Abstract: A computing system including a processor configured to receive a plurality of marginal distribution samples and a copula support sample including a plurality of copula sample points. The processor may divide the copula support sample into copula sample blocks and divide each of the marginal distribution samples into marginal sample blocks. For each of the copula sample blocks, within each copula dimension, the processor may assign a respective copula value rank to each sampled copula value included in that copula sample block. For each of the marginal sample blocks, the processor may sort the sampled marginal values to match an order of the copula value ranks of the corresponding copula sample block. The processor may generate a plurality of joint distribution sample vectors that each include the sampled marginal values located at corresponding positions across the marginal distribution samples. The processor may output the joint distribution sample vectors.

Abstract: According to examples, an apparatus may include a processor that may determine that a first sound to be outputted responsive to an occurrence of a first event on the apparatus is to be modified, in which the first sound corresponds to a first sound data. The processor may also select, from a pool of sound data, a second sound data that is to be combined with the first sound data, in which the second sound data is associated with an occurrence of a second event on the apparatus. The processor may further combine the first sound data and the second sound data to generate a modified sound data and may use the modified sound data to output a modified sound corresponding to the modified sound data responsive to an occurrence of the first event.

Abstract: Systems and methods for representing two-dimensional representations as three-dimensional avatars are provided herein. In some examples, one or more input video streams are received. A first subject, within the one or more input video streams, is identified. Based on the one or more input video streams, a first view of the first subject is identified. Based on the one or more input video streams, a second view of the first subject is identified. The first subject is segmented into a plurality of planar object. The plurality of planar objects are transformed with respect to each other. The plurality of planar objects are based on the first and second views of the first subject. The plurality of planar objects are output in an output video stream. The plurality of planar objects provide perspective of the first subject to one or more viewers.

Abstract: In examples, a notification is presented to a user of a computing device (e.g., based on identifying a trigger associated with a current context or as a result of receiving a push notification). The notification includes notification content and a dismiss user interface (UI) control, which indicates a status of an auto-dismiss timer associated with the notification. If user interaction associated with the notification is identified, the auto-dismiss timer may be paused. The user may actuate the notification to perform an associated action or may actuate the dismiss UI control to manually dismiss the notification (e.g., prior to expiration of the auto-dismiss timer). The notification may further include a snooze UI control, which, when actuated, may snooze the notification and/or future notifications. Thus, such aspects may offer an improved and consistent user experience for notifications associated with an application, a group of applications, or an operating system, among other examples.

Abstract: A computing system uses Advanced Encryption Standard XEX Based Tweaked Codebook Mode with Ciphertext Stealing (AES-XTS) encryption to encrypt a block of data using a tweak key, a data key, a modified tweak value, and the block of data to thereby generate an encrypted block of data. The modified tweak value is computed according to the expression DEC(0, CONST KEY), where DEC is an AES decryption algorithm, and CONST KEY is the tweak key. The encrypted block of data is thereby formatted according to the Advanced Encryption Standard with no extended mode and not according to the XEX Based Tweaked Codebook Mode with Ciphertext Stealing.

Abstract: The description relates to hinged devices, such as hinged computing devices. One example can include a first portion extending from a hinge end to a distal end that includes a first magnet and a second portion extending from a hinge end to a distal end and includes a second magnet. A translation mechanism can cause a button to both send control signals for electronic components of the device and create a force that acts on the first magnet to affect magnetic attraction between the first and second magnets.

Abstract: A method for training a neural network for facial expression recognition includes recognizing a plurality of digital human face models. For each of the plurality of digital human face models, a plurality of simulated facial expressions are simulated. Simulated capacitance measurements for an array of simulated radio frequency (RF) antennas are found for each of the plurality of simulated facial expressions. The simulated capacitance measurements for each simulated facial expression are provided as input training data to a neural network configured to output facial expression parameters based on input capacitance measurements.

Abstract: Examples are disclosed that relate to determining a capacitance based on a charge accumulated on a sense capacitor electrode. One example provides a head-mounted device comprising a facial-tracking sensor, a controller, and a charge sensing circuit connected to the facial-tracking sensor. The facial-tracking sensor comprises a sense capacitor electrode configured to be positioned proximate to a surface of a face and form a capacitance based upon a distance between the sense capacitor electrode and the surface of the face, and the controller configured to apply a reference voltage to the sense capacitor electrode. The charge sensing circuit is configured to determine the capacitance of the sense capacitor electrode by determining an amount of charge accumulated on the sense capacitor electrode resulting from the reference voltage.

