Systems and methods for training a control system based on prior audio inputs

Patent metadata

Field	Value
Publication number	US-11087740-B2
Application number	US-201816201679-A
Country	US
Kind code	B2
Filing date	Nov 27, 2018
Priority date	Nov 27, 2018
Publication date	Aug 10, 2021
Grant date	Aug 10, 2021

Abstract

Systems and methods are disclosed herein for training a control system based on prior audio inputs. The disclosed systems and methods receive a non-lexical or interjectional audio input. State change indications are also received and stored by the system within a predefined period of time starting from the time the system received the audio input. The system then receives a subsequent audio input. If the audio inputs of both the audio input and the subsequent audio input match, and contextual information for the audio input and the subsequent audio input match, the system stores a match association, comprising a confidence factor, for the subsequent audio input to the audio input in the associative data structure. If the confidence factor is greater than a preconfigured confidence level, the system executes one or more functions based on stored state change indications.

First claim

What is claimed is:

1. A method for training a control system based on prior audio inputs, the method comprising:
receiving, by control circuitry, an audio input, wherein the audio input comprises a non-lexical sound or an interjection;
receiving, by the control circuitry, one or more state change indications within a predefined period of time starting from the receipt of the audio input; and
in response to receiving one or more state change indications within the predefined period of time:
determining, by the control circuitry, contextual information for the audio input at the time of the audio input; and
storing, by the control circuitry, the audio input, the contextual information, and the one or more state change indications in an associative data structure in a memory.

2. The method of claim 1 further comprising:
receiving, by the control circuitry, a subsequent audio input, wherein the subsequent audio input comprises a non-lexical sound or an interjection;
determining, by the control circuitry, contextual information for the subsequent audio input at the time of the subsequent audio input; and
in response to determining, by the control circuitry, (a) the subsequent audio input matches the stored audio input and (b) the contextual information for the subsequent audio input matches the stored contextual information of the audio input:
storing, by the control circuitry, a match association for the subsequent audio input to the audio input in the associative data structure, wherein the match association comprises a confidence factor of the match association; and
adjusting, by the control circuitry, the confidence factor of the match association.

3. The method of claim 2 , wherein in response to determining, by the control circuitry, (a) the subsequent audio input matches the stored audio input and (b) the contextual information for the subsequent audio input matches the stored contextual information of the audio input, further comprises:
determining, by the control circuitry, whether the confidence factor of the match is greater than a preconfigured confidence level; and
in response to determining the confidence factor of the match is greater than a preconfigured confidence level, executing, by the control circuitry, one or more functions based on the stored one or more state change indications.

4. The method of claim 3 , wherein:
executing, by the control circuitry, one or more functions comprises providing a confirmatory function;
and the method further comprises:
receiving, by the control circuitry, a selection of the confirmatory function; and
storing, by the control circuitry, the confirmatory function, the subsequent audio input, and the contextual information for the subsequent audio input in the associative data structure in the memory.

5. The method of claim 3 , wherein:
executing, by the control circuitry, one or more functions comprises transmitting an instruction to a network-connected device, the instruction based on the stored one or more state change indications;
and the method further comprises storing, by the control circuitry, the one or more functions, the subsequent audio input, and the contextual information for the subsequent audio input in the associative data structure in the memory.

6. The method of claim 1 , wherein determining, by the control circuitry, contextual information for the audio input, further comprises:
parsing, by the control circuitry, the audio input into one or more audio components;
comparing, by the control circuitry, each component, of the one or more audio components, to a plurality of audio signal profiles, wherein the audio signal profiles comprise corresponding contextual information;
determining, by the control circuitry, whether one of the one or more components matches one of the plurality of audio signal profiles; and
in response to the determination that one of the one or more components matches one of the plurality of audio signal profiles, assigning, by the control circuitry, the corresponding contextual information of the matching audio signal profile to the contextual information of the audio input.

7. The method of claim 1 , wherein:
the audio input further comprises metadata including corresponding contextual information of the audio input; and
determining, by the control circuitry, contextual information for the audio input further comprises assigning the corresponding contextual information of the audio input as the contextual information for the audio input.

8. The method of claim 1 , wherein determining, by the control circuitry, contextual information for the audio input further comprises:
receiving, by the control circuitry, contextual information associated with the audio input from a contextual data structure; and
assigning, by the control circuitry, the received contextual information to the audio input as the contextual information for the audio input.

9. The method of claim 2 , wherein adjusting, by the control circuitry, the confidence factor of the match association further comprises:
determining, by the control circuitry, statistical information associated with the match association; and
calculating, by the control circuitry, the confidence factor by applying a mathematical operation to a previous confidence factor associated with the match association, wherein the mathematical operation utilizes the statistical information of the match association.

10. The method of claim 1 wherein receiving, by control circuitry, an audio input, wherein the audio input comprises a non-lexical sound or an interjection further comprises:
determining whether the audio input is a non-lexical sound or an interjection by:
comparing, by the control circuitry, the received audio input to a lexical sound data structure;
determining, by the control circuitry, if the audio input matches one of a plurality of lexical sounds within the lexical sound data structure; and
in response to the determination that the audio input does not match one of a plurality of lexical sounds within the lexical sound data structure, determining, by the control circuitry, that the received audio input is a non-lexical sound or an interjection.

11. A system for training a control system based on prior audio inputs, comprising:
control circuitry configured to:
receive an audio input, wherein the audio input comprises a non-lexical sound or an interjection;
receive one or more state change indications within a predefined period of time starting from the receipt of the audio input; and
in response to receiving one or more state change indications within the predefined period of time:
determine contextual information for the audio input at the time of the audio input; and
store the audio input, the contextual information, and the one or more state change indications in an associative data structure in a memory.

12. The system of claim 11 , wherein the control circuitry is further configured to:
receive a subsequent audio input, wherein the subsequent audio input comprises a non-lexical sound or an interjection;
determine contextual information for the subsequent audio input at the time of the subsequent audio input; and
in response to a determination that (a) the subsequent audio input matches the stored audio input and (b) the contextual information for the subsequent audio input matches the stored contextual information of the audio input:
store a match association for the subsequent audio input to the audio input in the associative data structure, wherein the match association comprises a confidence factor of the match association; and
adjust the confidence factor of the match asso

Assignees

• Rovi Guides Inc

Inventors

• James Bryan
• Malhotra Manik

Classifications

• G10L15/063 Primary

  Training · CPC title

• G06N20/00

  Machine learning · CPC title

• G10L15/02

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

• G10L15/22 Primary

  Procedures used during a speech recognition process, e.g. man-machine dialogue · CPC title

• G10L2015/223

  Execution procedure of a spoken command · CPC title