Granted Patent B2

US 12,499,112 · App. 18/675,784 · Granted Dec 16, 2025

Inventor: Yongsung Kim (Mountain View, CA)

Assignee: GOOGLE LLC

G06F16/2454G06F16/2282G06F16/23G06F16/2365G06F16/24549G06F16/24575G06F16/29G06F16/3322G06F16/9535G06F16/9538H04N21/278H04N21/47202

Abstract

Claims (72)

1 . A method comprising:

determining, based on a voice query received in a search domain, a first voice recognition term and a second voice recognition term;

determining that at least a portion of the first voice recognition term corresponds to an entity name associated with the search domain;

determining, based on the entity name, a feasibility score for the first voice recognition term;

ranking, based on the feasibility score, the first voice recognition term relative to the second voice recognition term; and

based on a determination that a first portion of the voice query does not correspond to an action command term, executing the voice query in the search domain in accordance with a default action command and using the first voice recognition term, the first voice recognition term being selected for use over the second voice recognition term based on the ranking.

2 . The method of claim 1 , wherein the determining the first voice recognition term and the second voice recognition term includes segmenting the voice query to generate:

the first voice recognition term as the first portion of the voice query; and

the second voice recognition term as a second portion of the voice query.

3 . The method of claim 1 , wherein:

the determining the first voice recognition term and the second voice recognition term includes segmenting an entirety of the voice query to generate the first voice recognition term; and

the determining that at least the portion of the first voice recognition term corresponds to the entity name includes determining that the entirety of the voice query corresponds to the entity name.

4 . The method of claim 1 , further comprising:

determining, based on an additional voice query received in the search domain, a third voice recognition term and a fourth voice recognition term, the determining the third voice recognition term and the fourth voice recognition term includes:

determining that a first portion of the additional voice query corresponds to an additional action command term, and

generating a set of terms, including the third voice recognition term and the fourth voice recognition term, based on the first portion of the additional voice query; and

determining that at least a portion of the third voice recognition term corresponds to an additional entity name associated with the search domain by comparing the set of terms with entities in an entity table associated with action command terms.

5 . The method of claim 4 , the entity table includes curated entity information for a plurality of action command terms.

6 . The method of claim 1 , wherein the default action command is based on the search domain.

7 . The method of claim 1 , wherein:

the entity name to which at least the portion of the first voice recognition term corresponds is of an entity type; and

the default action command is based on the entity type.

8 . The method of claim 1 , wherein:

the entity name to which at least the portion of the first voice recognition term corresponds is of a first entity type;

the method further comprises determining that at least a portion of the second voice recognition term corresponds to an additional entity name of a second entity type different from the first entity type; and

the default action command is determined based on:

a first entity score for the entity name,

a second entity score for the additional entity name, and

at least one of the first entity type or the second entity type.

9 . The method of claim 1 , wherein the determining the feasibility score for the first voice recognition term includes applying a penalty score to the feasibility score, the penalty score based on how many terms associated with the first voice recognition term are recognized as entities in an entity table.

10 . The method of claim 1 , wherein the determining the feasibility score for the first voice recognition term includes weighting an average entity score for a plurality of terms that are associated with the first voice recognition term and recognized as entities in an entity table.

11 . The method of claim 1 , wherein the voice query is received from at least one of:

a media playback device having audio input capabilities; or

an audio input device included within a mobile device.

12 . The method of claim 1 , further comprising detecting a language in which the voice query is spoken;

wherein the determining the first voice recognition term and the second voice recognition term includes interpreting the voice query based on the detected language.

13 . A system comprising:

a memory storing instructions; and

a processor communicatively coupled to the memory and configured to execute the instructions to perform a process comprising:

determining, based on a voice query received in a search domain, a first voice recognition term and a second voice recognition term;

determining that at least a portion of the first voice recognition term corresponds to an entity name associated with the search domain;

determining, based on the entity name, a feasibility score for the first voice recognition term;

ranking, based on the feasibility score, the first voice recognition term relative to the second voice recognition term; and

based on a determination that a first portion of the voice query does not correspond to an action command term, executing the voice query in the search domain in accordance with a default action command and using the first voice recognition term, the first voice recognition term being selected for use over the second voice recognition term based on the ranking.

14 . The system of claim 13 , wherein the determining the first voice recognition term and the second voice recognition term includes segmenting the voice query to generate:

the first voice recognition term as the first portion of the voice query; and

the second voice recognition term as a second portion of the voice query.

15 . The system of claim 13 , wherein the process further comprises:

determining, based on an additional voice query received in the search domain, a third voice recognition term and a fourth voice recognition term, the determining the third voice recognition term and the fourth voice recognition term includes:

determining that a first portion of the additional voice query corresponds to an additional action command term, and

generating a set of terms, including the third voice recognition term and the fourth voice recognition term, based on the first portion of the additional voice query; and

determining that at least a portion of the third voice recognition term corresponds to an additional entity name associated with the search domain by comparing the set of terms with entities in an entity table associated with action command terms.

16 . The system of claim 13 , wherein the default action command is based on the search domain.

17 . The system of claim 13 , wherein:

the entity name to which at least the portion of the first voice recognition term corresponds is of an entity type; and

the default action command is based on the entity type.

18 . The system of claim 13 , wherein:

the entity name to which at least the portion of the first voice recognition term corresponds is of a first entity type;

the process further comprises determining that at least a portion of the second voice recognition term corresponds to an additional entity name of a second entity type different from the first entity type; and

the default action command is determined based on:

a first entity score for the entity name,

a second entity score for the additional entity name, and

at least one of the first entity type or the second entity type.

19 . A non-transitory computer-readable medium storing instructions that, when executed, cause a processor of a computing device to perform a process comprising:

determining, based on a voice query received in a search domain, a first voice recognition term and a second voice recognition term;

determining that at least a portion of the first voice recognition term corresponds to an entity name associated with the search domain;

determining, based on the entity name, a feasibility score for the first voice recognition term;

ranking, based on the feasibility score, the first voice recognition term relative to the second voice recognition term; and

based on a determination that a first portion of the voice query does not correspond to an action command term, executing the voice query in the search domain in accordance with a default action command and using the first voice recognition term, the first voice recognition term being selected for use over the second voice recognition term based on the ranking.

20 . The non-transitory computer-readable medium of claim 19 , wherein the determining the first voice recognition term and the second voice recognition term includes segmenting the voice query to generate:

the first voice recognition term as the first portion of the voice query; and

the second voice recognition term as a second portion of the voice query.

Assignments (2)

Continuity (5)

References Cited (58)