating the basic context to the electronic speech-based document, wherein usage of the second set of words is related to the basic topic;
- identifying a sub topic as related to the first section of text of the electronic speech-based document based upon usage of the first set of words;
- and
- associating the first sub context to the first section of text of the electronic speech-based document.
A speech recognition system, to operate on an electronic speech-based document associated with a document template, wherein the electronic speech-based document comprises at least one section of text recognized or transcribed from at least one section of speech, wherein the at least one section of text comprises a first section of text associated with a first template section of the document template, the speech recognition system comprising:
- at least one processor programmed to:
- dynamically create and/or adapt at least one sub context from a basic context, wherein the at least one sub context comprises at least a first sub context comprising a first set of words and a first language model, and wherein the basic context comprises at least a second set of words and a second language model;
- associate the first sub context to the first template section of the document template;
- identify a basic topic related to the electronic speech-based document;
- associate the basic context to the electronic speech-based document, wherein usage of the second set of words is related to the basic topic;
- and
- identify a sub topic related to the first section of text of the electronic speech-based document based upon usage of the first set of words.
A non-transitory computer-readable medium having embodied thereon instructions that, when processed by a computer, perform a method for a speech recognition system to operate on an electronic speech-based document associated with a document template and comprises at least one section of text recognized or transcribed from at least one section of speech, wherein the at least one section of text comprises a first section of text associated with a first template section of the document template, the method comprising:
- dynamically creating and/or adapting at least one sub context from a basic context, wherein the at least one sub context comprises a first sub context comprising at least a first set of words and a first language model, and wherein the basic context comprises at least a second set of words and a second language model;
- associating the first sub context to the first template section of the document template;
- identifying a basic topic related to the electronic document;
- associating the basic context to the electronic speech-based document, wherein usage of the second set of words is related to the basic topic;
- and
- identifying a sub topic related to the first section of text of the electronic speech-based document based upon usage of the first set of words.
Dependencies are inferred from the claim text.