Abstract: A method includes receiving an automated speech recognition (ASR) request from a user device that includes a speech input captured by the user device and content metadata associated with the speech input. The content metadata is generated by the user device. The method also includes determining a priority score for the ASR request based on the content metadata associated with the speech input and caching the ASR request in a pre-processing backlog of pending ASR requests each having a corresponding priority score. The pending ASR requests in the pre-processing backlog are ranked in order of the priority scores. The method also includes providing, from the pre-processing backlog, one or more of the pending ASR requests to a backend-side ASR module, wherein pending ASR requests associated with higher priority scores are processed before pending ASR requests associated with lower priority scores.

Abstract: Techniques and apparatuses are described that implement a smart-device-based radar system capable of performing angular position estimation. A machine-learned module analyzes complex range data generated to estimate angular positions of objects. The machine-learned module is implemented using a multi-stage architecture. In a local stage, the machine-learned module splits the complex range data into different range intervals and separately processes subsets of the complex range data using individual branch modules. In a global stage, the machine-learned module merges the feature data generated from the individual branch modules using a symmetric function and generates angular position data. By using machine-learning techniques and processing the complex range data directly, the radar system can achieve higher angular resolutions compared to other radar systems that utilize other techniques, such as analog or digital beamforming.

Abstract: A single E2E multitask model includes a speech recognition model and an endpointer model. The speech recognition model includes an audio encoder configured to encode a sequence of audio frames into corresponding higher-order feature representations, and a decoder configured to generate probability distributions over possible speech recognition hypotheses for the sequence of audio frames based on the higher-order feature representations. The endpointer model is configured to operate between a VAD mode and an EOQ detection mode. During the VAD mode, the endpointer model receives input audio frames, and determines, for each input audio frame, whether the input audio frame includes speech. During the EOQ detection mode, the endpointer model receives latent representations for the sequence of audio frames output from the audio encoder, and determines, for each of the latent representation, whether the latent representation includes final silence.

Abstract: A method includes receiving training data that includes a set of unspoken textual utterances. For each respective unspoken textual utterance, the method includes, tokenizing the respective textual utterance into a sequence of sub-word units, generating a first higher order textual feature representation for a corresponding sub-word unit tokenized from the respective unspoken textual utterance, receiving the first higher order textual feature representation generated by a text encoder, and generating a first probability distribution over possible text units. The method also includes training an encoder based on the first probability distribution over possible text units generated by a first-pass decoder for each respective unspoken textual utterance in the set of unspoken textual utterances.

Abstract: A method includes receiving training data that includes unspoken textual utterances in a target language. Each unspoken textual utterance not paired with any corresponding spoken utterance of non-synthetic speech. The method also includes generating a corresponding alignment output for each unspoken textual utterance using an alignment model trained on transcribed speech utterance in one or more training languages each different than the target language. The method also includes generating a corresponding encoded textual representation for each alignment output using a text encoder and training a speech recognition model on the encoded textual representations generated for the alignment outputs. Training the speech recognition model teaches the speech recognition model to learn how to recognize speech in the target language.

Abstract: Systems and techniques are provided for determination of user presence and absence using WiFi connections. Reports may be received from WiFi access points in an environment. The reports may include an identifier of a WiFi device, an indication of a connection to or disconnection from a WiFi access point, a time of the connection or disconnection, and an identifier of the WiFi access point. A connection sequence for the WiFi device may be generated from the reports. Whether the WiFi device is present in or absent from the environment as of a specified time may be determined based on the connection sequence. An indication of presence for a user associated with the WiFi device may generated if the WiFi device is present in the environment. An indication of absence for the user associated with the WiFi device may be generated if the WiFi device is absent from the environment.

Abstract: A speaker verification method includes receiving audio data corresponding to an utterance, processing the audio data to generate a reference attentive d-vector representing voice characteristics of the utterance, the evaluation ad-vector includes ne style classes each including a respective value vector concatenated with a corresponding routing vector. The method also includes generating using a self-attention mechanism, at least one multi-condition attention score that indicates a likelihood that the evaluation ad-vector matches a respective reference ad-vector associated with a respective user. The method also includes identifying the speaker of the utterance as the respective user associated with the respective reference ad-vector based on the multi-condition attention score.

