Unsupervised labeling of sentence level accent

• US 7,844,457 B2
• Filed: 02/20/2007
• Issued: 11/30/2010
• Est. Priority Date: 02/20/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Chen, YiNing, Chu, Min, Soong, Frank Kao-ping
• Primary Examiner(s): Smits; Talivaldis I
• Assistant Examiner(s): BORSETTI, GREG
• Application Number:
  US11/708,442
• Publication Number:
  US 20080201145A1
• Time in Patent Office: 1,379 Days
• Field of Search: 704 9- 10, 704243-255, 704/E15.02, 704/E15.025
• US Class Current: 704/244
• CPC Class Codes: G10L 13/08 Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination

Abstract

Methods are disclosed for automatic accent labeling without manually labeled data. The methods are designed to exploit accent distribution between function and content words.

Claims

1. A computer-implemented method of training an acoustic model, the method comprising:

classifying each of a plurality of words as being either a content word or a function word;

utilizing a characteristic of at least one of the content words as a basis for identifying an accent characteristic of at least one of the function words;

utilizing a computer processor that is a functional component of a computer to train the acoustic model with a collection of data so as to be indicative of the accent characteristic of the at least one of the function words, to be indicative of accented vowels of words that have been classified as being content words, and to be indicative of unaccented vowels of words that have been classified as being function words; and

wherein accented vowels of words that have been labeled as function words are excluded from the collection of data utilized to train the acoustic model.

2. The method of claim 1, wherein training the acoustic model comprises training so as to add an indication of an unaccented vowel of a word that has been classified as a function word.

3. The method of claim 2, further comprising training the acoustic model so as to add an indication of an accented vowel of a word that has been classified as a content word.

4. The method of claim 1, further comprising utilizing the characteristic as a basis for labeling an accent characteristic of at least one of the function words.

5. The method of claim 1, wherein utilizing a characteristic of at least one of the content words as a basis for identifying an accent characteristic of at least one of the function words comprises utilizing accented and unaccented vowels.

6. The method of claim 1, further comprising utilizing the acoustic model as a basis for labeling the function word.

7. A computer-implemented method of training an acoustic model, the method comprising:

utilizing a computer processor that is a functional component of a computer and a first acoustic model to label accented and unaccented components of function words;

utilizing the unaccented components of the function words as a basis for training a second acoustic model; and

excluding the accented components of the function words from a collective set of data utilized as a basis for training the second acoustic model.

8. The method of claim 7, wherein the first acoustic model contains a representation of accented and unaccented components of words that have been identified as being content words.

9. The method of claim 7, further comprising utilizing the second acoustic model as a basis for labeling accented and unaccented components of words that have been identified as being function words.