Abstract: A method includes receiving a sequence of acoustic frames characterizing one or more utterances as input to a multilingual automated speech recognition (ASR) model. The method also includes generating a higher order feature representation for a corresponding acoustic frame. The method also includes generating a hidden representation based on a sequence of non-blank symbols output by a final softmax layer. The method also includes generating a probability distribution over possible speech recognition hypotheses based on the hidden representation generated by the prediction network at each of the plurality of output steps and the higher order feature representation generated by the encoder at each of the plurality of output steps. The method also includes predicting an end of utterance (EOU) token at an end of each utterance. The method also includes classifying each acoustic frame as either speech, initial silence, intermediate silence, or final silence.

Abstract: A method includes receiving a sequence of acoustic frames characterizing one or more utterances as input to a multilingual automated speech recognition (ASR) model. The method also includes generating a higher order feature representation for a corresponding acoustic frame. The method also includes generating a hidden representation based on a sequence of non-blank symbols output by a final softmax layer. The method also includes generating a probability distribution over possible speech recognition hypotheses based on the hidden representation generated by the prediction network at each of the plurality of output steps and the higher order feature representation generated by the encoder at each of the plurality of output steps. The method also includes predicting an end of utterance (EOU) token at an end of each utterance. The method also includes classifying each acoustic frame as either speech, initial silence, intermediate silence, or final silence.

Abstract: Techniques disclosed herein are directed towards streaming keyphrase detection which can be customized to detect one or more particular keyphrases, without requiring retraining of any model(s) for those particular keyphrase(s). Many implementations include processing audio data using a speaker separation model to generate separated audio data which isolates an utterance spoken by a human speaker from one or more additional sounds not spoken by the human speaker, and processing the separated audio data using a text independent speaker identification model to determine whether a verified and/or registered user spoke a spoken utterance captured in the audio data. Various implementations include processing the audio data and/or the separated audio data using an automatic speech recognition model to generate a text representation of the utterance.

Abstract: Techniques disclosed herein are directed towards streaming keyphrase detection which can be customized to detect one or more particular keyphrases, without requiring retraining of any model(s) for those particular keyphrase(s). Many implementations include processing audio data using a speaker separation model to generate separated audio data which isolates an utterance spoken by a human speaker from one or more additional sounds not spoken by the human speaker, and processing the separated audio data using a text independent speaker identification model to determine whether a verified and/or registered user spoke a spoken utterance captured in the audio data. Various implementations include processing the audio data and/or the separated audio data using an automatic speech recognition model to generate a text representation of the utterance.

Abstract: A centrifuge rotor includes tube retainers for holding two or more sample tubes at a 90 degree angle in a minimized configuration for effective separation of a blood sample using a mini centrifuge. The centrifuge rotor with sample tubes in place has an overall width that is less than two times the length of a sample tube.

Abstract: A method includes receiving, as input to a speech recognition model, audio data corresponding to a spoken utterance. The method also includes performing, using the speech recognition model, speech recognition on the audio data by, at each of a plurality of time steps, encoding, using an audio encoder, the audio data corresponding to the spoken utterance into a corresponding audio encoding, and decoding, using a speech recognition joint network, the corresponding audio encoding into a probability distribution over possible output labels. At each of the plurality of time steps, the method also includes determining, using an intended query (IQ) joint network configured to receive a label history representation associated with a sequence of non-blank symbols output by a final softmax layer, an intended query decision indicating whether or not the spoken utterance includes a query intended for a digital assistant.

Abstract: An automated speech recognition (ASR) model includes a first encoder, a first encoder, a second encoder, and a second decoder. The first encoder receives, as input, a sequence of acoustic frames, and generates, at each of a plurality of output steps, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The first decoder receives, as input, the first higher order feature representation generated by the first encoder, and generates a first probability distribution over possible speech recognition hypotheses. The second encoder receives, as input, the first higher order feature representation generated by the first encoder, and generates a second higher order feature representation for a corresponding first higher order feature frame. The second decoder receives, as input, the second higher order feature representation generated by the second encoder, and generates a second probability distribution over possible speech recognition hypotheses.