Abstract: A method includes generating a first asset token that represents control, by a first entity, of a portion of a physical communication asset. The method includes publishing, to a distributed ledger, ownership of the first asset token and receiving, from a second entity, a request to control the portion of the physical communication asset represented by the first asset token. In response to receiving the request, the method includes removing the first asset token from circulation on the distributed ledger and generating a second asset token representing control, by the second entity, of the portion of the physical communication asset. The method also includes publishing, to the distributed ledger, ownership of the second asset token and transferring, using the distributed ledger, ownership of the second asset token to the second entity.

Abstract: A method includes obtaining a dataset that includes health data in a Fast Healthcare Interoperability Resources (FHIR) standard. The health data includes a plurality of healthcare events. The method includes generating, using the dataset, an events table that includes the plurality of healthcare events and is indexed by time and a unique identifier per patient encounter. The method also includes generating, using the dataset, a traits table that includes static data and is indexed by the unique identifier per patient encounter. The method includes training a machine learning model using the events table and the traits table and predicting, using the trained machine learning model and one or more additional healthcare events associated with a patient, a health outcome for the patient.

Abstract: A method for arbitrarily grouping computer system resources includes receiving a resource policy to be applied to a group of resources associated with a user. The resource policy includes a unique user-defined label assigned to each resource of the group of resources. The method also includes identifying the group of resources among a plurality of resources using the user-defined label. Here, the plurality of resources is associated with the user and distributed across a plurality of different computing environments. For each identified resource of the group of resources, the method also includes applying the resource policy to a corresponding identified resource by accessing the respective computing environment where the corresponding identified resource resides.

Abstract: Methods, systems, and apparatus for data analysis in a distributed computing system by accessing data stored at a first processing zone associated with a distributed data processing job, detecting information identifying a particular child job associated with the distributed data processing job, comparing the identifying information to data stored at a second processing zone, and identifying an additional child job as associated with the distributed data processing job based on a result of the comparison. The methods, systems and apparatus are further for correlating particular output data associated with the particular child job and additional output data associated with the additional child job for the distributed data processing job, determining performance data for the distributed data processing job based on the output data associated with each of the particular child job and the additional child job, and providing for display the performance data for the distributed data processing job.

Abstract: A method for training a sub-model for contextual biasing for speech recognition includes obtaining a base speech recognition model trained on non-biased data. The method includes obtaining a set of training utterances representative of a particular domain, each training utterance in the set of training utterances including audio data characterizing the training utterances and a ground truth transcription of the training utterance. The method further includes, for each corresponding training utterance in the set of training utterances, determining, using an embedding encoder, a corresponding document embedding from the ground truth transcription of the corresponding training utterance. The method includes training, using the corresponding document embeddings determined from the ground truth transcriptions of the set of training utterances, a sub-model to bias the base speech recognition model to recognize speech in the particular domain.

Abstract: The subject matter of the present disclosure can be implemented by, among other things, a remote control device configured to receive an identifier, present a plurality of controllable devices for selection on a display of the remote control device, and receive a selection of two or more controllable devices via the user interface. Using the identifier, a session with the two or more controllable devices is generated and performances of a same task by the two or more controllable devices are controlled by the remote control device by transmitting a message containing control information and the identifier to a network service.

Abstract: Techniques and apparatuses are described for resource block-level index modulation. In aspects, a wireless transmitter modulates a first portion of data for a wireless receiver to provide modulation symbols that correspond to the first portion of the data. The wireless transmitter also selects, based on a value of a second portion of the data, respective index locations for one or more resource blocks by which to transmit the modulation symbols. The wireless transmitter then transmits the modulation symbols to the wireless receiver using the one or more resource blocks having the respective index locations to convey the first portion of the data and the second portion of the data to the wireless receiver. By so doing, the wireless transmitter conveys the second portion of the data without using additional time-frequency resources of a communication channel, which can be useful when concurrently transmitting small amounts of data to many wireless receivers.

