Abstract: In a method an orchestrator of a computing system determines that results of Machine Learning model computations are available and dispatches a worker to perform model computations that include computing gradients of the results. The orchestrator determines that a set of gradients of the results is available and dispatches a gradient worker to compute a sum of the gradients. The orchestrator determines that a second set of gradients of the results is available and dispatches a second gradient worker to compute a sum of the second set of gradients. The orchestrator determines that the sums of the first and second gradients are available and dispatches a third gradient worker to compute synchronized gradients. The gradient workers compute the sums and synchronized gradients concurrent with training workers computing additional model computations results and/or gradients. A computer program product can include the method and a computing system can include the orchestrator.

Abstract: Described herein are systems, methods, and software to generate user interfaces to indicate software-defined storage information in view of available hardware resources. In one example, a user interface service may obtain enclosure information associated with enclosures in a computing environment and may determine a storage overview for the plurality of enclosures based on the enclosure information. The user interface service may further determine a function for each storage device in the storage overview in relation to a software-defined storage configuration for the computing environment and generate a user interface to indicate the storage overview with the function for each storage device in the plurality of enclosures.

Abstract: A system may determine a first code version including a first plurality of executable steps, the first plurality of executable steps including one or more triggers and one or more actions triggered by the one or more triggers. The system may determine a second code version including a second plurality of executable steps, the second plurality of executable steps including at least one change from the first plurality of executable steps. In some implementations, the system may detect the at least one change between the first plurality of executable steps and the second plurality of executable steps. The system may generate a graphical user interface including one or more graphical elements graphically indicating the at least one change and may provide the graphical user interface for display on a client device.

Abstract: The embodiments of the present invention relate to a non-destructive testing method for CICC superconducting cable damage estimation.

Abstract: A package includes a first and a second package component. The first package component includes a first metal trace and a second metal trace at the surface of the first package component. The second metal trace is parallel to the first metal trace. The second metal trace includes a narrow metal trace portion having a first width, and a wide metal trace portion having a second width greater than the first width connected to the narrow metal trace portion. The second package component is over the first package component. The second package component includes a metal bump overlapping a portion of the first metal trace, and a conductive connection bonding the metal bump to the first metal trace. The conductive connection contacts a top surface and sidewalls of the first metal trace. The metal bump is neighboring the narrow metal trace portion.

Abstract: An example virtualized computing system includes a cluster of hosts having a virtualization layer executing thereon and configured to manage virtual machines (VMs); a local storage device in a first host of the cluster, the local storage device configured as direct-attach storage; and an orchestration control plane, integrated with the virtualization layer, the orchestration control plane including a master server managing state of the orchestration control plane, the state including objects representing the hosts and the VMs, the orchestration control plane deploying a persistent application executing on a first VM of the VMs, the persistent application storing persistent data on the direct-attach storage.

Abstract: A microfluidic chip can include a microfluidic network that comprises a port, one or more test volumes, and one or more channels through which fluid must flow from the port to the test volume(s). A crosslinkable material can also be disposed within the microfluidic network such that the crosslinkable material is flowable through the channel(s). The crosslinkable material of the microfluidic chip may be exposed to light and/or heat to crosslink the material within and thereby occlude the channel(s).

Abstract: A myopia management soft contact lens design derive from analysis of overnight orthokeratology procedure that is to obtain the corneal height data before and after treatment according to reference position of cornea before and after treatment by mathematical calculation and then superimpose data with reference position of cornea, and then obtain corresponding central optical zone size and peripheral treatment zone size by obtaining the difference between corneal changes before and after treatment using corneal height data and tangential curvature data, and then calculate power change data before and after corneal treatment by using position of central optical zone and peripheral treatment zone to match the axial curvature data, the algorithm, and then use the power change data to distinguish different myopia power correction ranges to analyze the corresponding different power profile, and then apply this power profile to front curved surface of lens for the manufacture of soft contact lenses.

Abstract: A computing system is configured to generate a transformer-transducer-based deep neural network. The transformer-transducer-based deep neural network comprises a transformer encoder network and a transducer predictor network. The transformer encoder network has a plurality of layers, each of which includes a multi-head attention network sublayer and a feed-forward network sublayer. The computing system trains an end-to-end (E2E) automatic speech recognition (ASR) model, using the transformer-transducer-based deep neural network. The E2E ASR model has one or more adjustable hyperparameters that are configured to dynamically adjust an efficiency or a performance of E2E ASR model when the E2E ASR model is deployed onto a device or executed by the device.

Abstract: A troubleshooting technique provides faster and more efficient troubleshooting of issues in a distributed system, such as a distributed storage system provided by a virtualized computing environment. The distributed system includes a plurality of hosts arranged in a cluster. The troubleshooting technique uses cluster-wide correlation analysis to identify potential causes of a particular issue in the distributed system, and executes workflows to remedy the particular issue.