Abstract: A computer-implemented method includes receiving a sequence of acoustic frames corresponding to an utterance and generating a reference speaker embedding for the utterance. The method also includes receiving a target speaker embedding for a target speaker and generating feature-wise linear modulation (FiLM) parameters including a scaling vector and a shifting vector based on the target speaker embedding. The method also includes generating an affine transformation output that scales and shifts the reference speaker embedding based on the FiLM parameters. The method also includes generating a classification output indicating whether the utterance was spoken by the target speaker based on the affine transformation output.

Abstract: Techniques disclosed herein are directed towards streaming keyphrase detection which can be customized to detect one or more particular keyphrases, without requiring retraining of any model(s) for those particular keyphrase(s). Many implementations include processing audio data using a speaker separation model to generate separated audio data which isolates an utterance spoken by a human speaker from one or more additional sounds not spoken by the human speaker, and processing the separated audio data using a text independent speaker identification model to determine whether a verified and/or registered user spoke a spoken utterance captured in the audio data. Various implementations include processing the audio data and/or the separated audio data using an automatic speech recognition model to generate a text representation of the utterance.

Abstract: A method includes receiving a sequence of acoustic frames characterizing one or more utterances as input to a multilingual automated speech recognition (ASR) model. The method also includes generating a higher order feature representation for a corresponding acoustic frame. The method also includes generating a hidden representation based on a sequence of non-blank symbols output by a final softmax layer. The method also includes generating a probability distribution over possible speech recognition hypotheses based on the hidden representation generated by the prediction network at each of the plurality of output steps and the higher order feature representation generated by the encoder at each of the plurality of output steps. The method also includes predicting an end of utterance (EOU) token at an end of each utterance. The method also includes classifying each acoustic frame as either speech, initial silence, intermediate silence, or final silence.

Abstract: Techniques disclosed herein are directed towards streaming keyphrase detection which can be customized to detect one or more particular keyphrases, without requiring retraining of any model(s) for those particular keyphrase(s). Many implementations include processing audio data using a speaker separation model to generate separated audio data which isolates an utterance spoken by a human speaker from one or more additional sounds not spoken by the human speaker, and processing the separated audio data using a text independent speaker identification model to determine whether a verified and/or registered user spoke a spoken utterance captured in the audio data. Various implementations include processing the audio data and/or the separated audio data using an automatic speech recognition model to generate a text representation of the utterance.

Abstract: Techniques disclosed herein are directed towards streaming keyphrase detection which can be customized to detect one or more particular keyphrases, without requiring retraining of any model(s) for those particular keyphrase(s). Many implementations include processing audio data using a speaker separation model to generate separated audio data which isolates an utterance spoken by a human speaker from one or more additional sounds not spoken by the human speaker, and processing the separated audio data using a text independent speaker identification model to determine whether a verified and/or registered user spoke a spoken utterance captured in the audio data. Various implementations include processing the audio data and/or the separated audio data using an automatic speech recognition model to generate a text representation of the utterance.

Abstract: Two-pass automatic speech recognition (ASR) models can be used to perform streaming on-device ASR to generate a text representation of an utterance captured in audio data. Various implementations include a first-pass portion of the ASR model used to generate streaming candidate recognition(s) of an utterance captured in audio data. For example, the first-pass portion can include a recurrent neural network transformer (RNN-T) decoder. Various implementations include a second-pass portion of the ASR model used to revise the streaming candidate recognition(s) of the utterance and generate a text representation of the utterance. For example, the second-pass portion can include a listen attend spell (LAS) decoder. Various implementations include a shared encoder shared between the RNN-T decoder and the LAS decoder.

