D ShinFollow


Users often invoke a virtual assistant to quickly accomplish routine tasks that require the corresponding components of a device to be woken up from an idle state prior to performing the desired action. Activating the components introduces an inherent latency that can noticeably raise the task completion time and degrade the user experience. This disclosure describes techniques to reduce the latency involved by proactively waking up idle device components by predicting the intended task based on the initial words in a user’s spoken command to a voice based virtual assistant. The net effect of starting the wake-up process based on partial decoding of the spoken query can reduce latency without affecting power consumption.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.