Abstract: The present disclosure selects third party content based on feedback. A selector identifies several content items including first and second content items (or more) responsive to a request. A machine learning engine determines a first feature of the first content item, a second feature of the second content item, and a third feature of the web page or a device associated with the request. The machine learning engine determines, responsive to the first feature and the third feature, a first score for the first content item based on a machine learning model generated using historical signals received from devices via a metadata channel formed from an electronic feedback interface. The machine learning engine determines a second score for the second content item responsive to the second feature and the third feature. A bidding module determines a price for the first content item based on the first and second scores.

Abstract: Systems and methods for validation of modeling and simulation systems that provide for the virtual fitting of wearable devices, such as glasses, by a user. Three-dimensional modeling and simulation of test subject both with and without fitting frames corresponding to a wearable device may be captured to validate the modeling and simulation modules and associated algorithms and machine learning modules used to simulate the fit of the wearable device on a user. Validation in this manner may provide for increased accuracy/realism of the modeling and simulation systems.

Abstract: Allocating network resources to one or more signals that are to be conveyed over the network by calculating a transport capacity for a sublink of the network based on a spectral efficiency of at least one subpath included in the sublink, and allocating the sublink to at least one signal based on the calculated transport capacity.

Abstract: This document describes systems and techniques for deriving identity and root keys for embedded systems. In aspects, a boot process and key manager of an embedded system may implement a secure or trusted boot process for embedded systems in which code of next-level boot loader or software image is verified using root keys or other protected information before execution of the boot process is passed to the next stage in the boot process. Alternatively or additionally, the key manager may enable sealing and attestation of various levels of root and identity keys to enable respective verification of software or hardware throughout a life cycle of a device to prevent unauthorized access to protected or private code of an embedded system. By so doing, the described aspects may enable an embedded system with a secure boot process and robust identity and root key management system.

Abstract: A recurrent neural network-transducer (RNN-T) model improves speech recognition by processing sequential non-blank symbols at each time step after an initial one. The model's prediction network receives a sequence of symbols from a final Softmax layer and employs a shared embedding matrix to create and map embeddings to each symbol, associating them with unique position vectors. These embeddings are weighted according to their similarity to their matching position vector. Subsequently, a joint network of the RNN-T model uses these weighted embeddings to output a probability distribution for potential speech recognition hypotheses at each time step, enabling more accurate transcriptions of spoken language.

Abstract: Systems and methods of the present disclosure can include a computer-implemented method for efficient machine-learned model training. The method can include obtaining a plurality of training samples for a machine-learned model. The method can include, for one or more first training iterations, training, based at least in part on a first regularization magnitude configured to control a relative effect of one or more regularization techniques, the machine-learned model using one or more respective first training samples of the plurality of training samples. The method can include, for one or more second training iterations, training, based at least in part on a second regularization magnitude greater than the first regularization magnitude, the machine-learned model using one or more respective second training samples of the plurality of training samples.

Abstract: A computerized method for associating cyberthreat actor groups responsible for different cyberthreats is described. The method involves generating a similarity matrix based on content from received clusters of cybersecurity information. Each received cluster of cybersecurity information is assumed to be associated with a cyberthreat. The similarity matrix is composed via an optimized equation combining separate similarity metrics, where each similarity metric of the plurality of similarity metrics represents a level of correlation between at least two clusters of cybersecurity information, with respect to a particular aspect of operations described in the clusters. The method further involves that, in response to queries directed to the similarity matrix, generating a listing of a subset of the clusters of cybersecurity information having a greater likelihood of being associated with cyberthreats caused by the same cyberthreat actor group.

Abstract: Systems and methods enable an entity to certify a web page address as being linked to the entity. The web page address includes semantic web mark-up identified attributes for the entity. A system may extract the attributes from the web page for the entity and use the attributes to generate an information card for the entity. The certification process ensures that the attributes are accurate, so that information cards generated for the entity are of high quality and reliable. Implementations may also simplify maintenance and quality assurances processes for an entity repository.

Abstract: Training and/or using both a high-level policy model and a low-level policy model for mobile robot navigation. High-level output generated using the high-level policy model at each iteration indicates a corresponding high-level action for robot movement in navigating to the navigation target. The low-level output generated at each iteration is based on the determined corresponding high-level action for that iteration, and is based on observation(s) for that iteration. The low-level policy model is trained to generate low-level output that defines low-level action(s) that define robot movement more granularly than the high-level action—and to generate low-level action(s) that avoid obstacles and/or that are efficient (e.g., distance and/or time efficiency).

Abstract: A mobile device includes a panel audio loudspeaker including a display panel and an actuator coupled to the display panel. The mobile device includes a temperature sensor arranged to sense a temperature of the display panel, and an electronic control module in communication with the actuator and the temperature sensor. The electronic control module is programmed to perform operations including: obtaining, from the temperature sensor, data indicating a temperature of the display panel; and based on the data indicating the temperature of the display panel, adjusting a power signal provided to the actuator to drive the panel audio loudspeaker. The power signal can be adjusted by selecting, based on the data indicating the temperature of the display panel, a target temperature of the display panel; mapping the target temperature to a target power level; and changing the power signal provided to the actuator to the target power level.