Abstract: This document describes systems and techniques directed at an external wide-angle lens for imagers in electronic devices. An imager is disclosed that includes an image sensor and a lens stack, the lens stack including an external wide-angle lens, an internal lens, and four or more intermediate lenses. The imager has a first ratio of a projection at a vertex of the external wide-angle lens divided by a maximum focused dimension of the focal area being less than or equal to 0.15, a second ratio of a total length of the lens stack divided by the maximum focused dimension being less than or equal to 7.0, or a third ratio of a total transmission length of the imager divided by an entrance pupil diameter of the external wide-angle lens being between 1.2 and 2.6.

Abstract: The described techniques and apparatuses enable memory-request priority up-leveling. A memory request is received over a virtual channel, VC, and is then added to a memory-request buffer with an original priority-level for the memory request and an indication that the memory request is associated with a virtual channel identification, VCID, of the VC. Related memory requests within the memory-request buffer are also indicated as being associated with the VCID. Responsive to determining that an up-level indication for the VCID is asserted over a side channel, the original priority-levels of the memory request, and other memory-requests in the memory-request buffer having an indication of the VCID, are increased to respective up-leveled priority levels. Responsive to determining that the up-level indication is no longer asserted, the up-leveled priority levels are returned to respective original priority-levels.

Abstract: This document describes techniques for flexible frequency band pairing for satellite communications. In aspects, a non-terrestrial communication system uses multiple frequency bands for a wireless link between a user equipment, UE, and a satellite of the non-terrestrial communication system. The non-terrestrial communication system determines to utilize two different frequency bands the wireless link between the satellite and the UE, the two different frequency bands being defined by a governing entity. In response, the non-terrestrial based communication system selects a first defined frequency band for downlink communications from the satellite to the UE and a second defined frequency band for uplink communications from the UE to the satellite. The non-terrestrial communication system then directs the satellite and the UE to communicate via the wireless link by using the first defined frequency band for the downlink communications and the second defined frequency band for the uplink communications.

Abstract: A method includes receiving a build request containing build step instructions from a user. The build step instructions specify a usage of containers within memory hardware for building an output container. The containers include at least one private container having private contents and/or at least one public container having public contents. The method also includes authenticating the user initiating the build request and determining whether the user is authorized to access the private containers. When the user is authenticated and authorized to access the private containers, the method includes obtaining the containers specified by the build step instructions from the memory hardware, executing the build step instructions to build the output container while using the received containers, and outputting the built output container.

Abstract: A method includes receiving follow-on audio data captured by an assistant-enabled device, the follow-on audio data corresponding to a follow-on query spoken by a user of the assistant-enabled device to a digital assistant subsequent to the user submitting a previous query to the digital assistant. The method also includes processing, using a speech recognizer, the follow-on audio data to generate multiple candidate hypotheses, each candidate hypothesis corresponding to a candidate transcription for the follow-on query and represented by a respective sequence of hypothesized terms. For each corresponding candidate hypothesis among the multiple candidate hypotheses, the method also includes determining a corresponding similarity metric between the previous query and the corresponding candidate hypothesis and determining a transcription of the follow-on query spoken by the user based on the similarity metrics determined for the multiple candidate hypotheses.

Abstract: This disclosure describes methods, devices, systems, and procedures in a computing system for capturing a configuration state of an operating system executing on a central processing unit (CPU), and offloading resource-related tasks, based on the configuration state, to a resource management unit such as a system-on-chip (SoC). The resource management unit identifies a status of each resource based on the captured configuration state of the operating system. The resource management unit then processes tasks associated with the status of the resources, such as modifying a clock rate of a clocked component in the computing system. This can alleviate the CPU from processing those tasks thereby improving overall computing system performance and dynamics.

Abstract: A method for multi-channel voice activity detection includes receiving a sequence of input frames characterizing streaming multi-channel audio captured by an array of microphones. Each channel of the streaming multi-channel audio includes respective audio features captured by a separate dedicated microphone. The method also includes determining, using a location fingerprint model, a location fingerprint indicating a location of a source of the multi-channel audio relative to the user device based on the respective audio features of each channel of the multi-channel audio. The method also includes generating an output from an application-specific classifier. The first score indicates a likelihood that the multi-channel audio corresponds to a particular audio type that the particular application is configured to process.

