Disambiguation of a spoken query term

• US 9,418,177 B1
• Filed: 08/05/2013
• Issued: 08/16/2016
• Est. Priority Date: 07/28/2010
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Lloyd, Matthew I., Schalkwyk, Johan, Risbood, Pankaj
• Primary Examiner(s): Cyr, Leonard Saint
• Application Number:
  US13/958,740
• Time in Patent Office: 1,107 Days
• Field of Search: 704/246, 704/247, 704/251, 704/252, 704270-271
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/90332 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing spoken query terms. In one aspect, a method includes performing speech recognition on an audio signal to select two or more textual, candidate transcriptions that match a spoken query term, and to establish a speech recognition confidence value for each candidate transcription, obtaining a search history for a user who spoke the spoken query term, where the search history references one or more past search queries that have been submitted by the user, generating one or more n-grams from each candidate transcription, where each n-gram is a subsequence of n phonemes, syllables, letters, characters, words or terms from a respective candidate transcription, and determining, for each n-gram, a frequency with which the n-gram occurs in the past search queries, and a weighting value that is based on the respective frequency.

Claims

1. A computer-implemented method comprising:

generating, by a server-based automated speech recognizer that is associated with a search engine, a candidate transcription of an utterance that was input through a computing device;

accessing a collection of past search queries that were previously submitted by a user of the computing device;

determining that at least a portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device;

selecting the candidate transcription of the utterance based at least on determining that at least the portion of the candidate transcription matches the query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device;

generating a search query that includes the selected candidate transcription as a query term; and

submitting the search query to the search engine.

2. The method of claim 1, comprising:

after selecting the candidate transcription of the utterance, performing a search using the particular search query of the collection of past search queries.

3. The method of claim 1, wherein determining that at least the portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device comprises determining that one or more portions of the candidate transcription are identical to one or more portions of the query term.

4. The method of claim 1, wherein determining that at least the portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device comprises:

determining a number of search queries that include the query term that matches at least the portion of the candidate transcription,

wherein the candidate transcription of the utterance is selected further based on the number of search queries that include the query term that matches at least the portion of the candidate transcription.

5. The method of claim 1, comprising:

obtaining a speech recognition confidence value associated with the candidate transcription of the utterance,

wherein the candidate transcription of the utterance is selected further based on the speech recognition confidence value associated with the candidate transcription of the utterance.

6. The method of claim 1, comprising:

obtaining a second candidate transcription of the utterance; and

determining that no portion of the second candidate transcription of the utterance matches a query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device,

wherein the candidate transcription of the utterance is selected further based on determining that no portion of the second candidate transcription of the utterance matches a query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device.

7. The method of claim 1, comprising:

determining a temporal context, geographic context, dock context, or device context associated with the utterance; and

determining that the temporal context, geographic context, dock context, or device context associated with the candidate utterance matches a temporal context, geographic context, dock context, or device context associated with a search query whose query term matches at least the portion of the candidate transcription,

wherein the candidate transcription of the utterance is selected further based on determining that the temporal context, geographic context, dock context, or device context associated with the candidate utterance matches the temporal context, geographic context, dock context, or device context associated with the search query whose query term matches at least the portion of the candidate transcription.

8. A system comprising:

one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising;

generating, by a server-based automated speech recognizer that is associated with a search engine, a candidate transcription of an utterance that was input through a computing device;

accessing a collection of past search queries that were previously submitted by a user of the computing device;

determining that at least a portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device;

selecting the candidate transcription of the utterance based at least on determining that at least the portion of the candidate transcription matches the query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device;

generating a search query that includes the selected candidate transcription as a query term; and

submitting the search query to the search engine.

9. The system of claim 8, wherein the operations comprise:

after selecting the candidate transcription of the utterance, performing a search using the particular search query of the collection of past search queries.

10. The system of claim 8, wherein determining that at least the portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device comprises determining that one or more portions of the candidate transcription are identical to one or more portions of the query term.

11. The system of claim 8, determining that at least the portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device comprises:

determining a number of search queries that include the query term that matches at least the portion of the candidate transcription,

wherein the candidate transcription of the utterance is selected further based on the number of search queries that include the query term that matches at least the portion of the candidate transcription.

12. The system of claim 8, wherein the operations comprise:

obtaining a speech recognition confidence value associated with the candidate transcription of the utterance,

wherein the candidate transcription of the utterance is selected further based on the speech recognition confidence value associated with the candidate transcription of the utterance.

13. The system of claim 8, wherein the operations comprise:

obtaining a second candidate transcription of the utterance; and

determining that no portion of the second candidate transcription of the utterance matches a query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device,

wherein the candidate transcription of the utterance is selected further based on determining that no portion of the second candidate transcription of the utterance matches a query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device.

14. The system of claim 8, wherein the operations comprise:

determining a temporal context, geographic context, dock context, or device context associated with the utterance; and

determining that the temporal context, geographic context, dock context, or device context associated with the candidate utterance matches a temporal context, geographic context, dock context, or device context associated with a search query whose query term matches at least the portion of the candidate transcription,

wherein the candidate transcription of the utterance is selected further based on determining that the temporal context, geographic context, dock context, or device context associated with the candidate utterance matches the temporal context, geographic context, dock context, or device context associated with the search query whose query term matches at least the portion of the candidate transcription.

15. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

generating, by a server-based automated speech recognizer that is associated with a search engine, a candidate transcription of an utterance that was input through a computing device;

accessing a collection of past search queries that were previously submitted by a user of the computing device;

determining that at least a portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device;

selecting the candidate transcription of the utterance based at least on determining that at least the portion of the candidate transcription matches the query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device;

generating a search query that includes the selected candidate transcription as a query term; and

submitting the search query to the search engine.

16. The medium of claim 15, wherein the operations comprise:

after selecting the candidate transcription of the utterance, performing a search using the particular search query of the collection of past search queries.

17. The medium of claim 15, wherein determining that at least the portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device comprises determining that one or more portions of the candidate transcription are identical to one or more portions of the query term.

18. The medium of claim 15, determining that at least the portion of the candidate transcription matches a query term of a particular search query of the collection of past search queries that were previously submitted by the user of the computing device comprises:

determining a number of search queries that include the query term that matches at least the portion of the candidate transcription,

wherein the candidate transcription of the utterance is selected further based on the number of search queries that include the query term that matches at least the portion of the candidate transcription.

19. The medium of claim 15, wherein the operations comprise:

obtaining a speech recognition confidence value associated with the candidate transcription of the utterance,

wherein the candidate transcription of the utterance is selected further based on the speech recognition confidence value associated with the candidate transcription of the utterance.

20. The medium of claim 15, wherein the operations comprise:

obtaining a second candidate transcription of the utterance; and

determining that no portion of the second candidate transcription of the utterance matches a query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device,

wherein the candidate transcription of the utterance is selected further based on determining that no portion of the second candidate transcription of the utterance matches a query term of the particular search query of the collection of past search queries that were previously submitted by the user of the computing device.