Interpretation of natural communication

Patent metadata

Field	Value
Publication number	US-9653067-B2
Application number	US-201414469365-A
Country	US
Kind code	B2
Filing date	Aug 26, 2014
Priority date	Mar 11, 2014
Publication date	May 16, 2017
Grant date	May 16, 2017

Abstract

A computer-implemented method, including receiving by one or more computer systems input information that represents a multi-dimensional communication; detecting, based on contents of the input information, a plurality of communication inputs; applying one or more weighted values to one or more of the communication inputs; assigning, based on application of the one or more weighted values, confidence levels to the communications inputs; determining which of the confidence levels are below a confidence threshold; executing one or more disambiguation rules to disambiguate the communication inputs with confidence levels below the confidence threshold; and generating a communication instruction to perform an action that is specified by the multi-dimensional communication.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
receiving by one or more computer systems input information that represents a multidimensional communication that comprises at least two different communication domains;
detecting, by the one or more computer systems based on contents of the input information, at least two different communication inputs, corresponding to the at least two different communication domains;
determining, by the one or more computer systems, corresponding weighted values for each to one or more of the at least two different communication inputs, with a weighted value based on a communication domain weight value divided by a time interval between receiving a first one of the at least two different communication inputs and a subsequent one of the at least two different communication inputs;
assigning, by the one or more computer systems, confidence levels to the at least two different communications inputs, with the confidence levels being based on the weighted values;
determining, by the one or more computer systems, which of the confidence levels are below a confidence threshold;
executing, by the one or more computer systems, one or more disambiguation rules to disambiguate between the at least two different communication inputs when the assigned confidence levels are below the confidence threshold; and
generating, by the one or more computer systems, a communication instruction to perform an action that is specified by the multi-dimensional communication.

2. The computer-implemented method of claim 1 , further comprising:
determining, based on execution of the one or more disambiguation rules, a meaning of the multi-dimensional communication.

3. The computer-implemented method of claim 2 , wherein the generated communication instruction is based on the determined meaning.

4. The computer-implemented method of claim 1 , wherein the first one of the two different domains is a speech domain and detecting further comprises:
detecting the speech input;
subsequently, detecting a second one of the at least two communication inputs; and
determining whether the speech input is free of conflicts with the second one of the at least two communication inputs.

5. The computer-implemented method of claim 4 wherein when there is a conflict, further comprising:
transmitting, to a client device, an audio prompt that requests additional information to promote disambiguation of at least first and second ones of the communication inputs.

6. The computer-implemented method of claim 1 , further comprising:
transmitting the communication instruction to a networked device for execution of the communication instruction.

7. The computer-implemented method of claim 1 , wherein the at least two different communication domains are selected from speech, facial gesture, eye gaze, physical motion, and biometric measurement.

8. A computer program product tangibly stored on a computer readable hardware storage device, the computer program product comprising instructions for causing one or more processing devices to:
receive input information that represents a multi-dimensional communication that comprises at least two different communication domains;
detect based on contents of the input information, at least two different communication inputs, corresponding to the at least two different communication domains;
determine one or more corresponding weighted values for each of the at least two different communication inputs, with a weighted value based on a communication domain weight value divided by a time interval between receiving a first one of the at least two different communication inputs and a subsequent one of the at least two different communication inputs;
assign confidence levels to the at least two different communications inputs, with the confidence levels being based on the weighted values;
determine which of the confidence levels are below a confidence threshold;
execute one or more disambiguation rules to disambiguate between the at least two different communication inputs when the assigned confidence levels are below the confidence threshold; and
generate a communication instruction to perform an action that is specified by the multidimensional communication.

9. The computer program product of claim 8 , further comprising instructions for causing the one or more processing devices to:
determine, based on execution of the one or more disambiguation rules, a meaning of the multi-dimensional communication.

10. The computer program product of claim 9 , wherein the generated communication instruction is based on the determined meaning.

11. The computer program product of claim 8 , wherein the first one of the two different domains is a speech domain and the instructions to detect further comprise instructions to:
detect the speech input;
subsequently, detect a second one of the at least two communication inputs; and
determine whether the speech input first one of the at least two communication inputs is free of conflicts with the second one of the at least two communication inputs.

12. The computer program product of claim 11 , further comprising instructions for causing the one or more processing devices to:
transmit, to a client device, an audio prompt that requests additional information to promote disambiguation of at least first and second ones of the communication inputs.

13. The computer program product of claim 8 , further comprising instructions for causing the one or more processing devices to:
transmit the communication instruction to a networked device for execution of the communication instruction.

14. The computer program product of claim 8 , wherein the at least two different communication domains are selected from more of speech, facial gesture, eye gaze, physical motion, and biometric measurement.

15. A system comprising:
one or more processing devices; and
a computer program product tangibly stored on a computer readable hardware storage device, the computer program product comprising instructions for causing the one or more processing devices to:
receive input information that represents a multi-dimensional communication, that comprises at least two different communication domains;
detect based on contents of the input information, at least two different communication inputs, corresponding to the at least two different communication domains;
determine one or more corresponding weighted values for each of the at least two different communication inputs, with a weighted value based on a communication domain weight value divided by a time interval between receiving a first one of the at least two different communication inputs and a subsequent one of the at least two different communication inputs;
assign confidence levels to the at least two different communications inputs, with the confidence levels being based on the weighted values;
determine which of the confidence levels are below a confidence threshold;
execute one or more disambiguation rules to disambiguate between the at least two different communication inputs when the assigned confidence levels are below the confidence threshold; and
generate a communication instruction to perform an action that is specified by the multi-dimensional communication.

16. The system of claim 15 , further comprising instructions for causing the one or more processing devices to:
determine, based on execution of the one or more disambiguation rules, a meaning of the multi-dimensional communication.

17. The system of claim 16 , wherein the generated

Assignees

• Fmr Llc

Inventors

• Brooks Seth Warren
• Raikula Ronald M

Classifications

• G10L15/1815 Primary

  Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning · CPC title

• G10L15/02 Primary

  Feature extraction for speech recognition; Selection of recognition unit · CPC title

• G06K9/00268

  Physics · mapped topic

• G06K9/0061

  Physics · mapped topic

• G06K9/00355

  Physics · mapped topic