Granted Patent B2

US 12,347,417 · App. 18/387,892 · Granted Jul 1, 2025

Inventors: Ankur Aher (Maharashtra, IN); Jeffry Copps Robert Jose (Tamil Madu, IN)

Assignee: Adeia Guides Inc.

G10L15/02G10L15/22G10L2015/025G10L2015/223

Abstract

Claims (40)

1. A method of determining hint words for automated speech recognition, the method comprising:

retrieving, from a database, terms from speech input entered into a voice interface system;

determining, using processing circuitry, a set of terms that are most frequently occurring terms from the terms from the speech input entered into the voice interface system;

comparing each term of the set of terms to a graph of terms, wherein the graph of terms is a graph data structure comprising entertainment-related terms;

for each term of the set of terms that is present in the graph of terms, identifying a respective subset of proximate terms by selecting a predetermined number of terms proximate to the respective term that is present in the graph of terms;

comparing each respective subset of proximate terms to each other respective subset of proximate terms;

determining a set of hint words by selecting proximate terms that are common to each respective subset of proximate terms; and

transmitting, using the processing circuitry, the set of hint words to an automated speech recognition application.

2. The method of claim 1 , wherein the selecting the predetermined number of proximate terms further comprises, for each of the terms that is present in the graph of terms:

selecting one or more nearest connected terms to the term of the set of terms that is on the graph of terms, and adding the selected one or more nearest connected terms to the respective subset of proximate terms; and

successively selecting next-nearest connected terms to the term of the set of terms that is in the graph of terms, and adding the successively selected next-nearest connected terms to the respective subset of proximate terms until the number of terms in the respective subset of proximate terms is at least equal to the predetermined number.

3. The method of claim 1 , wherein determining the set of hint words by selecting terms common to the respective subsets of proximate terms further comprises:

selecting terms common to one subset of proximate terms less than every subset of proximate terms, and adding the terms common to one subset of proximate terms less than every subset of proximate terms to the set of hint words; and

successively selecting terms common to another subset of proximate terms less than every subset of proximate terms, and adding the terms common to another subset of proximate terms less than every subset of proximate terms to the set of hint words, until the number of terms in the set of hint words is at least equal to a second predetermined number.

4. The method of claim 1 , wherein the set of terms comprises at least one natural language word.

5. The method of claim 1 , wherein the set of terms comprises at least one natural language phrase.

6. The method of claim 1 , wherein the set of hint words includes phonemes of the terms.

7. The method of claim 1 , wherein the set of terms comprises phonetic neighbors of the terms.

8. The method of claim 1 , wherein a number of terms in the set of terms is different than the predetermined number of proximate terms.

9. A system for determining hint words for automated speech recognition, the system comprising:

a storage device; and

control circuitry configured to:

retrieve, from a database, terms from speech input entered into a voice interface system;

determine a set of terms that are most frequently occurring terms from the terms from the speech input entered into a voice interface system;

compare each term of the set of terms to a graph of terms, wherein the graph of terms is a graph data structure comprising entertainment-related terms;

for each term of the set of terms that is present in the graph of terms, identify a respective subset of proximate terms by selecting a predetermined number of terms proximate to the respective term that is present in the graph of terms;

compare each respective subset of proximate terms to each other respective subset of proximate terms;

determine a set of hint words by selecting proximate terms that are common to each respective subset of proximate terms; and

transmit the set of hint words to an automated speech recognition application.

10. The system of claim 9 , wherein the selecting the predetermined number of proximate terms further comprises, for each of the terms that is present in the graph of terms:

selecting one or more nearest connected terms to the term of the set of terms that is on the graph of terms, and adding the selected one or more nearest connected terms to the respective subset of proximate terms; and

successively selecting next-nearest connected terms to the term of the set of terms that is in the graph of terms, and adding the successively selected next-nearest connected terms to the respective subset of proximate terms until the number of terms in the respective subset of proximate terms is at least equal to the predetermined number.

11. The system of claim 9 , wherein the control circuitry is further configured to select terms common to the respective subsets of proximate terms by:

selecting terms common to one subset of proximate terms less than every subset of proximate terms, and adding the terms common to one subset of proximate terms less than every subset of proximate terms to the set of hint words; and

successively selecting terms common to another subset of proximate terms less than every subset of proximate terms, and adding the terms common to another subset of proximate terms less than every subset of proximate terms to the set of hint words, until the number of terms in the set of hint words is at least equal to a third predetermined number.

12. The system of claim 9 , wherein the set of terms comprises at least one natural language word.

13. The system of claim 9 , wherein the set of terms comprises at least one natural language phrase.

14. The system of claim 9 , wherein the set of hint words includes phonemes of the terms.

15. The system of claim 9 , wherein the set of terms comprises phonetic neighbors of the terms.

16. The system of claim 9 , wherein a number of terms in the set of terms is different than the predetermined number of proximate terms.

Assignments (2)

Continuity (3)

References Cited (31)