Wakeword selection
A system and method are disclosed capable of parsing a spoken utterance into a natural language request and a speech audio segment, where the natural language request directs the system to use the speech audio segment as a new wakeword. In response to this wakeword assignment directive, the system and method are further capable of immediately building a new wakeword spotter to activate the device upon matching the new wakeword in the input audio. Different approaches to promptly building a new wakeword spotter are described. Variations of wakeword assignment directives can make the new wakeword public or private. They can also add the new wakeword to earlier wakewords, or replace earlier wakewords.
1. A method of modifying a wakeword set of a speech- enabled device, comprising:
receiving, from a user, a request to modify the wakeword set;
parsing the request into a modification request and a selection of a factory-installed wakeword, wherein the modification request instructs the device to accept the selection of the factory-installed wakeword as a wakeword; and
enabling a factory-installed wakeword spotter to recognize the factory-installed wakeword as an activation trigger.
2. The method of claim 1 , further comprising adding the wakeword to the wakeword set comprising an earlier wakeword, the factory-installed wakeword spotter activating the device upon receiving a speech audio segment comprising the wakeword or the earlier wakeword.
3. The method of claim 1 , further comprising generating a voiceprint of a user's voice, the voiceprint being suitable for speaker verification.
4. The method of claim 1 , wherein the step of enabling the factory-installed wakeword spotter comprises dividing the wakeword into phonetic segments and using spotters for at least some of the phonetic segments.
5. A method of modifying a wakeword set of a speech- enabled device, comprising:
receiving an utterance;
parsing the utterance into a natural language request and a speech audio segment, wherein the natural language request instructs the device to accept the speech audio segment as a new wakeword;
using automatic speech recognition to map the new wakeword to a phonetic sequence; and
configuring a wakeword spotter to recognize the phonetic sequence by:
dividing the phonetic sequence into a sequence of two or more successive partial phonetic segments;
selecting a plurality of corresponding factory-installed partial wakeword spotters; and
assembling sequentially the factory-installed partial wakeword spotters into the wakeword spotter for the phonetic sequence.
6. The method of claim 5 , wherein said step of dividing the phonetic sequence into phonetic segments comprises dividing the phonetic sequence into separate syllables.
7. The method of claim 5 , wherein said step of dividing the phonetic sequence into phonetic segments comprises dividing the phonetic sequence into phonetic segments smaller than syllables.
8. The method of claim 5 , wherein selecting a partial wakeword spotter for a phonetic segment comprises:
identifying a collection of cached wakeword spotters, indexed by wakeword; and
retrieving a cached wakeword spotter of the collection of cached wakeword spotters for the partial phonetic segment.
9. The method of claim 5 , wherein said step of selecting partial wakeword spotters comprises retrieving, from a cache of wakeword spotters, partial wakeword spotters for successive phonetic segments of the phonetic sequence, and assembling the partial wakeword spotters into the wakeword spotter.
10. The method of claim 5 , wherein said step of selecting partial wakeword spotters comprises:
checking memory for a cached wakeword spotter for a phonetic segment, and
building a wakeword spotter for the phonetic segment if the phonetic segment does not have a wakeword spotter cached in memory.
11. The method of claim 5 , further comprising adding the new wakeword to the wakeword set comprising an earlier wakeword, the wakeword spotter activating the device upon receiving a speech audio segment comprising the new wakeword or the earlier wakeword.
12. A method of modifying a wakeword set of a speech- enabled device, comprising:
receiving a request to modify the wakeword set;
parsing the request into a modification request and a selection of a factory-installed wakeword, wherein the modification request instructs the device to accept the selection of the factory-installed wakeword as a wakeword; and
defining the a wakeword spotter as a sequence of phonemes.
13. The method of claim 12 , wherein the wakeword is a private wakeword to a person that provided the spoken request.
14. The method of claim 12 , further comprising adding the wakeword to the wakeword set comprising an earlier wakeword, the wakeword spotter activating the device upon receiving a speech audio segment comprising the wakeword or the earlier wakeword.