Abstract: Generally discussed herein are devices, systems, and methods for on-device detection of a wake word. A device can include a memory including model parameters that define a custom wake word detection model, the wake word detection model including a recurrent neural network transducer (RNNT) and a lookup table (LUT), the LUT indicating a hidden vector to be provided in response to a phoneme of a user-specified wake word, a microphone to capture audio, and processing circuitry to receive the audio from the microphone, determine, using the wake word detection model, whether the audio includes an utterance of the user-specified wake word, and wake up a personal assistant after determining the audio includes the utterance of the user-specified wake word.

Abstract: Generally discussed herein are devices, systems, and methods for on-device detection of a wake word. A device can include a memory including model parameters that define a custom wake word detection model, the wake word detection model including a recurrent neural network transducer (RNNT) and a lookup table (LUT), the LUT indicating a hidden vector to be provided in response to a phoneme of a user-specified wake word, a microphone to capture audio, and processing circuitry to receive the audio from the microphone, determine, using the wake word detection model, whether the audio includes an utterance of the user-specified wake word, and wake up a personal assistant after determining the audio includes the utterance of the user-specified wake word.

Abstract: Generally discussed herein are devices, systems, and methods for on-device detection of a wake word. A device can include a memory including model parameters that define a custom wake word detection model, the wake word detection model including a recurrent neural network transducer (RNNT) and a lookup table (LUT), the LUT indicating a hidden vector to be provided in response to a phoneme of a user-specified wake word, a microphone to capture audio, and processing circuitry to receive the audio from the microphone, determine, using the wake word detection model, whether the audio includes an utterance of the user-specified wake word, and wake up a personal assistant after determining the audio includes the utterance of the user-specified wake word.

Abstract: Generally discussed herein are devices, systems, and methods for on-device detection of a wake word. A device can include a memory including model parameters that define a custom wake word detection model, the wake word detection model including a recurrent neural network transducer (RNNT) and a lookup table (LUT), the LUT indicating a hidden vector to be provided in response to a phoneme of a user-specified wake word, a microphone to capture audio, and processing circuitry to receive the audio from the microphone, determine, using the wake word detection model, whether the audio includes an utterance of the user-specified wake word, and wake up a personal assistant after determining the audio includes the utterance of the user-specified wake word.

Abstract: Disclosed is an apparatus for storing media. The apparatus comprises a bottom cover, a pillar assembly mounted inside the bottom cover, a handling plate mounted within the pillar assembly and coupled to the pillar assembly via a linear actuator assembly. Further, the bottom cover comprises a cassette indexing mechanism mounted on the bottom surface of the bottom cover and a media entry assembly mounted through the side wall of the bottom cover. Furthermore, the pillar assembly comprises a shaft extending through the bottom surface of the bottom cover and the shaft is coupled with the cassette indexing mechanism for indexing the pillar assembly in a predefined configuration. Further, the handling plate is indexed along with the pillar assembly during operation, and the handling plate is configured to store a plurality of media indexed with respect to each other.

Abstract: In some embodiments, Satellite Positioning System (SPS) time information associated with at least one SPS may be maintained at a UE, which may also receive time information from a Wireless Wide Area Network (WWAN). In some embodiments, the UE may determine a corrected SPS time information for a first time based, in part, on the received WWAN time information, where the corrected SPS time information corrects the SPS time information associated with the at least one SPS maintained at the UE. The UE may initiate transmission of SPS timing assistance information to an associated device over a Wireless Personal Area Network (WPAN), wherein the SPS timing assistance information comprises the corrected SPS time information for the first time.

Abstract: Disclosed is an apparatus for storing media. The apparatus comprises a bottom cover, a pillar assembly mounted inside the bottom cover, a handling plate mounted within the pillar assembly and coupled to the pillar assembly via a linear actuator assembly. Further, the bottom cover comprises a cassette indexing mechanism mounted on the bottom surface of the bottom cover and a media entry assembly mounted through the side wall of the bottom cover. Furthermore, the pillar assembly comprises a shaft extending through the bottom surface of the bottom cover and the shaft is coupled with the cassette indexing mechanism for indexing the pillar assembly in a predefined configuration. Further, the handling plate is indexed along with the pillar assembly during operation, and the handling plate is configured to store a plurality of media indexed with respect to each other.

Abstract: In some embodiments, Satellite Positioning System (SPS) time information associated with at least one SPS may be maintained at a UE, which may also receive time information from a Wireless Wide Area Network (WWAN). In some embodiments, the UE may determine a corrected SPS time information for a first time based, in part, on the received WWAN time information, where the corrected SPS time information corrects the SPS time information associated with the at least one SPS maintained at the UE. The UE may initiate transmission of SPS timing assistance information to an associated device over a Wireless Personal Area Network (WPAN), wherein the SPS timing assistance information comprises the corrected SPS time information for the first time.

Abstract: In some examples, compensating for hysteretic characteristics of a crystal oscillator in a timing circuit includes obtaining a plurality of successive temperature measurements. From the plurality of successive temperature measurements, a temperature gradient having a sign and a magnitude can be determined. A frequency compensation parameter can then be determined based on any combination of two or more factors chosen from a set of factors including a temperature measurement, the sign of the temperature gradient, and the magnitude of the temperature gradient. A frequency error of the timing circuit can then be compensated based on the frequency compensation parameter.

Abstract: In some examples, compensating for hysteretic characteristics of a crystal oscillator in a timing circuit includes obtaining a plurality of successive temperature measurements. From the plurality of successive temperature measurements, a temperature gradient having a sign and a magnitude can be determined. A frequency compensation parameter can then be determined based on any combination of two or more factors chosen from a set of factors including a temperature measurement, the sign of the temperature gradient, and the magnitude of the temperature gradient. A frequency error of the timing circuit can then be compensated based on the frequency compensation parameter.