Abstract: Examples of the present disclosure describe systems and methods for video genre classification. In one example implementation, video content is received. A plurality of sliding windows of the video content is sampled. The plurality of sliding windows comprises audio data and video data. The audio data is analyzed to identify a set of audio features. The video data is analyzed to identify a set of video features. The set of audio features and the set of video features is provided to a classifier. The classifier is configured to detect a genre for the video content using the set of audio features and the set of video features. The video content is indexed based on the genre.

Abstract: Example aspects include techniques for real-time detection of metric baseline behavior change. These techniques may include generating a reference distance signature based on historic time series information for a component metric, the historic time series information corresponding to a first period of time, generating a sample distance signature based on sample time series information for the component metric, the sample time series information corresponding to a second period of time, and comparing the reference distance signature to the sample distance signature to determine a signature difference. In addition, the techniques may include determining that the second period of time is a baseline change candidate based on the signature difference being greater than a distance threshold, and presenting, based at least in part on the signature difference, an alert notification identifying the second period of time as the baseline change candidate.

Abstract: Software-based authentication techniques may use mutable authentication data that changes to deter counterfeiters from making clones of authentic clients. An authentication service may issue mutable authentication data to a client and store the mutable authentication data in association with an identifier of the client. The client may authenticate by presenting the most recently issued mutable authentication data to the authentication service. Upon successful authentication, the authentication service may issue updated mutable authentication data to the client. Any clones that cannot present the most recent mutable authentication data may be rejected by the authentication service. By changing the mutable authentication data, counterfeiters cannot make a large number of clones that all work at the same time.

Abstract: Systems, methods, and computer-readable storage devices are disclosed for optimizing speech enhancement components to use in speech communication systems using non-intrusive speech quality assessment. One method including: receiving audio data, the audio data including speech; and the audio data having been processed by at least one speech enhancement component; detecting a first quality of the speech of the audio data using a trained non-intrusive speech quality assessment (NISQA) model, the trained NISQA model trained to detect quality of speech automatically; and changing one or more of the at least one speech enhancement component based on the detected first quality of the speech.

Abstract: Systems and methods are provided that are directed to generating video sequences including physio-realistic avatars. In examples, an albedo for an avatar is received, a sub-surface skin color associated with the albedo is modified based on physiological data associated with physiologic characteristic, and an avatar based on the albedo and the modified sub-surface skin color is rendered. The rendered avatar may then be synthesized in a frame of video. In some examples, a video including the synthesized avatar may be used to train a machine learning model to detect a physiological characteristic. The machine learning model may receive a plurality of video segments, where one or more of the video segments includes a synthetic physio-realistic avatar generated with the physiological characteristic. The machine learning model may be trained using the plurality of video segments. The trained model may be provided to a requesting entity.

Abstract: Embodiments described herein are directed to managing device compliance for devices that are connected to an enterprise network. For example, a mobile device manager may provide configuration settings to a computing device, which implements the settings in order to be compliant with an enterprise's data and/or security policy. The mobile device manager also maintains a local reference of each device's configuration settings implemented thereby. When the mobile device manager subsequently performs a determination as to whether the computing device is still in compliance, the mobile device manager simply needs to refer to the local reference to determine the computing device's settings instead of explicitly querying the computing device for its settings.

Abstract: The description relates to input devices that can be secured to and used to control various devices. One example input device can include a housing defining a top surface that includes controls. The housing can define a recess that includes generally opposing upper and lower surfaces that extends at least partially under and parallel to the top surface. A self-adjusting gripping mechanism can extend into the recess through the lower surface and can be biased toward the upper surface. The self-adjusting gripping mechanism can be configured to be contacted by devices having a range of thicknesses and configured to grip individual devices within the range of thicknesses between the self-adjusting gripping mechanism and the upper surface.

