Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for model training and duration prediction. The method includes acquiring a first set of parameter values related to a first snapshot of a data object, the first snapshot being deleted from a storage system through a first deletion operation. The method further includes acquiring a first duration during which the first deletion operation is performed. The method further includes generating a prediction model based on at least the first set of parameter values and the first duration, the prediction model being used for determining a predicted duration required for deleting the snapshot from the storage system.

Abstract: This application discloses a motion vector obtaining method and apparatus, a computer device, and a storage medium. In the method, an initial motion vector of a to-be-processed picture block is determined by using a location relationship between a reference block and the to-be-processed picture block. When the reference block and the to-be-processed picture block are located in a same coding tree block, a decoder uses an initial motion vector of the reference block as the initial motion vector of the to-be-processed picture block. When the reference block and the to-be-processed picture block are located in different coding tree blocks, the decoder uses a final motion vector of the reference block as the initial motion vector of the to-be-processed picture block.

Abstract: A method for producing a hollow composite structure, such as a spar beam for a wind turbine blade, includes placing a membrane within a mold tool, the membrane being permeable to air and impermeable to resin. A mandrel is placed within the mold tool, the mandrel enclosed in an air tight layer that includes a vent. Fiber reinforcement material is placed around the mandrel within the mold tool and the membrane is sealed at least partly around the fiber reinforcement material and mandrel. The mold tool is closed with the vent line from the mandrel extending through the sealed membrane to outside of the mold tool. A vacuum is drawn in the mold tool while the mandrel is vented to outside of the mold tool, and while the vacuum is being drawn, resin is infused into the mold tool around the mandrel such that the resin is drawn towards the membrane.

Abstract: Methods and devices are provided for allowing a mobile device (e.g., a key fob or a consumer electronic device, such as a mobile phone, watch, or other wearable device) to interact with a vehicle such that a location of the mobile device can be determined by the vehicle, thereby enabling certain functionality of the vehicle. A device may include both RF antenna(s) and magnetic antenna(s) for determining a location of a mobile device relative to the vehicle. Such a hybrid approach can provide various advantages. Existing magnetic coils on a mobile device (e.g., for charging or communication) may be re-used for distance measurements that are supplemented by the RF measurements. Any device antenna may provide measurements to a machine learning model that determines a region in which the mobile device resides, based on training measurements in the regions.

Abstract: The disclosure provides a speech recognition method, a device and a computer-readable storage medium. The method includes obtaining a first voice signal collected from a first microphone in a microphone array and a second voice signal collected from a second microphone in the microphone array, the microphone array including at least two microphones, such as two, three or six microphones. The method further includes extracting enhanced features associated with the first voice signal and the second voice signal through a neural network, and obtaining a speech recognition result based on the enhanced features extracted.

Abstract: This application discloses a face image processing method performed by an electronic device. The method includes: acquiring a face image of a source face and a face template image of a template face; performing three-dimensional face modeling on the face image and the face template image to obtain a three-dimensional face image feature of the face image and a three-dimensional face template image feature of the face template image; fusing the three-dimensional face image feature and the three-dimensional face template image feature to obtain a three-dimensional fusion feature; performing face replacement feature extraction on the face image based on the face template image to obtain an initial face replacement feature; transforming the initial face replacement feature based on the three-dimensional fusion feature to obtain a target face replacement feature; and replacing the template face with the source face based on the target face replacement feature to obtain a target face image.

Abstract: An image processing method and apparatus are provided. During image matting on an original image, a plurality of segmented images including different regions are first obtained through semantic segmentation. Further, according to the segmented images, lines of different widths are drawn on a contour line of a foreground region to obtain a target trimap. Finally, a target image is generated based on the target trimap. For the target trimap, because lines of different widths are drawn on the contour line of the foreground region, pertinent image matting may be implemented for different regions, so that image matting precision may be improved for a region requiring fine image matting, and image matting precision of another region can also be ensured. In this way, a final image obtained through image matting is fine and natural.