Abstract: An apparatus includes a directional coupler and an absorptive low pass filter, in which the directional coupler has a first transmission line extending from a first port to a second port and a second transmission line extending from a third port to a fourth port, the first transmission line and the second transmission line configured such that a portion of a signal travelling from the first port onto the first transmission line is coupled to the second transmission line and towards the third port. The second port is connected to the fourth port of the directional coupler via the absorptive low pass filter. When the signal is input into the first port of the directional coupler and output through the third port of the directional coupler, the signal is substantially unattenuated if the frequency of the signal is in a passband of the absorptive low pass filter.

Abstract: Techniques and apparatuses are described that enable a device to be unlocked and continuous user authentication without a touch input dedicated to fingerprint requisition. A touch input is received that comprises one or more touches to a touchscreen, and raw image data corresponding to the touches is retrieved from a fingerprint imaging sensor. A pixel-clustering technique is performed on the raw image data to determine a portion of the raw image data that corresponds to each of the touches. Touch embeddings are formed for each of the portions of the raw image data and compared to one or more stored fingerprint embeddings that correspond to respective fingerprints of one or more authorized users. An authentication result is then determined for the touch input based on the comparison results.

Abstract: Paxos transactions are pipelined in a distributed database formed by a plurality of replica servers. A leader server is selected by consensus of the replicas, and receives a lock on leadership for an epoch. The leader gets Paxos log numbers for the current epoch, which are greater than the numbers allocated in previous epochs. The leader receives database write requests, and assigns a Paxos number to each request. The leader constructs a proposed transaction for each request, which includes the assigned Paxos number and incorporates the request. The leader transmits the proposed transactions to the replicas. Two or more write requests that access distinct objects in the database can proceed simultaneously. The leader commits a proposed transaction to the database after receiving a plurality of confirmations for the proposed transaction from the replicas. After all the Paxos numbers have been assigned, inter-epoch tasks are performed before beginning a subsequent epoch.

Abstract: Examples for enhancing sensitivity to reflected GNSS signals are presented herein. An example may involve identifying, by a receiver, a particular positioning signal that reflected off a reflecting plane prior to reaching the receiver. The receiver may be in motion. The example may also involve determining a reflected satellite position for a satellite that transmitted the particular positioning signal based on identifying the particular positioning signal. The reflected satellite position may be determined by reflecting a position of the satellite about the reflecting plane. The example may also involve determining a direction vector to the reflected satellite position for the satellite and performing coherent integration over a threshold duration of time to increase a signal to noise ratio for the particular positioning signal based on the direction vector to the reflected satellite position.

Abstract: This document describes low-footprint dual-band ultra-wideband (UWB) antenna modules. A described UWB antenna module may be used as an internal part of a mobile device (e.g., cellphone, tablet, and/or other mobile devices). The UWB antenna module includes a multi-layer dual-band antenna that includes a set of multi-layer patch antennas, each patch antenna including a layer with a conductive ground plate, a feeding plate layer, and a parasitic strip layer with two parasitic strips, one configured to resonate at a frequency within a first band of the dual-band antenna, the other configured to resonate at a frequency within a second band of the dual-band antenna. The parasitic strips are electromagnetically coupled to the feeding plate.

Abstract: The technology provides a holistic approach to note-taking on client computing devices. An embedded app is integrated with an operating system-level user interface of a client device. The approach leverages system UI-level access to user data, supporting functionality including providing context for notes. This may include tagging a current context in a note, such as to indicate what the user was doing when taking the note and/or other information relating to how the note was triggered. Context can provide suggestions associated with a task when a user is interacting with a particular app or other program. The technology enables users to take notes quickly within their current task flow, via a system level tool. It offers users the opportunity to quickly jot down a thought and process it at a later time, while offering operating system-level easy access and also capturing the context from interactions that inspire the note.

Abstract: The present document describes systems and methods for on-device person recognition and provision of intelligent alerts. The system includes a decentralized multi-camera system for on-device facial recognition. A device (e.g., security camera, video doorbell) captures images/video of a person, processes input image frames, detects face images, filters static faces, and aligns a rotation of the face to be upright and frontal. The device then filters low-quality face images and/or images having a large portion of the face occluded. The device computes a face embedding, compares it against a set of locally stored reference embeddings, and sends matching results to cloud services, which, based on the matching result, notifies the device owner whether the observed person is a known person or a stranger. Face detection and recognition computations are performed on device, not at the cloud. No sensitive information is transmitted off device and privacy is thus preserved.

Abstract: The present document describes a battery configuration for reducing electromagnetic interference (EMI). The battery configuration includes a coin cell (e.g., stacking battery cell, button cell) with predefined external-tab configurations and a predefined internal-tab angle for reducing electromagnetic (EM) coupling. In particular, internal tabs are positioned to be separated by an angle (e.g., a predefined internal-tab angle) of approximately 90 degrees. External tabs include (i) a first external tab connected to a side of the coin cell and extending to overlap a top surface of the coin cell and (ii) a second external tab connected to the top surface of the coin cell. Both external tabs are positioned relative to the internal tabs to reduce an H-field and/or an E-field created by current running through the coin cell.

Abstract: Features described herein pertain to extending the functionality of an application executing on an electronic device. A device trait specification associated with a smart device can be used to generate an application extension template for defining a layout of a graphical user interface page of the application and binding at least one device trait of the set of device traits to the application. An application extension can be generated based on the application extension template being populated with layout defining values and binding values. The application extension can be provided to the application. The application extension can cause the graphical user interface page of the application to be displayed on a display of the electronic device with the defined layout and elements bound to the layout and configured to cause the smart device to be controlled based on the device traits.