Abstract: Aspects of the present disclosure relate to headset virtual presence techniques. For example, a participant of a communication session may not have an associated video feed, for example as a result of a user preference to disable video communication or a lack of appropriate hardware. Accordingly, a virtual presence may be generated for such a non-video participant, such that the non-video participant may be represented within the communication session similar to video participants. The virtual presence may be controllable using a headset device, for example such that movements identified by the headset device cause the virtual presence to move. In some instances, user input may be received to control emotions conveyed by the virtual presence, for example specifying an emotion type and/or intensity.

Abstract: Data compression techniques are described for saving memory space by using fewer bits to store information while achieving high fidelity. A data set may be partitioned into a plurality of regions. Locally varying numerical ranges of data values (e.g., the minimum and maximum extents) may be determined for the plurality of regions. The data in the individual regions may be encoded using a lower number of bits as interpolation values in reference to the local extents rather than being encoded using a higher number of bits as absolute values. Where there are multiple channels of data in the regions, the number of available bits for encoding the data may be dynamically allocated per region based on the relative degrees of variance in data among the multiple channels.

Abstract: Methods and systems for estimating a number of people at a point of interest using vehicle data are provided. In some examples, vehicle data is received, from a plurality of vehicles, that corresponds to a geographic boundary encompassing a point of interest. From the vehicle data, a first subset of data corresponding to door sensor information is extracted. Based on the first subset of data, an estimated number of people within the geographic boundary, during a specified period of time, is determined. An indication corresponding to the estimated number of people within the geographic boundary is generated.

Abstract: Aspects of the present disclosure relate to storage policy change usage estimation. In examples, a base storage utilization is generated for a storage object according to a current storage policy. The current storage policy may comprise a RAID (“redundant array of independent disks”) level and/or a fault tolerance level with which the storage object is stored. An estimated storage utilization may then be generated for the storage object based on the base storage utilization and a new storage policy accordingly. Thus, the base storage utilization may omit or otherwise account for overhead associated with the RAID level and/or fault tolerance level indicated by the current storage policy, while the resulting estimated storage utilization may indicate an estimated physical storage utilization for the storage object if the new storage policy is implemented in place of the current storage policy.

Abstract: Systems and methods for selectively allowing or blocking operating system upgrades based on probabilities that the upgraded operating system may result in a fault condition. In an example, the technology may perform a method that includes receiving usage data from a client device to evaluate providing an operating system upgrade for the client device; accessing a root-cause factor that is likely to lead to a fault condition with the operating system upgrade; based on the usage data, identifying a fault probability for the root-cause factor; comparing the fault probability to a fault threshold for the root-cause factor; based on the comparison of the fault probability and the fault threshold, performing one of: blocking the operating system upgrade from being installed on the client device; or providing the operating system upgrade for installation on the client device.

Abstract: The performance of a neural network (NN) and/or deep neural network (DNN) can limited by the number of operations being performed as well as management of data among the various memory components of the NN/DNN. A sparsity-inducing regularization optimization process is performed on a machine learning model to generate a compressed machine learning model. A machine learning model is trained using a first set of training data. A sparsity-inducing regularization optimization process is executed on the machine learning model. Based on the sparsity-inducing regularization optimization process, a compressed machine learning model is received. The compressed machine learning model is executed to generate one or more outputs.

Abstract: Methods and systems for identifying vehicle filling stations and generating related visual indications are provided. In some examples, vehicle data is received that corresponds to a plurality of vehicles. From the vehicle data, a first subset of data is extracted that corresponds to location data and one of a fuel or charge level state change of each of the plurality of vehicles. One or more locations of one or more of the plurality of vehicles are identified, based on the location data, when the fuel or charge level state change corresponds to an increase in a fuel or charge level of the one or more of the plurality of vehicles. One or more visual indications are generated that correspond to the identified one or more locations.

Abstract: A method for evaluating gesture input comprises receiving input data for sequential data frames, including hand tracking data for hands of a user. A first neural network is trained to recognize features indicative of subsequent gesture interactions and configured to evaluate input data for a sequence of data frames and to output an indication of a likelihood of the user performing gesture interactions during a predetermined window of data frames. A second neural network is trained to recognize features indicative of whether the user is currently performing one or more gesture interactions and configured to adjust parameters for gesture interaction recognition during the predetermined window based on the indicated likelihood. The second neural network evaluates the predetermined window for performed gesture interactions based on the adjusted parameters, and outputs a signal as to whether the user is performing one or more gesture interactions during the predetermined window.