Abstract: This application is directed to a method for correct temperature measurement. An electronic device includes a temperature sensor that measures an ambient temperature of an environment and a display that is driven by a display driver. The electronic device determines a brightness setting of the display, estimates a display driver current based on the brightness setting, estimates a driver efficiency of the display driver based on the display driver current, and combines a predetermined display driver voltage, the display driver current, and the driver efficiency to determine a power consumption of the display driver. An ambient temperature correction is determined in accordance with the determined power consumption of the display driver, and the measured ambient temperature is thereby corrected using the ambient temperature correction. In some implementations, a power consumption of a distinct heat-generating electronic component is also monitored for adjusting the ambient temperature correction.

Abstract: Techniques and apparatuses are described for intra-user equipment-coordination set (intra-UECS) communication via an adaptive phase-changing device (APD) are described. In aspects, a base station selects an APD for use by a first user equipment-coordination set, UECS), in an intra-UECS communication path. The base station communicates APD information about the APD to a first coordinating user equipment, UE, of the first UECS. In aspects, the base station the apportions APD-access to the APD for the first UECS and indicates the apportioned APD-access to the first coordinating UE of the first UECS.

Abstract: In aspects, a non-terrestrial communication system communicates with a user equipment, UE, using repetitive communications. The non-terrestrial communication system determines (905, 940) a repetition configuration for repetitive communications with the UE and indicates (910, 915) the repetition configuration to the UE. The non-terrestrial communication system communicates (920) with the UE using the repetitive communications in accordance with the repetition configuration.

Abstract: A method includes obtaining a corpus of unlabeled training data including a plurality of spoken utterances, each corresponding spoken utterance of the plurality of spoken utterances includes audio data characterizing the corresponding spoken utterance. The method also includes receiving a target domain. The method also includes selecting, using a contrastive data selection model, a subset of the utterances from the corpus of unlabeled training data that correspond to the target domain. The method includes training an automatic speech recognition (ASR) model on the subset of utterances.

Abstract: This document describes systems and techniques for modulating credit allocations in memory subsystems. The described systems and techniques can provide a feedback mechanism to a credit controller to improve the bandwidth at a memory interface. The memory controller monitors statistics associated with transaction requests served to one or more random access memories (RAMs) of the memory subsystem. The memory controller can then provide suggestions to the credit controller or to the one or more clients to modulate the number of credits allocated to one or more clients. In this way, the described systems and techniques can improve the efficiency of the memory controller in managing the transaction requests and the bandwidth at the memory interface.

Abstract: Various arrangements for controlling multiple wireless audio devices with a Bluetooth device are presented herein. The Bluetooth device can receive a wireless audio device identifier from each wireless audio device of a plurality of wireless audio devices. The Bluetooth device can duplicate an audio stream into multiple audio sub-streams based on a number of wireless audio devices. The Bluetooth device can apply an audio effect to at least one audio sub-stream based on a characteristic of communication signals between the wireless communication circuit of the Bluetooth device and each wireless device of the plurality of wireless audio devices. The Bluetooth device can transmit each audio sub-stream to the respective wireless audio device for synchronized emission.

Abstract: A method includes, while a user device is using a first presentation mode to present content to a user, obtaining a current state of the user of the user device. The method also includes, based on the current state of the user, providing, as output from a user interface of the user device, a user-selectable option that when selected causes the user device to use a second presentation mode to present the content to the user. The method further includes, in response to receiving a user input indication indicating selection of the user-selectable option, initiating presentation of the content using the second presentation mode.

