METHOD AND SYSTEM FOR DYNAMIC CREATION OF CONTEXTS

• US 20140303976A1
• Filed: 06/20/2014
• Published: 10/09/2014
• Est. Priority Date: 12/08/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Papai, Miklos, GROBAUER, GERHARD
• Granted Patent: US 9,412,370 B2
• Time in Patent Office: Days
• US Class Current: 704/235
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

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.

Claims

1. A method for a speech recognition system, comprising an electronic document, which is a speech based document comprising one or more sections of text recognized or transcribed from sections of speech, wherein said sections of speech are dictated by an author and processed by a speech recognizer in the speech recognition system into corresponding sections of text of said speech based document, said method comprising:

dynamically creating and/or adapting sub contexts by said speech recognizer and associating said sub context to said sections of text.

2. The method according to claim 1, wherein said method comprising:

identifying a basic topic related to said electronic document;

associating a basic context to said electronic document, said basic context comprising at least a set of words and the usage of said set of words related to said basic topic;

identifying sub topics, each related to a specific section of text in said electronic document;

creating said sub contexts from said basic context, said sub contexts comprising at least a specified set of words and the usage of said set of specified words related to said sub topic, respectively; and

associating each of said sub contexts to a matching section of text in said electronic document.

3. The method according to claim 2, wherein creating said sub contexts to each section of text and adapting said sub context in said electronic document is automatically performed.

4. The method according to claim 2, wherein creating said sub contexts from said basic context comprises:

identifying words related to said specific section of text differing from the remainder of said electronic document; and

creating said set of specified words by using said basic context.

8. A system according to claim 2, wherein means for creating sub contexts from said basic context comprising means for identifying words related to said specific section of text differing from the remainder of said electronic document;

and

means for creating said set of specified words by using said basic context.

5. A system for a speech recognition system, comprising an electronic document, which is a speech based document comprising one or more sections of text recognized or transcribed from sections of speech, wherein said sections of speech are dictated by an author and processed by a speech recognizer in the speech recognition system into corresponding sections of text of said speech based document, said system comprises:

means for creating and/or adapting sub contexts by said speech recognizer and means for associating said sub context to said sections of text.

6. The system according to claim 5, wherein said system comprising:

means for identifying a basic topic related to said electronic document;

means for associating a basic context to said electronic document, said basic context comprising at least a set of words and the usage of said set of words related to said basic topic;

means for identifying sub topics, each related to a specific section of text in said electronic document;

means for creating sub contexts from said basic context, said sub contexts comprising a least a specified set of words and the usage of said set of specified words related to said sub topic, respectively; and

means for associating each of said sub contexts to a matching section of text in said electronic document.

7. A system according to claim 6, wherein means for creating and/or adapting sub contexts to each section of text in said electronic document is automatic means.

9. A computer-readable medium having embodied thereon a computer program for processing by a computer, the computer program being configured for a speech recognition system, comprising an electronic document, which is a speech based document comprising one or more sections of text recognized or transcribed from sections of speech, wherein said sections of speech are dictated by an author and processed by a speech recognizer in the speech recognition system into corresponding sections of text of said speech based document, said computer program comprising:

a code segment for creating and adapting sub contexts in said speech recognizer and a code segment for associating said sub context to said sections of text.

10. The computer-readable medium according of claim 9, wherein said computer program comprises:

a code segment for determining a basic topic related to said electronic document;

a code segment for associating a basic context to said electronic document, said basic context comprising a set of words and the usage of said set of words related to said basic topic;

a code segment for identifying sub topics, each related to a specific section of text in said electronic document;

a code segment for creating sub contexts by using said basic context, said sub contexts comprising at least a specified set of words and the usage of said set of specified words related to said sub topic, respectively; and

a code segment for associating each of said sub context to a matching section of text in said electronic document.