Abstract: Systems, methods, and computer-readable storage devices are disclosed for optimizing speech enhancement components to use in speech communication systems using non-intrusive speech quality assessment. One method including: receiving, from a computing device over a network, audio data, the audio data including speech; detecting a first quality of the speech of the audio data using a trained non-intrusive speech quality assessment (NISQA) model, the trained NISQA model trained to detect quality of speech automatically; determining whether the computing device is a low-quality endpoint based on the first quality of speech of the audio data; and transferring, from the computing device over the network, at least one speech enhancement component to at least one server device when the computing device is determined to be a low-quality endpoint.

Abstract: Systems and methods for transferring link context to a web browser. An example method includes receiving, by a web browser, a web address from a desktop application based on a selection of a link in document displayed in the desktop application; receiving, by a web browser from the application, link context for the selected link, the link context including an identification of the desktop application and an identification of the document; loading a web page corresponding to the received web address; based on the link context, identifying a browser-based application corresponding to the desktop application; launching the browser-based application within the web browser; transmitting the link context to the browser-based application; and displaying at least a portion of the document within the web browser.

Abstract: Systems and methods for opening files on a client device include receiving a request to open a file and identifying a file type pertaining to the file. A list of software applications is identified that are capable of opening the identified file type, and metadata pertaining to each of the software applications on the list is retrieved. User preference data pertaining to a user requesting to open the file is also retrieved. Scores for the software applications on the list are then determined based at least on the metadata and the user preference data. At least one software application on the list is then selected to open the file on the client device based on the determined scores.

Abstract: Systems and methods for providing colorblind-accessible versions of a color image by generating a color translation table for color vision deficiency (CVD) corrections are disclosed. For a given image, the pixel values as would be seen by a colorblind user are determined. Using these values, a corrected rendered image that appears to colorblind users as depicting a similar color scheme as the original image is produced and presented to the colorblind user. The proposed systems can be implemented with previously captured image data, real-time video, and as an add-on to other applications.

Abstract: A set of adversarial training examples for training a query interpreter are generated by: obtaining a target data table for a natural language query; identifying a primary entity of the target data table; for a target domain of the target data table, generating a set of candidate identifiers that are each semantically associated with an identifier of the target domain; for each candidate identifier, providing a premise-hypothesis pair to an NLI model to generate an entailment score; selecting a first subset of candidate identifiers from among the set of candidate identifiers based on the entailment score generated for each premise-hypothesis pair; for each candidate identifier of the first subset, applying the candidate identifier to an instance of the target data table as a table perturbation to generate a perturbed data table; and outputting each perturbed data table as part of an adversarial training example.

Abstract: In some examples, systems and methods for generating predicted ink strokes, using text-based semantics, are provided. Ink stroke data may be received, the ink stroke data may be input into a first model, and text data may be received from the first model. The text data may correspond to the ink stroke data. The text data and a semantic context may be input into a second model. A predicted ink stroke may be determined, from the second model. Further, an indication of the predicted ink stroke may be generated.

Abstract: In some examples, systems and methods for generating predicted ink strokes, using ink-based semantics, are provided. Ink stroke data may be received, the ink stroke data and a semantic context may be input into a model. From the model, one or more predicted ink strokes may be determined. Further, an indication of the one or more predicted ink strokes may be generated.

Abstract: In some examples, systems and methods for formatting ink are provided. Ink stroke data may be received, letters may be identified from the ink stroke data, and spacing may be identified between the letters. A user command and semantic context may be received. An action may be determined based on the user command, the semantic context, and the spacing between the letters. Further, the action may be performed.

Abstract: A method for selectively etching a metal component of a workpiece comprises: forming a hard mask over the metal component; and etching the metal component using an etchant solution, whereby the hard mask controls the etching; wherein the etchant solution is a basic etchant solution; and wherein the workpiece includes a semiconductor component comprising a material of Formula 1: InAsxSb1-x wherein x is in the range 0 to 1. It has surprisingly been found that basic etchants do not damage the semiconductor material. Another aspect provides the use of a basic etchant to etch aluminium selectively in the presence of a semiconductor comprising a material of Formula 1, where x is in the range 0 to 0.25.