Abstract: This disclosure describes apparatuses, methods, and techniques that enable a computing device to support a dynamic range of audio quality, varying bandwidths, varying sampling rates, numerous effective number of bits (ENOBs) resolutions, conserve power during an overall usage of the computing device, and enhance a user experience. To do so, the computing device utilizes a reconfigurable analog-to-digital converter (ADC). The reconfigurable ADC includes a successive-approximation-register (SAR) ADC, a noise-canceling circuit, and a noise-shaping circuit. The reconfigurable ADC can selectively operate in different modes of operation, in part, by enabling or disabling the noise-canceling circuit and the noise-shaping circuit.

Abstract: In the presence of a network fault condition preventing communication between a communication device and a network device of a first network, a method includes receiving a heartbeat request from the communication device requesting permission for the communication device to perform a device operation. The method further includes obtaining device state data associated with the communication device and determining whether the communication device is permitted to perform the device operation based on the device state data associated with the communication device. When the communication device is permitted to perform the device operation, the method also includes transmitting a heartbeat confirmation signal to the communication device. The heartbeat confirmation signal permits the communication device to perform the device operation.

Abstract: Systems and techniques are described for implementing testing-and-manufacturing keys for a system-on-chip (SoC). A hardware test portion of the SoC is configured to exercise features of domains that process data being communicated across the fabrics during an externally initiated test. In response to receiving a testing-and-manufacturing token from an external test system, a testing-and-manufacturing key support component of the SoC generates a testing-and-manufacturing key. The hardware test portion is configured to execute a test function to promote security of the SoC, however, only in response to the testing-and-manufacturing security component authenticating the testing-and-manufacturing key. Through implementing testing-and-manufacturing keys this way, the system-on-chip secures access to potentially sensitive functions and secrets, while allowing their unencumbered and authorized access for testing the system-on-chip during various life cycle states.

Abstract: Implementations relate to automatic sharing of images with designated users over a communication network. In some implementations, a method includes causing display of images in a user interface on a first device, each of the images depicting a different person. The images are obtained from a library of images associated with a first user. A selection of a particular image of the images is received based on user input, and a person identifier is determined indicating a particular person depicted in the selected image. The person identifier is designated as a person sharing criterion. A first image not included in the images is obtained and programmatically analyzed to determine that the first image depicts a person that matches the person sharing criterion, and an access permission is updated to grant access to the first image to a second user of a second device over a communication network.

Abstract: Techniques and apparatuses are described that use transaction queue lengths to alter a clock frequency that controls access to a memory of an electronic device. Techniques include detecting that a transaction queue threshold has been violated, initiating a counter to measure a time duration, determining that the transaction queue threshold continues to be violated for the time duration and, in response, altering the clock frequency that controls access to the memory of the electronic device.

Abstract: A method for an in-memory distributed cache includes receiving a write request from a client device to write a block of client data in random access memory (RAM) of a memory host and determining whether to allow the write request by determining whether the client device has permission to write the block of client data at the memory host, determining whether the block of client data is currently saved at the memory host, and determining whether a free block of RAM is available. When the client device has permission to write the block of client data at the memory host, the block of client data is not currently saved at the memory host, and a free block of RAM is available, the write request is allowed and the client is allowed to write the block of client data to the free block of RAM.

Abstract: The present disclosure describes various aspects of adaptive sounding reference signal mapping that a user equipment (UE) implements to improve channel estimation. In aspects, a set of sounding reference signal (SRS) symbols are generated that include at least first and second SRS symbols. An offset for the second SRS symbol is determined based on a difference between a first radio chain and a second radio chain of the UE. The first and second SRS symbols are then mapped to an antenna port of the first radio chain. The UE transmits the first SRS symbol to a base station via the antenna port of the first radio chain and transmits the second SRS symbol to the base station via the antenna port of the first radio chain while the offset is applied to the first radio chain. By so doing, the UE may improve channel estimation for uplink and/or downlink communications.

Abstract: This disclosure describes apparatuses, methods, and techniques for enabling a user to safeguard a computing device with a fingerprint identification system by using biometric data. The fingerprint identification system includes a fingerprint sensor used during an enrollment process of the user's biometric data. The biometric data may include fingerprint data from the user's thumb, finger, a plurality of fingers, palm, and so forth. The computing device uses a collocation of a user's touch, for example, a thumb-tap, and a fingerprint sensor's location to guide the user to complete the enrollment process of a full fingerprint with ease and with fewer thumb-taps. Consequently, the techniques enable biometric security with an enrollment process having a good user experience.

