Word-Dependent Language Model

• US 20120166196A1
• Filed: 12/23/2010
• Published: 06/28/2012
• Est. Priority Date: 12/23/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Ju, Yun-Cheng, Tashev, Ivan J., Heinemann, Chad R.
• Granted Patent: US 8,838,449 B2
• Time in Patent Office: Days
• US Class Current: 704/254
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules

Abstract

This document describes word-dependent language models, as well as their creation and use. A word-dependent language model can permit a speech-recognition engine to accurately verify that a speech utterance matches a multi-word phrase. This is useful in many contexts, including those where one or more letters of the expected phrase are known to the speaker.

Claims

1. A computer-implemented method comprising:

determining, based on a multi-word phrase, a word-dependent language model specific to the multi-word phrase and enabling word-level speech verification;

providing the word-dependent language model of the multi-word phase to a speech-recognition engine; and

receiving, from the speech-recognition engine and responsive to the speech-recognition engine receiving one or more human utterances associated with the multi-word phrase, an indication that the one or more utterances is verified or not verified to be the multi-word phrase.

2. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines, based on a length of a word in the multi-word phrase, a set of words having the length, and further comprising including the set of words in the word-dependent language model.

3. A computer-implemented method as described in claim 2, wherein the set of words can be used as part of an in-vocabulary recognition algorithm for a portion of the one or more human utterances associated with the word of the multi-word phrase.

4. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines, based on one or more letters and letter locations of multiple words in the multi-word phrase that are known to a person from which the one or more human utterances are expected, different accuracy thresholds for the multiple words and further comprising including the different accuracy thresholds for the multiple words in the word-dependent language model.

5. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines, based on one or more letters and letter locations of a word in the multi-word phrase that are known to a person from which the one or more human utterances are expected, a set of words having the one or more letters, and further comprising including the set of words in the word-dependent language model.

6. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines, based on a complete word of the multi-word phrase having been revealed to a person from which the one or more human utterances are expected, that the complete word does not need to be verified and further comprising including an indication that a portion of the one or more human utterances associated with this complete word does not need to be verified in the word-dependent language model.

7. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines higher-probability wrong words based on a length of a word in the multi-word phrase, and further comprising including the higher-probability wrong words in the word-dependent language model.

8. A computer-implemented method as described in claim 7, wherein the higher-probability wrong words are intended to be used as part of an out-of-vocabulary recognition algorithm for a portion of the one or more human utterances associated with the word of the multi-word phrase.

9. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines higher-probability wrong words based on a history of prior human utterances associated with a word in the multi-word phrase, and further comprising including the higher-probability wrong words in the word-dependent language model.

10. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines higher-probability wrong words based on common usage in a language in which the one or more human utterances are expected, and further comprising including the higher-probability wrong words in the word-dependent language model.

11. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model determines higher-probability wrong words for a word in the multi-word phrase based on one or more other words of the multi-word phrase that are adjacent to the word in the multi-word phrase, and further comprising including the higher-probability wrong words in the word-dependent language model.

12. A computer-implemented method as described in claim 1, wherein determining the word-dependent language model includes assigning different weights to different branches, each of the different branches associated with different words of the multi-word phrase and having one or more imposter words for respective different words of the multi-word phrase.

13. A computer-implemented method as described in claim 1, further comprising receiving the multi-word phrase from an entity and, responsive to receiving the indication that the one or more utterances is verified or not verified, passing the indication to the entity.

14. A computer-implemented method comprising:

receiving a word-dependent language model enabling sequential, word-level speech verification of a multi-word phrase;

receiving an utterance; and

verifying, sequentially, word-by-word, and based on the word-dependent language model, that the utterance does or does not include each word of the multi-word phrase.

15. A computer-implemented method as described in claim 14, wherein verifying that the utterance does or does not include each word performs an out-of-vocabulary recognition algorithm for one of the words of the multi-word phrase and based on a length of the one of the words.

16. A computer-implemented method as described in claim 14, wherein verifying that the utterance does or does not include each word performs an in-vocabulary recognition algorithm for one of the words of the multi-word phrase and based on a set of words in the word-dependent language model making up a vocabulary usable in performing the in-vocabulary recognition algorithm for the one of the words.

17. A computer-implemented method as described in claim 14, wherein verifying that the utterance does or does not include each word verifies different words of the multi-word phrase using different accuracy thresholds, the different accuracy thresholds included within the word-dependent language model.

18. A computer-implemented method as described in claim 14, wherein the word-dependent language model includes higher-probability wrong words for one of the words of the multi-word phrase and wherein verifying that the utterance does or does not include each word assigns a higher likelihood to one of the higher-probability wrong words for the one of the words of the multi-word phrase.

19. A computer-implemented method comprising:

receiving a multi-word answer to a question and context indicating information about the multi-word answer, the question and the context exposed to a user from which a human utterance in response to the question is expected;

determining, based on the context and the multi-word answer, a word-dependent language model specific to the multi-word answer and enabling word-level speech verification;

providing the word-dependent language model of the multi-word phase to a speech-recognition engine; and

receiving, responsive to the speech-recognition engine receiving the human utterance, an indication that the human utterance is verified or not verified to be the multi-word phrase.

20. A computer-implemented method as described in claim 19, further comprising presenting the question and the context to the user within a video game interface, the context exposing a grammar, structure, word, or letter of the multi-word answer, and presenting within the video game interface and responsive to receiving the indication that the human utterance is verified or not verified, that the utterance is correct or not correct, respectively.