Abstract: A resonance conversion device and a universal serial bus circuit are provided. The resonance conversion device includes an input filter circuit, a full-bridge LLC converter circuit, a transformer circuit, a rectifier filter circuit, and a controller. The controller is configured to determine whether an indication voltage of a voltage command is less than or equal to a threshold value and to perform the following steps: in response to the indication voltage being less than or equal to the threshold value, controlling the full-bridge LLC converter circuit into a half-bridge operation mode, and regulating the DC output voltage by performing half-bridge burst mode control on the full-bridge LLC converter circuit based on the DC output voltage; and in response to the indication voltage being greater than the threshold value, regulating the DC output voltage by performing full-bridge burst mode control on the full-bridge LLC converter circuit based on the DC output voltage.

Abstract: A method, a system, and a device for full-waveform inversion deghosting for a marine variable depth streamer data acquisition are provided for solving existing problems that deghosted seismic data has low accuracy and is accompanied by artifacts due to a large error in ghost prediction. The provided method includes: acquiring seismic data, jointly solving Lippmann-Schwinger equations to obtain normal derivatives of an incident wave field and a wave field of a receiver surface, performing a wave field extrapolation by a Kirchhoff equation that includes only an integral on the receiver surface to obtain a wave field of a sea surface recorded by a horizontal streamer, calculating a ghost operator, and subjecting the ghosted wave field of the sea surface recorded by the horizontal streamer to full-waveform inversion deghosting to obtain deghosted seismic data. The provided method improves the accuracy and signal-to-noise ratio (SNR) of deghosted seismic data.

Abstract: The disclosure relates to processing application programming interface (API) requests. Embodiments include receiving, at an API wrapper, from a first caller, a first call to an API and sending the first call to the API. Embodiments include receiving, by the API wrapper, from one or more second callers, a second one or more calls to the API prior to receiving a response from the API to the first call. Embodiments include receiving, by the API wrapper, the response from the API to the first call and responding to the first call from the first caller with the response from the API to the first call. Embodiments include responding, by the API wrapper, to the second one or more calls from the one or more second callers with the response from the API to the first call without sending the second one or more calls to the API.

Abstract: The utility model discloses an encoder switch with side pressing function, which includes an encoder body and a touch switch arranged on the back of the encoder body, wherein the end of the rotating shaft of the encoder body extends to the end face of the touch switch. The encoder switch with the side press function provided by the utility model has the encoder function, and also gives the side press touch switch function, so that the encoder switch function is increased, the practicability is improved, and the application range is wider.

Abstract: Provided are a user verifying bicycle control system and a user verification method thereof, including a sensing module and a control module installed on a bicycle. The sensing module obtains a torque signal and an angle signal from a crank sensing component, further obtains a speed signal from a speed sensing unit, and outputs the torque signal, the angle signal, and the speed signal to the control module; the control module assembles the torque signal, the angle signal, and the speed signal into a signal sequence as a key to verify a user identity for the bicycle.

Abstract: A driving circuit for an electrowetting on dielectric (EWOD) pixel. The driving circuit includes a latch circuit for transmitting a source data pulse to a storage capacitor in response to an activation gate signal applied to the gate of the switch transistor and generating a latch voltage and an inversion circuit for outputting a driving voltage at either a first power voltage or a second power voltage based on the latch voltage generated by the latch circuit.

Abstract: Systems and methods for developing a classification model for classifying medical reports, such as radiology reports. One method includes selecting, from a corpus of reports, a training set and a testing set, assigning labels of a modality and an anatomical focus to the reports in both sets, and extracting a sparse representation matrix for each set based on features in the training set. The method also includes learning, with one or more electronic processors, a correlation between the features of the training set and the corresponding labels using a machine learning classifier, thereby building a classification model and testing the classification model on the reports in the testing set for accuracy using the sparse representation matrix of the testing set. The method further includes predicting, with the classification model, labels of an anatomical focus and a modality for remaining reports in the corpus not included in the sets.

Abstract: An electric vehicle charging management system has a dashboard apparatus, a vehicle controller, and a power battery apparatus. The dashboard apparatus stores multiple charging parameters that are adjustable. The multiple charging parameters include a charging power, a charging capacity, and a charging time frame. The vehicle controller is connected to the dashboard apparatus. The vehicle controller enters a charging operation mode when receiving a wake-up signal to read the multiple charging parameters from the dashboard apparatus and output the multiple charging parameters. The power battery apparatus receives a charging power source. The power battery apparatus has a control unit and an energy storage unit. The control unit is connected to the vehicle controller. When the control unit receives the multiple charging parameters from the vehicle controller, the control unit controls the charging power source to charge the energy storage unit according to the multiple charging parameters.

Abstract: The present disclosure describes an exemplary method that can prevent or reduce out-diffusion of Cu from interconnect layers to magnetic tunnel junction (MTJ) structures. The method includes forming an interconnect layer over a substrate that includes an interlayer dielectric stack with openings therein; disposing a metal in the openings to form corresponding conductive structures; and selectively depositing a diffusion barrier layer on the metal. In the method, selectively depositing the diffusion barrier layer includes pre-treating the surface of the metal; disposing a precursor to selectively form a partially-decomposed precursor layer on the metal; and exposing the partially-decomposed precursor layer to a plasma to form the diffusion barrier layer. The method further includes forming an MTJ structure on the interconnect layer over the diffusion barrier layer, where the bottom electrode of the MTJ structure is aligned to the diffusion barrier layer.