Method and system for automatically detecting morphemes in a task classification system using lattices

Patent metadata

Field	Value
Publication number	US-9514126-B2
Application number	US-201414540179-A
Country	US
Kind code	B2
Filing date	Nov 13, 2014
Priority date	Nov 5, 1999
Publication date	Dec 6, 2016
Grant date	Dec 6, 2016

Abstract

The invention concerns a method and corresponding system for building a phonotactic model for domain independent speech recognition. The method may include recognizing phones from a user's input communication using a current phonotactic model, detecting morphemes (acoustic and/or non-acoustic) from the recognized phones, and outputting the detected morphemes for processing. The method also updates the phonotactic model with the detected morphemes and stores the new model in a database for use by the system during the next user interaction. The method may also include making task-type classification decisions based on the detected morphemes from the user's input communication.

First claim

What is claimed is:

1. A method comprising:
recognizing, via a processor and with a current phonotactic model, phonemes using a combination of speech and touchscreen input from a first user input, to yield recognized phonemes;
transforming a prior probability distribution from a first domain to a second domain without speech transcriptions to yield a transformed prior probability distribution, wherein the first domain and the second domain comprise word sequence spaces;
creating, via the processor, a new phonotactic model associated with the second domain using morphemes detected from:
the recognized phonemes; and
the transformed prior probability distribution associated with the first domain, wherein the first domain is distinct from the second domain;

replacing, via the processor, the current phonotactic model with the new phonotactic model in a database;
recognizing, via the processor, further phonemes in second input using the new phototactic model.

2. The method of claim 1 , wherein the recognizing of the phonemes is further based on an environment of a user.

3. The method of claim 1 , wherein the morphemes are expressed in multimodal form.

4. The method of claim 1 , wherein the morphemes in the speech are derived from an action of a user.

5. The method of claim 4 , wherein the action of the user comprises a focus of attention of the user.

6. The method of claim 1 , operating in conjunction with one of a telephone network, Internet, an intranet, a cable television network, a local area network, and a wireless communication network.

7. A system comprising:
a processor; and
a computer-readable storage medium having instructions stored which, when executed by the processor, result in the processor performing operations comprising:
recognizing, using a current phonotactic model, phonemes using a combination of speech and touchscreen input from a first user input, to yield recognized phonemes;
transforming a prior probability distribution from a first domain to a second domain without speech transcriptions to yield a transformed prior probability distribution, wherein the first domain and the second domain comprise word sequence spaces;
creating a new phonotactic model associated with the second domain using morphemes detected from:
the recognized phonemes; and
the transformed prior probability distribution associated with the first domain, wherein the first domain is distinct from the second domain;
replacing the current phonotactic model with the new phonotactic model in a database;

recognizing further phonemes in second input using the new phototactic model.

8. The system of claim 7 , wherein the recognizing of the phonemes is further based on an environment of a user.

9. The system of claim 7 , wherein the morphemes are expressed in multimodal form.

10. The system of claim 7 , wherein the morphemes in the speech are derived from an action of a user.

11. The system of claim 10 , wherein the action of the user comprises a focus of attention of the user.

12. The system of claim 7 , operating in conjunction with one of a telephone network, Internet, an intranet, a cable television network, a local area network, and a wireless communication network.

13. A computer-readable storage device having instructions stored which, when executed by a computing device, result in the computing device performing operations comprising:
recognizing with a current phonotactic model, phonemes using a combination of speech and touchscreen input from a first user input, to yield recognized phonemes;
transforming a prior probability distribution from a first domain to a second domain without speech transcriptions to yield a transformed prior probability distribution, wherein the first domain and the second domain comprise word sequence spaces;
creating a new phonotactic model associated with the second domain using morphemes detected from:
the recognized phonemes; and
the transformed prior probability distribution associated with the first domain, wherein the first domain is distinct from the second domain;
replacing the current phonotactic model with the new phonotactic model in a database;
recognizing further phonemes in second input using the new phototactic model.

14. The computer-readable storage device of claim 13 , wherein the recognizing of the phonemes is further based on an environment of a user.

15. The computer-readable storage device of claim 13 , wherein the morphemes are expressed in multimodal form.

Assignees

• At & T Ip Ii Lp

Inventors

• Riccardi Giuseppe

Classifications

• G10L15/02 Primary

  Feature extraction for speech recognition; Selection of recognition unit · CPC title

• G06F40/40 Primary

  Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30) · CPC title

• G10L15/187

  Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams · CPC title

• G10L25/27

  characterised by the analysis technique · CPC title

• G10L15/07

  to the speaker · CPC title