Abstract: A method for recognizing persons in video streams includes obtaining a live video stream, detecting a first person in the live video stream, determining from analysis of the live video stream first information that identifies an attribute of the first person, determining based on at least some of the first information that the first person is not identifiable to the computing system, storing at least some of the first information, receiving a user classification of the first person as being a stranger, and deleting the stored first information.

Abstract: A method including: transmitting, by a control device, a first portion of content comprising a first portion of a signal corresponding to a multimedia presentation characteristic of a peripheral device; receiving, when an adjustment of the signal is below an adjustment threshold, a first instance of an input indicating a request to change the multimedia presentation characteristic; in response to receiving the first instance of the input, adjusting a second portion of the signal and transmitting a second portion of the content comprising the adjusted second portion of the signal; receiving, when the adjustment of the signal is at or above the adjustment threshold, a second instance of the input; and transmitting, in response to receiving the second instance of the input, a communication signal to the peripheral device to adjust a peripheral device control of an output of the multimedia presentation characteristic.

Abstract: A method includes receiving video data that includes a series of frames of image data. Here, the video data is representative of an actor performing an activity. The method also includes processing the video data to generate a spatial input stream including a series of spatial images representative of spatial features of the actor performing the activity, a temporal input stream representative of motion of the actor performing the activity, and a pose input stream including a series of images representative of a pose of the actor performing the activity. Using at least one neural network, the method also includes processing the temporal input stream, the spatial input stream, and the pose input stream. The method also includes classifying, by the at least one neural network, the activity based on the temporal input stream, the spatial input stream, and the pose input stream.

Abstract: A method includes obtaining a document with textual fields and a visual element. For each textual field, the method includes determining a textual offset for the textual field that indicates a location of the textual field relative to each other textual field in the document. The method includes detecting, using a machine learning vision model, the visual element and determining a visual element offset indicating a location of the visual element relative to each textual field in the document. The method includes assigning the visual element a visual element anchor token and inserting the visual element anchor token into the textual fields in an order based on the visual element offset and the respective textual offsets. The method also includes, after inserting the visual element anchor token, extracting, using a text-based extraction model, from the textual fields, structured entities representing the series of textual fields and the visual element.

Abstract: A method for managing cloud infrastructure includes receiving, from a user of a user device, a cloud infrastructure modification request requesting modification to cloud infrastructure. The cloud infrastructure modification request includes abstract configuration data derived from a user interaction with a graphical user interface (GUI) executing on the user device. The method includes translating the abstract configuration data into a configuration command. The configuration command describes a configuration of the cloud infrastructure. The method includes updating a configuration file with the configuration command. The configuration file includes one or more cloud infrastructure specifications for the cloud infrastructure and is controlled by a source control management system. The method includes provisioning, using the updated configuration file, the cloud infrastructure.

Abstract: A method includes receiving a series of writes that each requests respective data to be stored at a database. For each respective write, the method includes storing the respective data in a first buffer associated with a first data streaming application and a second buffer associated with a second data streaming application. The method includes transmitting the respective data to the database. The method includes receiving, from the database, confirmation that the respective data of the respective write has been committed to the database. The confirmation includes an order indicator indicating an order the respective data of the respective write was committed to the database relative to other writes in the series of writes. In response to receiving the confirmation, the method includes sending the respective data from the first buffer to the first data streaming application and from the second buffer to the second data streaming application.

Abstract: This document describes systems and techniques for narrow display borders using a display-to-frame interposer (“interposer”). In aspects, the interposer is bonded, on a first side, to a mechanical frame. On a second side, the interposer is bonded to a backside of a display panel module, including a folded portion, and supports a cover layer through the display panel module. In such a configuration, a display border, which extends from an edge of the cover layer to an active area of the display panel module and resides above the folded portion, can be reduced.