Abstract: A resonant power conversion apparatus including a transformer-based resonant converter and first and second switch control units is provided. The transformer-based resonant converter includes a primary switch circuit and a secondary output circuit configured to provide an output voltage to a load. The first switch control unit is configured to control an ON/OFF operation of the primary switch circuit in response to a status of the load. The second switch control unit is configured to determine whether to activate or inactivate the first switch control unit. When the status of the load is the light-loading or the no-loading, the first switch control unit intermittently controls the ON/OFF operation of the primary switch circuit, and meanwhile, the first switch control unit is inactivated during the primary switch circuit is disabled, so as to substantially reduce the light-loading or no-loading loss of the resonant power conversion apparatus.

Abstract: An AC-to-DC conversion apparatus is provided, and which includes a first switch-element, an output capacitor and a bridgeless power-factor-correction (PFC) circuit. The bridgeless PFC circuit is coupled to an AC input, and includes a first inductor, a second inductor and a bridge circuit constructed by second to fifth switch-elements. The first switch-element is connected between bridgeless PFC circuit and the output capacitor. Under such circuit configuration and suitable control manner, the common-mode interference in the provided AC-to-DC conversion apparatus is lowered and thus reducing the power loss.

Abstract: A bootstrap gate driver including a load indication unit, a bootstrap gate-drive unit and a drive-control unit is provided. The load indication unit is configured to generate a load indication signal in response to a state of a load. The bootstrap gate-drive unit is configured to drive a switch-transistor circuit in response to an inputted pulse-width-modulation (PWM) signal, wherein the switch-transistor circuit has a high-side driving path and a low-side driving path. The drive-control unit is coupled to the load indication unit and the bootstrap gate-drive unit, and configured to enable or disable the high-side driving path in response to the load indication signal. In the invention, the operation of the low-side driving path is not affected by enabling or disabling the high-side driving path.

Abstract: A resonant converter equipped with a phase shifting output circuit includes a resonant circuit to receive input power and regulate to become at least one resonant power, a switch unit to switch an ON period for the input power to pass through the resonant circuit and a power transformation circuit to regulate the resonant power and output a transformed power. The resonant converter further has a primary output circuit and at least one secondary output circuit. The primary output circuit regulates the transformed power to become a primary output power. A resonant control unit captures a feedback signal from the primary output circuit and generates a resonant control signal. A phase shifting control unit receives the resonant control signal and regulate to become a phase shifting driving signal. The secondary output circuit is controlled by the phase shifting driving signal and provides a secondary output power.

Abstract: A resonant power conversion apparatus including a transformer-based resonant converter and first and second switch control units is provided. The transformer-based resonant converter includes a primary switch circuit and a secondary output circuit configured to provide an output voltage to a load. The first switch control unit is configured to control an ON/OFF operation of the primary switch circuit in response to a status of the load. The second switch control unit is configured to determine whether to activate or inactivate the first switch control unit. When the status of the load is the light-loading or the no-loading, the first switch control unit intermittently controls the ON/OFF operation of the primary switch circuit, and meanwhile, the first switch control unit is inactivated during the primary switch circuit is disabled, so as to substantially reduce the light-loading or no-loading loss of the resonant power conversion apparatus.

Abstract: An AC-to-DC conversion apparatus is provided, and which includes a first switch-element, an output capacitor and a bridgeless power-factor-correction (PFC) circuit. The bridgeless PFC circuit is coupled to an AC input, and includes a first inductor, a second inductor and a bridge circuit constructed by second to fifth switch-elements. The first switch-element is connected between bridgeless PFC circuit and the output capacitor. Under such circuit configuration and suitable control manner, the common-mode interference in the provided AC-to-DC conversion apparatus is lowered and thus reducing the power loss.

Abstract: A bootstrap gate driver including a load indication unit, a bootstrap gate-drive unit and a drive-control unit is provided. The load indication unit is configured to generate a load indication signal in response to a state of a load. The bootstrap gate-drive unit is configured to drive a switch-transistor circuit in response to an inputted pulse-width-modulation (PWM) signal, wherein the switch-transistor circuit has a high-side driving path and a low-side driving path. The drive-control unit is coupled to the load indication unit and the bootstrap gate-drive unit, and configured to enable or disable the high-side driving path in response to the load indication signal. In the invention, the operation of the low-side driving path is not affected by enabling or disabling the high-side driving path.