Generating a task-adapted acoustic model from one or more different corpora

• US 7,263,487 B2
• Filed: 09/29/2005
• Issued: 08/28/2007
• Est. Priority Date: 03/20/2002
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Hwang, Mei Yuh
• Primary Examiner(s): Storm; Donald L.
• Application Number:
  US11/238,807
• Publication Number:
  US 20060036444A1
• Time in Patent Office: 698 Days
• Field of Search: None
• US Class Current: 704/243
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models

Abstract

The present invention generates a task-dependent acoustic model from a supervised task-independent corpus and further adapted it with an unsupervised task dependent corpus. The task-independent corpus includes task-independent training data which has an acoustic representation of words and a sequence of transcribed words corresponding to the acoustic representation. A relevance measure is defined for each of the words in the task-independent data. The relevance measure is used to weight the data associated with each of the words in the task-independent training data. The task-dependent acoustic model is then trained based on the weighted data for the words in the task-independent training data.

Claims

1. A method of generating a task-dependent acoustic model from a task-independent (TI) training corpus that includes an acoustic representation of an utterance and a sequence of transcribed words corresponding to the acoustic representation, the method comprising:

deriving a task relevance measure for each word in the TI training corpus, indicative of a relevance of the words to a task;

training a task-independent (TI) AM based on the TI training corpus, the TI AM including words from the TI training corpus and associated AM parameters;

generating a task-dependent (TD) acoustic model (AM) based on the TI AM trained from the TI training corpus and the task relevance measures for the words in the TI training corpus; and

storing the TD AM for use in performing the task.

2. The method of claim 1 wherein generating a TD AM comprises:

weighting words in the TI training corpus.

3. The method of claim 2 wherein generating a TD AM comprises:

training the TD AM using the weighted words.

4. The method of claim 1 wherein generating a TD AM comprises:

extracting relevant data from the TI training corpus based on the task relevance measures.

5. The method of claim 4 wherein generating a TD AM comprises:

training the TD AM based on the relevant data.

6. The method of claim 1 wherein deriving a task relevance measure comprises:

selecting a word from the TI training corpus; and

defining the task relevance measure for the selected word based on a portion of the selected word that is found in the task.

7. The method of claim 6 wherein defining the task relevance measure for the selected word, comprise:

defining the task relevance measure for the selected word based on a number of relevant triphones in the selected word, the number of relevant triphones being triphones in the selected word that are found in the task.

8. A system for generating a task-dependent (TD) acoustic model (AM) from a task-independent (TI) training corpus, comprising:

a task relevance generator receiving a task input indicative of words relevant to a task, the task comprising an application in which the TD AM is to be used, and the task relevance generator configured to generate a relevance measure for each word in the TI training corpus based on a relationship of each word in the TI training corpus to the task input; and

an AM generator, coupled to the TI training corpus and the task relevance generator and configured to generate the TD AM based on the TI training corpus and the relevance measure.

9. The system of claim 8 wherein the task relevance generator is configured to generate the relevance measure for a selected word based on a number of phonetic units in the selected word that are in the task input.

10. A method of generating a task-dependent acoustic model, comprising:

receiving a task input indicative of phonetic units used in a task for which the task-dependent acoustic model is to be used;

calculating a relevance, to the task for which the task-dependent acoustic model is to be used, of phonetic units in a task-independent training corpus of phonetic units, the phonetic units in the task-independent training corpus having no predefined relevance to the task for which the task-dependent acoustic model is to be used;

generating the task-dependent acoustic model by applying the calculated relevance to the phonetic units in the task-independent training corpus; and

storing the task-dependent acoustic model for use in performing the task.

11. The method of claim 10 wherein generating the task-dependent acoustic model includes:

weighting acoustic model parameters for the phonetic units in the task-independent training corpus based on the relevance calculated; and

generating the task-dependent acoustic model using the weighted acoustic model parameters.