Abstract: Systems and methods for deferring interactions with web data. An example computer-implemented method includes displaying, by a web browser, a web page; receiving, by a web browser, a selection of a quick note link displayed within the web browser; receiving a payload of the quick note link, the payload including note properties for generating a note; based on receiving the payload of the quick note link, launching a browser-based notes application within the web browser; transmitting the payload to the browser-based notes application to cause a new note to be created; and displaying the note within the web browser concurrently with the web page.

Abstract: Methods and systems for estimating a number of people at a point of interest using vehicle data are provided. In some examples, vehicle data is received, from a plurality of vehicles, that corresponds to a geographic boundary encompassing a point of interest. From the vehicle data, a first subset of data corresponding to door sensor information is extracted. Based on the first subset of data, an estimated number of people within the geographic boundary, during a specified period of time, is determined. An indication corresponding to the estimated number of people within the geographic boundary is generated.

Abstract: A computing system for pixel-based automated testing of a navigable simulated environment includes one or more processors configured to execute, in a run-time inference phase, an application testing program. The application testing program includes a machine learning model trained to detect errors in the application under test. A screen image of the simulated environment is processed by an object detection module to determine if a key object is present in the screen image. If a key object is present in the screen image, the application testing program executes an object investigation module to generate investigation inputs to investigate the key object. If a key object is not present in the screen image, the application testing program executes an environment exploration module to generate an environment exploration action to be provided to the application under test as simulated user input.

Abstract: Methods and systems for generating automation rules based on natural language inputs. In an example, the technology relates to a computer-implemented method for generating automation rules from natural language input. The method includes receiving a natural language input into a communications application for performing an action on communications received by the communications application; providing the natural language input into a trained machine learning model; receiving, as output from the trained machine learning model, a tagged primitive and an identified action from the natural language input; generating an automation rule for performing the action on a subset of communications received by the communications application, the subset of communications corresponding to the tagged primitives; and executing the generated automation rule to perform action on the subset of communications.

Abstract: An image sensor is disclosed. The image sensor includes a photo detector and a readout structure electronically coupled to the photodetector. The photodetector is configured to accumulate one or more photo charges responsive to one or more incident photons during an integration period. The readout structure is configured to repeatedly and nondestructively assess an amount of minority carrier photo charges accumulated at the photodetector during the integration period.

Abstract: This document relates to distributed teleconferencing. Some implementations can employ personalized enhancement models to enhance microphone signals for participants in a call. Further implementations can perform proximity-based mixing, where microphone signals received from devices in a particular room can be omitted from playback signals transmitted to other devices in the same room. These techniques can allow enhanced call quality for teleconferencing sessions where co-located users can employ their own devices to participate in a call with other users.

Abstract: A method of forming a thin film of material on a surface of a substrate, the substrate comprising a semiconductor, comprises: depositing a thin film of metal on the surface of the substrate, wherein the deposition is performed in an ultra-high vacuum, and wherein the substrate is at a temperature of less than or equal to 260 K during the deposition. Cooling the substrate during deposition of the thin film of metal may allow for an atomically flat and very uniform thin film to be obtained. Also provided is a device obtainable by the method.

Abstract: An apparatus comprising a crystal; at least three first wires arranged substantially parallel to each other; at least two second wires arranged substantially parallel to each other, each second wire arranged non-parallel to each first wire, such that each second wire crosses each first wire, forming at least six cross points; and at least six colour centres, each colour centre located within the crystal adjacent a different one of the at least six cross points.

Abstract: A system and method for preventing circumvention of content security measures includes receiving a request to display a content on a display screen and determining whether the content includes sensitive information. Upon determining that the content includes sensitive information, a request is transmitted to an operating system associated with the display screen to modify a plurality of display settings of the display screen from original display settings to modified display settings, the plurality of display settings including a screen refresh rate and a screen refresh line width, where the modified display settings prevent circumvention of content security through taking a photograph of the display screen.

Abstract: Techniques and solutions are described for executing a video processing task. A video processing task is received that includes one or more operations to be performed on a digital video file and an identifier of the digital video file. The video processing task is divided into subtasks of operations to be performed on fragments of the video, such as fragments having a particular duration. The duration can correspond to a duration used for video streaming Compared with video processing that is performed as a single task, disclosed techniques can provide improved fault tolerance, as only failed tasks need to be reprocessed. Video processing subtasks can be distributed to a plurality of workers, which can further improve fault tolerance, and can increase the computing power available for video processing, including allowing for the use of heterogenous or unreliable workers.

