Dynamic enrollment of user-defined wake-up key-phrase for speech enabled computer system

Patent metadata

Field	Value
Publication number	US-10672380-B2
Application number	US-201715855379-A
Country	US
Kind code	B2
Filing date	Dec 27, 2017
Priority date	Dec 27, 2017
Publication date	Jun 2, 2020
Grant date	Jun 2, 2020

Abstract

Techniques are provided for wake-on-voice (WOV) key-phrase enrollment. A methodology implementing the techniques according to an embodiment includes generating a WOV key-phrase model based on identification of the sequence of sub-phonetic units of a user-provided key-phrase. The WOV key-phrase model is employed by a WOV processor for detection of the user spoken key-phrase and triggering operation of an automatic speech recognition (ASR) processor in response to the detection. The method further includes updating an ASR language model based on the user-provided key-phrase. The update includes one of embedding the WOV key-phrase model into the ASR language model, converting sub-phonetic units of the WOV key-phrase model and embedding the converted WOV key-phrase model into the ASR language model, or generating an ASR key-phrase model by applying a phoneme-syllable based statistical language model to the user-provided key-phrase and embedding the generated ASR key-phrase model into the ASR language model.

First claim

What is claimed is:

1. A processor-implemented method for wake-on-voice (WOV) key-phrase enrollment, the method comprising:
generating, by a processor-based system, a WOV key-phrase model based on a user-provided WOV enrollment key-phrase, the WOV key-phrase model employed by a WOV processor for detecting of a correct sequence of sub-phonetic units of the WOV key-phrase spoken by the user and triggering operation of an automatic speech recognition (ASR) processor in response to the WOV key-phrase detection; and
updating, by the processor-based system, an ASR language model based on the user-provided WOV enrollment key-phrase, the ASR language model employed by the ASR processor for recognizing speech utterances spoken by the user, wherein updating the ASR language model comprises generating an ASR key-phrase model by applying a phoneme-syllable based statistical language model to the user-provided WOV enrollment key-phrase and incorporating the generated ASR key-phrase model into the ASR language model.

2. The method of claim 1 , wherein the user-provided WOV enrollment key-phrase is provided as a text entry, the method further comprising performing a grapheme to phoneme conversion on the text entry for the generation of the WOV key-phrase model.

3. The method of claim 1 , wherein the triggering of the ASR processor comprises waking the ASR processor from a lower power consuming idle state to a higher power consuming recognition state.

4. The method of claim 3 , wherein the WOV processor consumes less power than the ASR processor when the ASR processor is in the higher power consuming recognition state.

5. A system for wake-on-voice (WOV) key-phrase enrollment, the system comprising:
a WOV key-phrase model generation circuit to generate a WOV key-phrase model based on a user-provided WOV enrollment key-phrase, the WOV key-phrase model employed by a WOV processor for detecting of a correct sequence of sub-phonetic units of the WOV key-phrase spoken by the user and triggering operation of an automatic speech recognition (ASR) processor in response to the WOV key-phrase detection;
an ASR model update circuit to update an ASR language model based on the user-provided WOV enrollment key-phrase, the ASR language model employed by the ASR processor for recognizing speech utterances spoken by the user; and
an ASR key-phrase model generation circuit to generate an ASR key-phrase model by applying a phoneme-syllable based statistical language model to the user-provided WOV enrollment key-phrase and incorporate the generated ASR key-phrase model into the ASR language model.

6. The system of claim 5 , wherein the user-provided WOV enrollment key-phrase is provided as a text entry, the system further comprises a grapheme to phoneme conversion circuit to convert the text entry to phonemes for the generation of the WOV key-phrase model.

7. The system of claim 5 , wherein the triggering of the ASR processor comprises waking the ASR processor from a lower power consuming idle state to a higher power consuming recognition state.

8. The system of claim 7 , wherein the WOV processor consumes less power than the ASR processor when the ASR processor is in the higher power consuming recognition state.

9. A processor-implemented method for wake-on-voice (WOV) key-phrase enrollment, the method comprising:
generating, by a processor-based system, a WOV key-phrase model based on a user-provided WOV enrollment key-phrase, the WOV key-phrase model employed by a WOV processor for detecting of a correct sequence of sub-phonetic units of the WOV key-phrase spoken by the user and triggering operation of an automatic speech recognition (ASR) processor in response to the WOV key-phrase detection; and
updating, by the processor-based system, an ASR language model based on the user-provided WOV enrollment key-phrase, the ASR language model employed by the ASR processor for recognizing speech utterances spoken by the user, wherein updating the ASR language model comprises performing a sub-phonetic conversion of the WOV key-phrase model and incorporating the converted WOV key-phrase model into the ASR language model.

10. The method of claim 9 , wherein the user-provided WOV enrollment key-phrase is provided as a text entry, the method further comprising performing a grapheme to phoneme conversion on the text entry for the generation of the WOV key-phrase model.

11. The method of claim 9 , wherein the triggering of the ASR processor comprises waking the ASR processor from a lower power consuming idle state to a higher power consuming recognition state.

12. The method of claim 11 , wherein the WOV processor consumes less power than the ASR processor when the ASR processor is in the higher power consuming recognition state.

Assignees

• Intel Ip Corp

Inventors

• Georges Munir Nikolai Alexander
• Bocklet Tobias
• Stemmer Georg
• Hofer Joachim
• Bauer Josef G

Classifications

• G10L15/02

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

• G10L2015/223

  Execution procedure of a spoken command · CPC title

• G10L15/063 Primary

  Training · CPC title

• G10L2015/088

  Word spotting · CPC title

• G10L2015/025

  Phonemes, fenemes or fenones being the recognition units · CPC title