Speech recognition for keywords

• US 10,055,767 B2
• Filed: 05/13/2015
• Issued: 08/21/2018
• Est. Priority Date: 05/13/2015
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Aleksic, Petar, Moreno Mengibar, Pedro J.
• Primary Examiner(s): Reagan, James A
• Application Number:
  US14/710,928
• Publication Number:
  US 20160335677A1
• Time in Patent Office: 1,196 Days
• Field of Search: None
• CPC Class Codes: G06Q 30/0256 User search
• CPC Class Codes: G06Q 30/0275 Auctions
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/0638 Interactive procedures
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speech recognition are disclosed. In one aspect, a method includes receiving a candidate adword from an advertiser. The method further includes generating a score for the candidate adword based on a likelihood of a speech recognizer generating, based on an utterance of the candidate adword, a transcription that includes a word that is associated with an expected pronunciation of the candidate adword. The method further includes classifying, based at least on the score, the candidate adword as an appropriate adword for use in a bidding process for advertisements that are selected based on a transcription of a speech query or as not an appropriate adword for use in the bidding process for advertisements that are selected based on the transcription of the speech query.

Claims

1. A computer-implemented method comprising:

receiving, by a computing device and from an advertiser, a textual representation of a candidate adword for use in a bidding process for selecting advertisements based on a transcription of a speech query;

generating, by a text to speech module of the computing device, data that represents an audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

based on the data that represents the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, generating a transcription of the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

applying an acoustic model for a speech recognizer to the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

applying a language model for the speech recognizer to the transcription of the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

based on applying the acoustic model for the speech recognizer to the audible representation of the textual representation of the candidate adword and based on applying a language model for the speech recognizer to the transcription of the audible representation of the textual representation of the candidate adword, generating, for the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, a score that reflects a likelihood of the speech recognizer generating, in response to receiving an utterance of the candidate adword, a transcription that includes a word that is associated with the textual representation of the candidate adword; and

based at least on the score that reflects the likelihood of the speech recognizer generating, in response to receiving the utterance of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, the transcription that includes the word that is associated with the textual representation of the candidate adword, classifying the textual representation of the candidate adword as an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query or as not an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query.

2. The method of claim 1, wherein generating the score comprises:

applying the language model for the speech recognizer to the textual representation of the candidate adword.

3. The method of claim 1, comprising:

receiving, from the advertiser, a bid that is associated with the textual representation of the candidate adword.

4. The method of claim 1, comprising:

receiving, from the advertiser, a bid that is associated with a transcription of the audible representation of the textual representation of the candidate adword.

5. The method of claim 1, wherein receiving, from an advertiser, a textual representation of a candidate adword for use in a bidding process for selecting advertisements based on a transcription of a speech query comprises:

receiving an utterance spoken by the advertiser of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of a speech query; and

designating a transcription of the utterance spoken by the advertiser as the textual representation of the candidate adword for use in a bidding process for selecting advertisements based on a transcription of a speech query.

6. The method of claim 1, comprising:

receiving data indicating that the language model is updated;

based on the updated language model, generating an updated score for the textual representation of the candidate adword based on an updated likelihood of a speech recognizer generating, based on the utterance of the textual representation of the candidate adword, a transcription that includes a word that is associated with the textual representation of the candidate adword; and

classifying, based at least on the updated score, the textual representation of the candidate adword as an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query or as not an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query.

7. The method of claim 1, comprising:

determining one or more terms that are used in the bidding process and that are similar to the audible representation of the textual representation of the candidate adword;

providing, for display to the advertiser, the one or more terms; and

receiving, from the advertiser, a bid that is associated with one or more of the one or more terms.

8. The method of claim 7, wherein providing, for display to the advertiser, the one or more terms comprises:

providing misrecognition frequencies that each indicate a frequency that the expected pronunciation of the textual representation of the candidate adword is misrecognized as a respective term.

9. The method of claim 8, comprising:

providing data indicating whether an advertisement was displayed when the expected pronunciation of the textual representation of the candidate adword is misrecognized as the respective term.

10. 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;

receiving, by a computing device and from an advertiser, a textual representation of a candidate adword for use in a bidding process for selecting advertisements based on a transcription of a speech query;

generating, by a text to speech module of the computing device, data that represents an audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

based on the data that represents the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, generating a transcription of the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

applying an acoustic model for a speech recognizer to the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

applying a language model for the speech recognizer to the transcription of the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

based on applying the acoustic model for the speech recognizer to the audible representation of the textual representation of the candidate adword and based on applying a language model for the speech recognizer to the transcription of the audible representation of the textual representation of the candidate adword, generating, for the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, a score that reflects a likelihood of the speech recognizer generating, in response to receiving an utterance of the candidate adword, a transcription that includes a word that is associated with the textual representation of the candidate adword; and

based at least on the score that reflects the likelihood of the speech recognizer generating, in response to receiving the utterance of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, the transcription that includes the word that is associated with the textual representation of the candidate adword, classifying the textual representation of the candidate adword as an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query or as not an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query.

11. The system of claim 10, wherein generating the score comprises:

applying the language model for the speech recognizer to the textual representation of the candidate adword.

12. The system of claim 10, wherein the operations further comprise:

receiving, from the advertiser, a bid that is associated with the textual representation of the candidate adword.

13. The system of claim 10, wherein the operations further comprise:

receiving, from the advertiser, a bid that is associated with a transcription of the audible representation of the textual representation of the candidate adword.

14. The system of claim 10, wherein receiving, from an advertiser, a textual representation of a candidate adword for use in a bidding process for selecting advertisements based on a transcription of a speech query comprises:

receiving an utterance spoken by the advertiser of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of a speech query; and

designating a transcription of the utterance spoken by the advertiser as the textual representation of the candidate adword for use in a bidding process for selecting advertisements based on a transcription of a speech query.

15. The system of claim 10, wherein the operations further comprise:

determining one or more terms that are used in the bidding process and that are similar to the audible representation of the textual representation of the candidate adword;

providing, for display to the advertiser, the one or more terms; and

receiving, from the advertiser, a bid that is associated with one or more of the one or more terms.

16. 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:

receiving, by a computing device and from an advertiser, a textual representation of a candidate adword for use in a bidding process for selecting advertisements based on a transcription of a speech query;

generating, by a text to speech module of the computing device, data that represents an audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

based on the data that represents the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, generating a transcription of the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

applying an acoustic model for a speech recognizer to the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

applying a language model for the speech recognizer to the transcription of the audible representation of the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query;

based on applying the acoustic model for the speech recognizer to the audible representation of the textual representation of the candidate adword and based on applying a language model for the speech recognizer to the transcription of the audible representation of the textual representation of the candidate adword, generating, for the textual representation of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, a score that reflects a likelihood of the speech recognizer generating, in response to receiving an utterance of the candidate adword, a transcription that includes a word that is associated with the textual representation of the candidate adword; and

based at least on the score that reflects the likelihood of the speech recognizer generating, in response to receiving the utterance of the candidate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query, the transcription that includes the word that is associated with the textual representation of the candidate adword, classifying the textual representation of the candidate adword as an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query or as not an appropriate adword for use in the bidding process for selecting advertisements based on the transcription of the speech query.