Abstract: A head protection device for wrapping a shoulder belt and a length adjustment device of a carrier, the shoulder belt connected to a backrest and a seat of the carrier, and the shoulder belt wound around the length adjustment device, the length adjustment device slideable relative to the shoulder belt along a length extension direction of the shoulder belt. The head protection device includes a protection sheet, including a first part located between the length adjustment device and the backrest, and a limiting structure, arranged on the first part, wherein, the shoulder belt passes through the protection sheet twice in opposite directions at at least one position on the first part through the limiting structure, and the protection sheet is slidable relative to the shoulder belt along a length extension direction of the shoulder belt, the length adjustment device synchronously slides with the protection sheet on the shoulder belt.

Abstract: Embodiments of the present disclosure provide a method and system for positioning a robot and adjusting a posture. The method may include obtaining a first image and a second image of a target object. The first image may be captured using an image capturing apparatus, and the second image may be captured using a medical imaging device. The method may also include determining at least one target region corresponding to at least one target portion of the target object from the first image. The at least one target portion may be less affected by physiological motions than other portions. The method may further include determining positioning information of the robot based on the at least one target region and the second image.

Abstract: A computer-implemented method for generating a unified machine learning model using a neural network on a data processing apparatus is described. The method includes the data processing apparatus determining respective learning targets for each of a plurality of object verticals. The data processing apparatus determines the respective learning targets based on two or more embedding outputs of the neural network. The method also includes the data processing apparatus training the neural network to identify data associated with each of the plurality of object verticals. The data processing apparatus trains the neural network using the respective learning targets and based on a first loss function. The data processing apparatus uses the neural network trained to generate a unified machine learning model, where the model is configured to identify particular data items associated with each of the plurality of object verticals.

Abstract: The present disclosure provides a shoulder strap height adjusting mechanism including a seat pad and shoulder straps. The seat pad is provided with shoulder strap perforations through which the shoulder straps pass and shielding sheets located at the shoulder strap perforations. The shielding sheets shield the shoulder strap perforations and divide the shoulder strap perforations into an upper perforation and a lower perforation. A communication hole is formed between the two perforations, communicating the upper perforation and the lower perforation. Two shoulder strap perforations are arranged side by side on the seat pad, with two shielding sheets respectively located at two shoulder strap perforations. Two sliding straps are disposed in the two shoulder strap perforations and extend respectively along the height direction. The two shoulder straps pass through the two shoulder strap perforations respectively and are slidably connected to the corresponding sliding straps through fasteners.

Abstract: This application provides example electronic devices. One example electronic device includes a first display, a first sound-emitting area, a second sound-emitting area, and a second display. The first display includes a first display area and a second display area, and when the electronic device is in folded state, the first display area of the first display is opposite to the second display area of the first display. The first display area of the first display and the second display are stacked. The first display is disposed on a first surface of the electronic device, the second display is disposed on a second surface of the electronic device, the first surface is connected to the second surface through a bezel, the first sound-emitting area is provided on the first surface and/or the bezel, and the second sound-emitting area is provided on the second surface and/or the bezel.

Abstract: Techniques regarding generating a synthetic dataset of objects are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can include a defining component that can define a tractable forward process associated with a diffusion model, with defining the tractable forward process including inputting noise to compromise training data, resulting in compromised training data. The computer executable components can further include a training component that, using the compromised training data, trains the diffusion model to reverse process the tractable forward process, wherein the training results in a compromised diffusion model.

Abstract: According to embodiments of the disclosure, a method, an apparatus, a device and a storage medium for order processing are provided. The method includes: determining a group of orders associated with a target time window, the group of orders being uncompleted; requesting a business party associated with the group of orders to execute the group of orders; in response to receiving a message from the business party indicating that the group of orders are completed, querying an order database maintained by the business party to verify whether the group of orders are completed; an in response to the group of orders being verified as completed, updating a state of the group of orders to indicate that the orders are completed.

Abstract: In a configuration management process, a configuration commission request sent by a client to a server includes an identifier of a specified data source. The server, even if configuration data in a candidate configuration datastore comes from a plurality of data sources, does not commit all configuration data that comes from data sources other than the specified data source in the plurality of data sources to a running configuration datastore. That is, some configuration data that is not expected to be committed is not forced to be committed when configuration data from a data source needs to be committed.

Abstract: A computer-implemented method for generating a unified machine learning model using a neural network on a data processing apparatus is described. The method includes the data processing apparatus determining respective learning targets for each of a plurality of object verticals. The data processing apparatus determines the respective learning targets based on two or more embedding outputs of the neural network. The method also includes the data processing apparatus training the neural network to identify data associated with each of the plurality of object verticals. The data processing apparatus trains the neural network using the respective learning targets and based on a first loss function. The data processing apparatus uses the neural network trained to generate a unified machine learning model, where the model is configured to identify particular data items associated with each of the plurality of object verticals.

