Abstract: System and method for operating an always-on ASR (automatic speech recognition) system by selecting target keywords and continuously detecting the selected target keywords in voice commands in a mobile device are provided. In the mobile device, a processor is configured to collect keyword candidates, collect usage frequency data for keywords in the keyword candidates, collect situational usage frequency data for the keywords in the keyword candidates, select target keywords from the keyword candidates based on the usage frequency data and the situational usage frequency data, and detect one or more of the target keywords in a voice command using continuous detection of the target keywords.

Abstract: A method performed by an artificial neural network (ANN) includes determining, at a first stage of a multi-stage cross-attention model of the ANN, a first cross-correlation between a first representation of each modality of a number of modalities associated with a sequence of inputs. The method still further includes determining, at each second stage of one or more second stages of the multi-stage cross-attention model, a second cross-correlation between first attended representations of each modality. The method also includes generating a concatenated feature representation associated with a final second stage of the one or more second stages based on the second cross-correlation associated with the final second stage, the first attended representation of each modality, and the first representation of each modality. The method further includes determining a probability distribution between a set of background actions and a set of foreground actions from the concatenated feature representation.

Abstract: Certain aspects of the present disclosure are generally directed to apparatus and techniques for event state detection. One example method generally includes receiving a plurality of sensor signals at a computing device, determining, at the computing device, probabilities of sub-event states based on the plurality of sensor signals using an artificial neural network for each of a plurality of time intervals, and detecting, at the computing device, the event state based on the probabilities of the sub-event states via a state sequence model.

Abstract: A device to process an audio signal representing input sound includes a user voice verifier configured to generate a first indication based on whether the audio signal represents a user's voice. The device includes a speaking target detector configured to generate a second indication based on whether the audio signal represents at least one of a command or a question. The device includes an activation signal unit configured to selectively generate an activation signal based on the first indication and the second indication. The device also includes an automatic speech recognition engine configured to be activated, responsive to the activation signal, to process the audio signal.

Abstract: A device to perform end-of-utterance detection includes a speaker vector extractor configured to receive a frame of an audio signal and to generate a speaker vector that corresponds to the frame. The device also includes an end-of-utterance detector configured to process the speaker vector and to generate an indicator that indicates whether the frame corresponds to an end of an utterance of a particular speaker.

Abstract: A device includes a screen and one or more processors configured to provide, at the screen, a graphical user interface (GUI) configured to display data associated with multiple devices on the screen. The GUI is also configured to illustrate a label and at least one control input for each device of the multiple devices. The GUI is also configured to provide feedback to a user. The feedback indicates that a verbal command is not recognized with an action to be performed. The GUI is also configured to provide instructions for the user on how to teach the one or more processors which action is to be performed in response to receiving the verbal command.

Abstract: In one embodiment, an electronic device includes an input device configured to provide an input stream, a first processing device, and a second processing device. The first processing device is configured to use a keyword-detection model to determine if the input stream comprises a keyword, wake up the second processing device in response to determining that a segment of the input stream comprises the keyword, and modify the keyword-detection model in response to a training input received from the second processing device. The second processing device is configured to use a first neural network to determine whether the segment of the input stream comprises the keyword and provide the training input to the first processing device in response to determining that the segment of the input stream does not comprise the keyword.

Abstract: A device to process an audio signal representing input sound includes a user voice verifier configured to generate a first indication based on whether the audio signal represents a user's voice. The device includes a speaking target detector configured to generate a second indication based on whether the audio signal represents at least one of a command or a question. The device includes an activation signal unit configured to selectively generate an activation signal based on the first indication and the second indication. The device also includes an automatic speech recognition engine configured to be activated, responsive to the activation signal, to process the audio signal.

Abstract: A device to process an audio signal representing input sound includes a hand detector configured to generate a first indication responsive to detection of at least a portion of a hand over at least a portion of the device. The device also includes an automatic speech recognition system configured to be activated, responsive to the first indication, to process the audio signal.

Abstract: A method for operating a neural network includes receiving an input sequence at an encoder. The input sequence is encoded to produce a set of hidden representations. Attention-heads of the neural network calculate attention weights based on the hidden representations. A context vector is calculated for each attention-head based on the attention weights and the hidden representations. Each of the context vectors correspond to a portion of the input sequence. An inference is output based on the context vectors.