Abstract: This document describes systems and techniques for a non-contact vibration testing system for enhanced component placement. In one example a testing system includes a mounting device configured to receive an electronic device having a magnetic-field-sensitive component, the magnetic-field-sensitive component configured to vibrate in response to a variable-frequency magnetic field. A magnetic coil is configured to generate the variable frequency magnetic field in response to receiving an alternating electric current. The magnetic coil is disposed proximate to the mounting device to cause the variable-frequency magnetic field to propagate into a region of the mounting device in which the electronic device is configured to be received. The variable-frequency magnetic field configured to cause the magnetic-field-sensitive component to vibrate.

Abstract: This document describes improved driving circuits for piezoelectric speakers (piezos). In aspects, a charging case for a hearable device includes a piezo and a circuit for driving the piezo having a battery, an inductor, first and second switches, and a microcontroller. The microcontroller is configured to first toggle the first switch for a number of times to charge the piezoelectric speaker via the inductor to a target voltage level and second toggle the second switch to discharge and thus sound the piezo. In such aspects, the first toggling of the first switch and the second toggling of the second switch control a volume and frequency, respectively, of the sounding, improving audio quality and providing finder features to users. Further, charging the piezoelectric speaker via the inductor enables the inductor to be smaller, conserving space within and power consumption of the charging case, leading to improved battery life.

Abstract: This document describes systems and techniques for encoding a video frame using different compression ratios or compression algorithms for text blocks and non-text blocks. The described systems and techniques can determine, using a machine-learned model, which blocks of a frame include and do not include text. The described systems and techniques can then use a different compression ratio or compression algorithm for text blocks than the compression ratio or compression algorithm used for non-text blocks. For example, the systems and techniques can encode the text blocks using a first compression ratio that results in higher video quality than a second compression ratio used on at least some non-text blocks. In this way, the described systems and techniques can improve text legibility in a video file without significantly increasing the bandwidth requirements to transmit the video file to remote computing devices.

Abstract: Methods, systems, and apparatus for performing speech recognition. In some implementations, acoustic data representing an utterance is obtained. The acoustic data corresponds to time steps in a series of time steps. One or more computers process scores indicative of the acoustic data using a recurrent neural network to generate a sequence of outputs. The sequence of outputs indicates a likely output label from among a predetermined set of output labels. The predetermined set of output labels includes output labels that respectively correspond to different linguistic units and to a placeholder label that does not represent a classification of acoustic data. The recurrent neural network is configured to use an output label indicated for a previous time step to determine an output label for the current time step. The generated sequence of outputs is processed to generate a transcription of the utterance, and the transcription of the utterance is provided.

Abstract: Disclosed is a cavity-stacked printed circuit board (PCB) assembly that includes a first PCB formed of glass-reinforced epoxy material and has a first and a second side. The first side includes an open cavity with a cavity floor and at least one side wall that extends between the floor and the first side. The open cavity defines a cavity perimeter, which includes a base defined around the perimeter. The second PCB has a top side opposite a bottom side. The top side has an electrical component around which an interposer region is defined. Solder is disposed between the interposer region of the second PCB and the base of the first PCB to couple the first PCB to the second PCB with the electrical component received in the cavity to form the cavity-stacked PCB assembly.

Abstract: Embodiments are provided for managing the operation of sensors in an electronic device. According to certain aspects, the electronic device may detect a change in motion from a set of lower-sensitivity sensor data generated by a sensor(s) operating in a lower-sensitivity mode. When the change in motion is detected and during a timeout window, the sensor(s) may generate an additional set of lower-sensitivity sensor data and a set of higher-sensitivity sensor data. The electronic device may initially confirm the change in motion based on analyzing the set of higher-sensitivity sensor data. Further, the electronic device may determine that the additional set of lower-sensitivity does not indicate an additional change in motion, and may deem the confirmation of the change in motion as a false positive.

Abstract: A control system includes an energy management system in operation with intelligent, network-connected thermostats located in structures. The thermostats are operable to control heating, ventilation, and air conditioning (HVAC) systems. Control during a demand response (DR) event period may be performed based on an optimal setpoint schedule of the HVAC system. Candidate setpoint schedules may be simulated to identify an optimal schedule.