Abstract: A computing system including one or more processing devices configured to generate a training data set. Generating the training data set may include generating training molecular structures, respective training Hamiltonians, and training energy terms. Computing the training energy terms may include, for each of the training Hamiltonians, computing a kinetic energy term, a nuclear potential energy term, an electron repulsion energy term, and an exchange energy term using Hartree-Fock (HF) estimation. Computing the training energy terms may further include, for a first subset of the training Hamiltonians, computing dynamical correlation energy terms using coupled cluster estimation. Computing the training energy terms may further include, for a second subset of the first subset, generating truncated Hamiltonians and computing static correlation energy terms using complete active space (CAS) estimation.

Abstract: Techniques are described in which storage nodes within a storage cluster are configured to support graph-based storage management. For example, a storage cluster comprises a network and a plurality of computing systems, each interconnected over the network, wherein the plurality of computing systems includes a plurality of storage nodes. The computing system of the plurality of computing systems is configured to allocate a volume of storage within the storage cluster, generate a volume graph of the volume, wherein the volume graph represents one or more functional elements in a data plane of the volume, and manage the volume based on the volume graph.

Abstract: An apparatus comprising: a crystal; and a colour centre located within the crystal and adapted to receive: a magnetic field, such that a projection of the magnetic field onto a colour centre axis of the colour centre causes a change in an electron spin resonance frequency of the colour centre, and an electromagnetic signal having a frequency that substantially matches the electron spin resonance frequency of the colour centre as changed by the projection of the magnetic field onto the colour centre axis.

Abstract: Examples of the present disclosure describe systems and methods for on-device, in-browser AI processing. In examples, a selection of an AI pipeline is received. Content associated with the AI pipeline is also received. The content is segmented into multiple data segments and a set of data features is generated for the data segments. AI modules associated with the AI pipeline are loaded to create the AI pipeline. The set of data features is provided to the AI pipeline. The AI pipeline is executed to generate insights for the set of data features. The insights are then provided to a user.

Abstract: Systems and methods are provided for determining contextual objects related to web content. Web content may include one or more elements, which may be extracted from the web content and analyzed to identify any contextual objects. The web content elements may be hyperlink elements, and contextual objects may be identified based on the hyperlink elements. A dominant entity may be identified for the web content, and contextual objects relating to the dominant entity may be identified. When a request for web content is received, contextual objects may be provided with the requested content. Personalized contextual objects may be selected from the identified contextual objects based on data associated with the request for web content. Selection of a contextual object may result in navigation to a browsing environment associated with the selected contextual object.

Abstract: Aspects of the present disclosure provide systems and methods which utilizes machine learning techniques to provide enhanced accessibility features to a game. An accessibility service is provided which is capable of instantiating one or more machine learning models which can process current gameplay states and generate commands to assist users during gameplay. The accessibility commands may be provided to a game and used to supplement or modify user provided inputs in order to compensate for specific user needs. In further aspects, an accessibility user interface is provided which allows a user to dynamically enable or disable accessibility features during gameplay. The user interface is operable to receive accessibility selections and provide the selection data to an accessibility service during gameplay.

Abstract: The present concepts relate to contiguously writing data blocks of a data group in storage media in a one-step writing operation without defragmenting. A host may assign a group identifier to the data group, tag the data blocks in the data group with the group identifier, and send the data blocks to a storage controller. In turn, the storage controller may receive the data blocks that are tagged with the common group identifier, and contiguously write the data blocks in contiguous free space in the storage media. As such, reading the data group from the storage media may be performed faster than if the data group has been allowed to be written with fragmentation. The present concepts may also avoid the costs associated with a two-step writing operation that involves defragmenting.

Abstract: A system and method for securing a portion of a communication includes receiving a request to mark a portion of a communication as sensitive, where the communication includes a sensitive portion and an unsensitive portion. Upon receiving the request, a data property of the sensitive portion is set to a sensitive content property. After setting the data property, the method includes, securing the sensitive portion by initiating security calls for handling the sensitive portion, and transmitting the communication to a recipient. After the communication is transmitted to the recipient, access to the unsensitive portion is provided to the recipient, while the sensitive portion is redacted until the recipient is authenticated for receiving access to the sensitive portion.