Abstract: An image processing method and apparatus, an electronic device, and a storage medium are provided. The method comprises: obtaining a sliding trajectory inputted by a user on a display assembly; generating, according to the sliding trajectory, a first virtual model corresponding to the sliding trajectory; and displaying the first virtual model at a first position in a target image.

Abstract: An image processing method includes performing additional image feature extraction on a training source face image to obtain a source additional image feature, performing identity feature extraction on the training source face image to obtain a source identity feature, inputting a training template face image into an encoder in a to-be-trained face swapping model to obtain a face attribute feature, inputting the source additional image feature, the source identity feature, and the face attribute feature into a decoder in the face swapping model for decoding to obtain a decoded face image, obtaining a target model loss value based on an additional image difference between the decoded face image and a comparative face image, and adjusting the model parameters of the encoder and the decoder based on the target model loss value to obtain the trained face swapping model.

Abstract: A negative stiffness generating mechanism and a quasi-zero stiffness vibration isolator are provided. A housing is mounted on a base, and the axial relative positions of the housing and the base can be adjusted; a negative stiffness unit comprises inner-ring magnets, outer-ring magnets and a supporting shaft, the supporting shaft axially slides on the base and passes through the housing, the inner-ring magnets fixedly sleeve the supporting shaft, and the outer-ring magnets sleeve outside the inner-ring magnets and are divided into upper and lower groups of outer-ring magnets; the upper and lower groups of outer-ring magnets can synchronously move through a negative stiffness adjusting device; and the axial relative positions of the middle planes of the outer-ring and inner-ring magnets can be adjusted by adjusting the axial relative positions of the housing and the base. The isolator comprises a negative stiffness generating mechanism and a positive stiffness unit.

Abstract: Techniques provide for data synchronization. For example, such a technique may involve: obtaining respective synchronization characteristics of a group of synchronization jobs to be processed, each synchronization characteristic indicating at least one of an expected completion time instant and an amount of data to be synchronized of a corresponding synchronization job; prioritizing the group of the synchronization jobs based on the synchronization characteristics; and controlling execution of the group of the synchronization jobs based on a result of the prioritizing. Accordingly, high priority is given to the synchronization jobs which can be rapidly completed thereby improving the Recovery Point Objective (RPO) achievement rate before occurrence of a failure.

Abstract: An apparatus for housing a storage processor includes first and second side portions each extending in a plane defined by a first direction and a second direction, and separated from each other in a third direction; first and second flapper components disposed between the first side portion and the second side portion, extending in a first plane and a second plane, respectively, and rotatable about a first axis and a second axis, respectively, the first flapper component coupled to the first side portion and having a first end face, the second flapper component coupled to the second side portion and having a second end face; and first and second magnetic components disposed in the first end face and the second end face, respectively, and configured to be coupled to each other by a magnetic force so as to couple the first end face to the second end face.

Abstract: A system of fans ventilates heated air from within an IHS (Information Handling System), such as a rack-mounted server, when operated during normal conditions at a rated fan speed. A controller detects a failure of a fan of this fan system and identifies the functioning fans of the system. One or more of the functioning fans are selected for boosting by operation of a fan failure compensation circuit that has been configured for delivery of additional power to the selected boost fans. The fan failure compensation circuit delivers an output voltage that boost the airflow output of the system to compensate for the failed fan. By increasing the output voltage by approximately twenty percent, the boosted fans operate at approximately fifteen percent above rated speeds, which has been demonstrated to compensate for a failed fan while avoiding further failures during the expected lifespan of the fan system.

Abstract: The described aspects and implementations enable fast and accurate verification of radar detection of objects in autonomous vehicle (AV) applications using combined processing of radar data and camera images. In one implementation, disclosed is a method and a system to perform the method that includes obtaining a radar data characterizing intensity of radar reflections from an environment of the AV, identifying, based on the radar data, a candidate object, obtaining a camera image depicting a region where the candidate object is located, and processing the radar data and the camera image using one or more machine-learning models to obtain a classification measure representing a likelihood that the candidate object is a real object.

