Unsupervised and active learning in automatic speech recognition for call classification
View Patent ↗Utterance data that includes at least a small amount of manually transcribed data is provided. Automatic speech recognition is performed on ones of the utterance data not having a corresponding manual transcription to produce automatically transcribed utterances. A model is trained using all of the manually transcribed data and the automatically transcribed utterances. A predetermined number of utterances not having a corresponding manual transcription are intelligently selected and manually transcribed. Ones of the automatically transcribed data as well as ones having a corresponding manual transcription are labeled. In another aspect of the invention, audio data is mined from at least one source, and a language model is trained for call classification from the mined audio data to produce a language model.
1. A method comprising:
performing, via a processor, automatic speech recognition using a bootstrap model on utterance data not having a corresponding manual transcription, to produce automatically transcribed utterances, wherein the bootstrap model is based on text data mined from a website relevant to a specific domain;
selecting, via the processor, a predetermined number of utterances not having a corresponding manual transcription based on a geometrically computed n-tuple confidence score;
receiving transcriptions of the predetermined number of utterances, wherein the transcriptions are made by a human being; and
generating a language model based on the automatically transcribed utterances, the predetermined number of utterances, and the transcriptions.
2. The method of claim 1 , further comprising:
performing automatic speech recognition using the language model.
3. The method of claim 2 , further comprising:
iteratively repeating the performing of automatic speech recognition using the bootstrap model, the selecting, the receiving, the generating, and the performing of speech recognition using the language model until a word accuracy converges.
4. The method of claim 1 , wherein the predetermined number of utterances correspond to a specific number of utterances having lowest confidence scores.
5. The method of claim 1 , wherein the predetermined number of utterances used in generating the language model are equal in number to the automatically transcribed utterances.
6. The method of claim 1 , wherein the predetermined number of utterances are randomly selected.
7. The method of claim 1 , wherein the language model is further based on the bootstrap model.
8. A system comprising:
a processor; and
a computer-readable storage medium having instructions stored which, when executed by the processor, cause the processor to perform operations comprising:
performing automatic speech recognition using a bootstrap model on utterance data not having a corresponding manual transcription, to produce automatically transcribed utterances, wherein the bootstrap model is based on text data mined from a website relevant to a specific domain;
selecting a predetermined number of utterances not having a corresponding manual transcription based on a geometrically computed n-tuple confidence score;
receiving transcriptions of the predetermined number of utterances, wherein the transcriptions are made by a human being; and
generating a language model based on the automatically transcribed utterances, the predetermined number of utterances, and the transcriptions.
9. The system of claim 8 , the computer-readable storage medium having additional instructions stored which, when executed by the processor, result in operations comprising:
performing automatic speech recognition using the language model.
10. The system of claim 9 , the computer-readable storage medium having additional instructions stored which, when executed by the processor, result in operations comprising:
iteratively repeating the performing of automatic speech recognition using the bootstrap model, the selecting, the receiving, the generating, and the performing of speech recognition using the language model until a word accuracy converges.
11. The system of claim 8 , wherein the predetermined number of utterances correspond to a specific number of utterances having lowest confidence scores.
12. The system of claim 8 , wherein the predetermined number of utterances used in generating the language model are equal in number to the automatically transcribed utterances.
13. The system of claim 8 , wherein the predetermined number of utterances are randomly selected.
14. The system of claim 8 , wherein the language model is further based on the bootstrap model.
15. A computer-readable storage device having instructions stored which, when executed by a computing device, cause the computing device to perform operations comprising:
performing automatic speech recognition using a bootstrap model on utterance data not having a corresponding manual transcription, to produce automatically transcribed utterances, wherein the bootstrap model is based on text data mined from a website relevant to a specific domain;
selecting a predetermined number of utterances not having a corresponding manual transcription based on a geometrically computed n-tuple confidence score;
receiving transcriptions of the predetermined number of utterances, wherein the transcriptions are made by a human being; and
generating a language model based on the automatically transcribed utterances, the predetermined number of utterances, and the transcriptions.
16. The computer-readable storage device of claim 15 , having additional instructions stored which, when executed by the processor, result in operations comprising:
performing automatic speech recognition using the language model.
17. The computer-readable storage device of claim 16 , having additional instructions stored which, when executed by the processor, result in operations comprising:
iteratively repeating the performing of automatic speech recognition using the bootstrap model, the selecting, the receiving, the generating, and the performing of speech recognition using the language model until a word accuracy converges.
18. The computer-readable storage device of claim 15 , wherein the predetermined number of utterances correspond to a specific number of utterances having lowest confidence scores.
19. The computer-readable storage device of claim 15 , wherein the predetermined number of utterances used in generating the language model are equal in number to the automatically transcribed utterances.
20. The computer-readable storage device of claim 15 , wherein the predetermined number of utterances are randomly selected.