Updating phonetic dictionaries

• US 9,135,912 B1
• Filed: 09/19/2012
• Issued: 09/15/2015
• Est. Priority Date: 08/15/2012
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Strope, Brian, Velikovich, Leonid
• Primary Examiner(s): BAKER, MATTHEW H
• Application Number:
  US13/622,547
• Time in Patent Office: 1,091 Days
• Field of Search: 704/9, 704/231, 704/235
• US Class Current: 1/1
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting
• CPC Class Codes: G10L 2015/0638 Interactive procedures

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for updating phonetic dictionaries. In one aspect, a method includes accessing a phonetic dictionary that identifies terms and one or more phonetic representations associated with each term, determining that a particular term that is identified in the phonetic dictionary is a spelling correction for another term that is identified in the phonetic dictionary, and storing, in the phonetic dictionary, one or more of the phonetic representations associated with the other term, with the particular term that is a spelling correction for the other term.

Claims

1. A computer-implemented method comprising:

identifying, from among a set of query terms, a particular query term that (i) does not occur in a lexicon of terms, and (ii) has no designated, canonical phonetic representation in a pronunciation phonetic dictionary, wherein a canonical phonetic representation comprises a sequence of phonemes;

generating a phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the pronunciation phonetic dictionary;

transmitting data identifying at least a portion of a term that does occur in the lexicon of terms and the particular query term to a spelling correction server;

receiving, from the spelling correction server, data that specifies a spelling correction confidence score, wherein the spelling correction confidence score reflects a probability that the term that does occur in the lexicon of terms is a correct spelling of the particular query term;

determining that the spelling correction confidence score satisfies a predetermined threshold; and

in response to determining that the spelling correction confidence score satisfies a predetermined threshold, designating, by one or more computing devices, the phonetic representation estimate for the particular query term as a canonical phonetic representation, in the phonetic dictionary, of the term that does occur in the lexicon of terms.

2. The method of claim 1, wherein generating a phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the phonetic pronunciation dictionary, comprises:

estimating a pronunciation of the particular query term based on one or more phonic rules for graphemes included in the particular query term.

3. The method of claim 1, wherein the phonetic representation estimate comprises a phonemic representation.

4. The method of claim 1, wherein the phonetic representation estimate comprises a sequence of phonemes.

5. The method of claim 1, wherein the term that does occur in the lexicon of terms comprises two or more words.

6. The method of claim 1, comprising:

after generating the phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the phonetic pronunciation dictionary, adding the particular query term to the lexicon of terms; and

after designating, by one or more computing devices, the phonetic representation estimate for the particular query term as the canonical phonetic representation, in the phonetic pronunciation dictionary, of the term that does occur in the lexicon of terms, deleting the particular query term from the lexicon of terms.

7. The method of claim 1, wherein designating, by one or more computing devices, the phonetic representation estimate for the particular query term as a canonical phonetic representation, in the phonetic pronunciation dictionary, of the term that does occur in the lexicon of terms comprises:

replacing one or more phonetic representations associated with the term that does occur in the lexicon of terms with the phonetic representation estimate.

8. The method of claim 1, comprising:

receiving a data representation of an utterance in which a user has spoken the term that does occur in the lexicon of terms; and

using the phonetic representation estimate for the particular query term, in outputting the term that does occur in the lexicon of terms as part of a transcription of the utterance.

9. 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;

identifying, from among a set of query terms, a particular query term that (i) does not occur in a lexicon of terms, and (ii) has no designated, canonical phonetic representation in a pronunciation phonetic dictionary, wherein a canonical phonetic representation comprises a sequence of phonemes;

generating a phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the pronunciation phonetic dictionary;

transmitting data identifying at least a portion of a term that does occur in the lexicon of terms and the particular query term to a spelling correction server;

receiving, from the spelling correction server, data that specifies a spelling correction confidence score, wherein the spelling correction confidence score reflects a probability that the term that does occur in the lexicon of terms is a correct spelling of the particular query term;

determining that the spelling correction confidence score satisfies a predetermined threshold; and

in response to determining that the spelling correction confidence score satisfies a predetermined threshold, designating, by one or more computing devices, the phonetic representation estimate for the particular query term as a canonical phonetic representation, in the phonetic dictionary, of the term that does occur in the lexicon of terms.

10. The system of claim 9, wherein generating a phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the phonetic pronunciation dictionary, comprises:

estimating a pronunciation of the particular query term based on one or more phonic rules for graphemes included in the particular query term.

11. The system of claim 9, the operations comprising:

after generating the phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the phonetic pronunciation dictionary, adding the particular query term to the lexicon of terms; and

after designating, by one or more computing devices, the phonetic representation estimate for the particular query term as the canonical phonetic representation, in the phonetic pronunciation dictionary, of the term that does occur in the lexicon of terms, deleting the particular query term from the lexicon of terms.

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

identifying, from among a set of query terms, a particular query term that (i) does not occur in a lexicon of terms, and (ii) has no designated, canonical phonetic representation in a pronunciation phonetic dictionary, wherein a canonical phonetic representation comprises a sequence of phonemes;

generating a phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the pronunciation phonetic dictionary;

transmitting data identifying at least a portion of a term that does occur in the lexicon of terms and the particular query term to a spelling correction server;

receiving, from the spelling correction server, data that specifies a spelling correction confidence score, wherein the spelling correction confidence score reflects a probability that the term that does occur in the lexicon of terms is a correct spelling of the particular query term;

determining that the spelling correction confidence score satisfies a predetermined threshold; and

in response to determining that the spelling correction confidence score satisfies a predetermined threshold, designating, by one or more computing devices, the phonetic representation estimate for the particular query term as a canonical phonetic representation, in the phonetic dictionary, of the term that does occur in the lexicon of terms.

13. The medium of claim 12, wherein generating a phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the phonetic pronunciation dictionary, comprises:

estimating a pronunciation of the particular query term based on one or more phonic rules for graphemes included in the particular query term.

14. The medium of claim 12, the operations comprising:

after generating the phonetic representation estimate for the particular query term that (i) does not occur in the lexicon of terms, and (ii) has no designated, canonical phonetic representation in the phonetic pronunciation dictionary, adding the particular query term to the lexicon of terms; and

after designating, by one or more computing devices, the phonetic representation estimate for the particular query term as the canonical phonetic representation, in the phonetic pronunciation dictionary, of the term that does occur in the lexicon of terms, deleting the particular query term from the lexicon of terms.