Abstract: A speaker and an electronic device are provided. The speaker includes a counterweight body, and a lower shell, a middle shell, an electric vibration plate, and a dome where the lower shell, the middle shell, and the electric vibration plate are matched to form an accommodating cavity, and the counterweight body is disposed in the accommodating cavity and is connected to a first surface of the electric vibration plate; the dome is disposed on a second surface of the electric vibration plate away from the counterweight body; and when a voltage is applied to the electric vibration plate, the electric vibration plate drives the dome to move.

Abstract: A multi-purpose labeling system can include a conveyor, a visual analysis module, and a labeling assembly. The conveyor can move an object in a first direction. The visual analysis module can include an optical sensor directed toward the conveyor to generate image data depicting the object. The labeling assembly can be spaced from the conveyor in a second direction and include a printer, a labeling module, and an alignment assembly. The printer can print a label based on the image data, and the labeling module can have a labeling plate for receiving the label. The alignment assembly can include a lateral-motion module, a vertical-motion module, and a rotary module for moving the labeling module along or about the first, the second, and a third direction, and can place the labeling plate adjacent to an object surface.

Abstract: A model training method and an apparatus thereof are provided. The method includes reading a portion of sample data in a sample full set to form a sample subset; mapping a model parameter related to the portion of sample data from a first feature component for the sample full set to a second feature component for the sample subset; and training a model based on the portion of sample data having the second feature component. A size of a copy of model parameters(s) on a sample computer can be reduced after mapping, thus greatly reducing an amount of training data and minimizing the occupancy of memory of the computer. Memory of a sample computer is used to place vectors, and store and load samples, thereby performing machine learning and training large-scale models with relatively low resource overhead under a condition of minimizing the loss of efficiency.

Abstract: The described aspects and implementations enable fast and accurate object identification in autonomous vehicle (AV) applications by combining radar data with camera images. In one implementation, disclosed is a method and a system to perform the method that includes obtaining a radar image of a first hypothetical object in an environment of the AV, obtaining a camera image of a second hypothetical object in the environment of the AV, and processing the radar image and the camera image using one or more machine-learning models MLMs to obtain a prediction measure representing a likelihood that the first hypothetical object and the second hypothetical object correspond to a same object in the environment of the AV.

Abstract: A picture decoding method includes: determining motion information of a to-be-decoded block; obtaining a first decoding prediction block of the to-be-decoded block based on the motion information; performing motion search with first precision in the prediction reference picture block to obtain at least two second decoding prediction blocks; performing downsampling on the first decoding prediction block and the at least two second decoding prediction blocks to obtain a first sampling pixel array and at least two second sampling pixel arrays, respectively; calculating a difference between the first sampling pixel array and each of the second sampling pixel arrays, and using, as a target prediction motion vector, a motion vector corresponding to a second sampling pixel array with a minimum difference; and obtaining a target decoding prediction block of the to-be-decoded block based on the target prediction motion vector, and decoding the to-be-decoded block based on the target decoding prediction block.

Abstract: Wireless communication between two electronic devices may be used to determine a distance between the two devices, even in the presence of an otherwise-disruptive attacker. A wireless receiver system of one device may receive a true wireless ranging signal from a first transmitting device and a false wireless ranging signal from an attacker. The wireless receiver system may correlate the wireless signals with a known preamble sequence and perform channel estimation using the result, obtaining a channel impulse response for the wireless signals. The wireless receiver system may filter the channel impulse response for the plurality of wireless signals by removing at least part of the channel impulse response due to the false wireless ranging signal while not removing at least part of the channel impulse response due to the true wireless ranging signal. The receiver system may perform a wireless ranging operation using the filtered channel impulse response.