Abstract: A device includes a screen and one or more processors configured to provide, at the screen, a graphical user interface (GUI) configured to display data associated with multiple devices on the screen. The GUI is also configured to illustrate a label and at least one control input for each device of the multiple devices. The GUI is also configured to provide feedback to a user. The feedback indicates that a verbal command is not recognized with an action to be performed. The GUI is also configured to provide instructions for the user on how to teach the one or more processors which action is to be performed in response to receiving the verbal command.

Abstract: A device to perform end-of-utterance detection includes a speaker vector extractor configured to receive a frame of an audio signal and to generate a speaker vector that corresponds to the frame. The device also includes an end-of-utterance detector configured to process the speaker vector and to generate an indicator that indicates whether the frame corresponds to an end of an utterance of a particular speaker.

Abstract: A method of determining, by an electronic device, an audio event is disclosed. The method may include receiving an input sound from a sound source by a plurality of sound sensors. The method may also extracting, by a processor, at least one sound feature from the received input sound, determining, by the processor, location information of the sound source based on the input sound received by the sound sensors, determining, by the processor, the audio event indicative of the input sound based on the at least one sound feature and the location information, and transmitting, by a communication unit, a notification of the audio event to an external electronic device.

Abstract: A processor is configured to transition in and out of a low-power state at a first rate and to operate in a first mode or a second mode. In a particular method, the processor while coupled to a coder/decoder (CODEC) retrieves audio feature data from a buffer after transitioning out of the low-power state. The CODEC is configured to operate at a second rate in the first mode and at a third rate in the second mode, the second rate and the third rate each greater than the first rate. The audio feature data indicates features of audio data received during the low-power state of the processor. A ratio of CODEC activity to processor activity in the second mode is less than the ratio in the first mode.

Abstract: A method, performed by an electronic device, for verifying a user to allow access to the electronic device is disclosed. In this method, sensor data may be received from a plurality of sensors including at least an image sensor and a sound sensor. Context information of the electronic device may be determined based on the sensor data and at least one verification unit may be selected from a plurality of verification units based on the context information. Based on the sensor data from at least one of the image sensor or the sound sensor, the at least one selected verification unit may calculate at least one verification value. The method may determine whether to allow the user to access the electronic device based on the at least one verification value and the context information.

Abstract: A method for controlling voice activation by a target keyword in a mobile device is disclosed. The method includes receiving an input sound stream. When the input sound stream indicates speech, the voice activation unit is activated to detect the target keyword and at least one sound feature is extracted from the input sound stream. Further, the method includes deactivating the voice activation unit when the at least one sound feature indicates a non-target keyword.

Abstract: According to an aspect of the present disclosure, a method for controlling access to a plurality of electronic devices is disclosed. The method includes detecting whether a first device is in contact with a user, adjusting a security level of the first device to activate the first device when the first device is in contact with the user, detecting at least one second device within a communication range of the first device, and adjusting a security level of the at least one second device to control access to the at least one second device based on a distance between the first device and the at least one second device.

Abstract: A method of determining, by an electronic device, an audio event is disclosed. The method may include receiving an input sound from a sound source by a plurality of sound sensors. The method may also extracting, by a processor, at least one sound feature from the received input sound, determining, by the processor, location information of the sound source based on the input sound received by the sound sensors, determining, by the processor, the audio event indicative of the input sound based on the at least one sound feature and the location information, and transmitting, by a communication unit, a notification of the audio event to an external electronic device.

Abstract: A method and apparatus for performing a function in a mobile device are disclosed. A media sound from a sound output device external to the mobile device is captured and a sound feature is extracted from the captured media sound. A function to be performed in the mobile device is determined by identifying at least one reference sound feature in a set of reference sound features based on the extracted sound feature, each reference sound feature in the set of reference sound features being associated with at least one of a plurality of media sounds and at least one of a plurality of functions. Further, the determined function is performed in the mobile device.

Abstract: A method for recognizing an environmental sound in a client device in cooperation with a server is disclosed. The client device includes a client database having a plurality of sound models of environmental sounds and a plurality of labels, each of which identifies at least one sound model. The client device receives an input environmental sound and generates an input sound model based on the input environmental sound. At the client device, a similarity value is determined between the input sound model and each of the sound models to identify one or more sound models from the client database that are similar to the input sound model. A label is selected from labels associated with the identified sound models, and the selected label is associated with the input environmental sound based on a confidence level of the selected label.