Abstract: A foldable display device includes a first frame, a second frame, a rotating shaft assembly, a flexible heat sink and a flexible circuit board. The flexible heat sink and the flexible circuit board passes through the rotating shaft assembly, and the rotating shaft assembly includes a shaft seat, a door plate, a first support plate, and a second support plate. Two opposite ends of the flexible heat sink are connected to the first frame and the second frame, and two opposite ends of a part of the flexible heat sink in the rotating shaft assembly are respectively fastened to a first fastening surface that is of the first support plate and that faces the shaft seat and a second fastening surface that is of the second support plate and that faces the shaft seat.

Abstract: The present disclosure provides an array substrate and a liquid crystal display panel, the array substrate includes a display area and a peripheral area surrounding the display area, the display area is divided into a plurality of heating zones, and each heating zone is provided with a heating electrode; the array substrate further includes a zone temperature control module and a heating wire module, the heating wire module includes a plurality of heating wire groups, the heating wire groups are arranged corresponding to the heating zones one to one, and a first terminal of each heating wire group is electrically connected with the zone temperature control module; the zone temperature control module is located in the peripheral area, and is configured to control each heating wire group to input a heating signal into the heating zone corresponding thereto.

Abstract: An electronic device includes a screen, a hinge assembly, a first housing, a second housing, and an elastic component. A material of an elastic portion of the elastic component is an elastic material or a non-Newtonian fluid material. A material of a support portion of the elastic component is a rigid material. The electronic device has an unfolded state and a folded state. When the electronic device is in the unfolded state, the elastic component is unfolded. The screen, the elastic component, and the hinge assembly are stacked in a thickness direction of the electronic device.

Abstract: A computer-implemented method for fine-grained referring expression comprehension is provided. The computer-implemented method includes receiving, at a processor, a textual expression and an image as inputs and executing, at the processor, fine-grained referring expression comprehension. The executing includes decomposing the textual expression into different textual modules, extracting visual regional proposals from the image, using language-guided graph neural networks to mine fine-grained object relations from the visual regional proposals and aggregating different matching similarities between the different textual modules and the fine-grained object relations.

Abstract: A method, computer system and computer program product is presented for providing a self-supervised speech representation. In one embodiment, audio input is received including speech utterances. A label sequence is generated from these speech utterances by a teacher label generator. A speech representation is generated of a partially masked version of the speech utterance using a speech representation network. The speech utterance is passed into two random transformations that alter only speaker information prior to the partial masking. A predictor will then predict the label sequence. In one embodiment performance-based assessment is made on a cross-entropy loss between the generated label sequence and a predicted label sequence.

Abstract: Some embodiments of the present disclosure are directed to systems, computer-readable media, and computer-implemented methods for neural network training. Some embodiments are directed to determining an attack order schedule for the data sample that includes a plurality of adversarial perturbation attacks associated with the data sample, and performing a composite adversarial attack process against the data set using the determined attack order schedule to generate a perturbed data sample for the data sample. Other embodiments may be disclosed or claimed.

Abstract: This application discloses a method for identifying a source address of a packet and an apparatus, and pertains to the field of network security. A protecting device intercepts a packet whose destination address is an IP address of a protected device, where the protected device provides a service according to the QUIC protocol. The protecting device sends a retry packet to a source address of the packet in response to that the packet is an initial packet. The protecting device identifies the source address of the packet as an attack source address if the protecting device receives no response packet corresponding to the retry packet. The protecting device determines the source address of the intercepted initial packet to identify an attack source address used by an attacker, and subsequently prevents only packets from the attack source address from being sent to the protected device.

Abstract: A method for constructing a design concept generation network (DCGN) and a method for automatically generating a conceptual scheme are provided. A DCGN includes a Transformer encoder, a Transformer decoder, an importance constraint matrix generation module, an importance constraint embedding layer, a cross-attention (CA) layer, and an optimization module. A word importance constraint is ingeniously introduced based on an attention mechanism of a Transformer to record input word constraint information contained in a generated text sequence. This can effectively ensure the reliability and effectiveness of a generated conceptual scheme and is conducive to capturing potential semantic importance information and implementing semantic knowledge reasoning.

Abstract: A cross-polarized dipole radiating element includes: (i) a substrate having a quad-arrangement of radiating arms on a forward-facing surface thereof, and a radio-frequency (RF) coupled loop on a rear-facing surface thereof, and (ii) a feed stalk electrically coupled to the quad-arrangement of radiating arms. The RF coupled loop can include a quad-arrangement of rear-facing coupling elements, which extend opposite corresponding ones of the quad-arrangement of radiating arms, and a quad-arrangement of metal traces, which are configured to electrically couple the quad-arrangement of coupling elements together into a closed electrical loop.

Abstract: Provided are a computer program product, system, and method for training a pose estimation model to determine anatomy keypoints in images. A teacher network, implementing machine learning, processes images representing anatomies to produce heatmaps representing keypoints of the anatomies. An anatomy parsing network, implementing machine learning, processes the images to produce segmentation representations labeling anatomies represented in the images. The segmentation representations from the anatomy parsing network and the heatmaps from the teacher network are concatenated to produce mixed heatmaps. A pose estimation model, implementing machine learning, is trained to process the images to output predicted heatmaps to minimize a loss function of the output predicted heatmaps from the pose estimation model and the mixed heatmaps.