Abstract: Systems and methods for object detection generate a feature pyramid corresponding to image data, and rescaling the feature pyramid to a scale corresponding to a median level of the feature pyramid, wherein the rescaled feature pyramid is a four-dimensional (4D) tensor. The 4D tensor is reshaped into a three-dimensional (3D) tensor having individual perspectives including scale features, spatial features, and task features corresponding to different dimensions of the 3D tensor. The 3D tensor is used with a plurality of attention layers to update a plurality of feature maps associated with the image data. Object detection is performed on the image data using the updated plurality of feature maps.

Abstract: Provided are an adjustable bracket and a high-feet sofa. In the adjustable bracket, the transmission rod is in a straight rod structure. A first end of the transmission rod is connected to the fourth connection rod. A second end of the transmission rod is connected to the pedal rod group. The present disclosure provides an adjustable bracket. In the case of the same sitting height, the adjustable mechanism of a sofa has a more compact structure so that the cleaning work of a mop, a vacuum cleaner, or a floor-sweeping robot is facilitated. Moreover, the height of the sofa from the ground is higher, making the sofa more attractive.

Abstract: Sensing circuit, detection control method, fingerprint identification module and display apparatus are provided. The sensing circuit comprises a photoelectric conversion element, an energy storage circuit, a switch control circuit, a discharge control circuit and a voltage conversion circuit. The photoelectric conversion element controls potential of first terminal of energy storage circuit according to received optical signal; the switch control circuit controls first terminal of energy storage circuit to be connected to or disconnected from conversion control node; the discharge control circuit controls conversion control node be connected to or disconnected from first voltage terminal; the voltage conversion circuit controls to output voltage signal outputted by output terminal according to second and/or third voltage signals.

Abstract: A sorting vehicle is provided that includes a wire harness structure and at least two vehicle bodies. The at least two vehicle bodies are pivotally connected to one another in sequence, the wire harness structure is connected to the two adjacent vehicle bodies, and the wire harness structure includes an elastic deformation portion located at a junction of the two adjacent vehicle bodies.

Abstract: An electronic device is provided. The electronic device may be a mobile phone, a wearable device, a vehicle-mounted device, or another device. The electronic device in this application uses a design in which a base protrudes in a Y direction or is flush with a shaft cover, thereby improving structural rigidity. The base can effectively bear all or a part of impact force, thereby effectively ameliorating a problem that the shaft cover is deformed and a flexible screen is impacted, and improving shatter resistance of the entire electronic device.

Abstract: The application discloses a processing method for graphite acid purification wastewater which comprises processing steps of neutralization and precipitation preprocessing, supernatant electrically distilling and vaporizing, electric energy recycling, condensate recycling and vaporization residue regularly discharging. The processing method performs neutralization and precipitation preprocessing to obtain neutralized supernatant, pumps the supernatant into an electric distiller to be vaporized and obtain electric energy for recycling through steam turbine, and acquires vaporization condensate by residual heat exchange. The vaporization condensate reaches the standard of graphite acid purification impurities such as fluorine, chlorine, aluminum, silicon, iron, magnesium, copper, and zinc and the wastewater discharge standard. The steam heat energy can be recycled and transformed into electric energy to save processing cost.

Abstract: A system is provided for measuring a power radiated pattern of an incident radio frequency (RF) signal indicating spatial distribution characteristics of a device under test (DUT). The system includes multiple antenna elements arranged in an array, where the antenna elements include antennas and diodes coupled to the antennas, respectively. Each antenna has a radar cross-section (RCS) that is too small for the antenna elements to operate in a reflect mode. Each diode is zero biased, such that the diodes receive the incident RF signal through the antennas to which the diodes are respectively coupled, and rectify the incident RF signal to DC voltages that are proportional to power of the incident RF signal, enabling the antenna elements to operate in a detect mode.

Abstract: A light-emitting substrate includes: a substrate; and at least one coupling portion and at least one auxiliary cathode pattern disposed on the substrate. Each auxiliary cathode region is provided with a coupling portion and an auxiliary cathode pattern coupled to the coupling portion. The auxiliary cathode pattern includes at least one disconnection portion, and each disconnection portion includes first, second and third conductive pattern layers. An orthographic projection of an edge of the second conductive pattern layer on the substrate is located within orthographic projections of edges of the third and first conductive pattern layers on the substrate. A total perimeter of at least one orthographic projection, on the substrate, of at least one edge of at least one third conductive pattern layer in the at least one disconnection portion is greater than a perimeter of an orthographic projection of an edge of the auxiliary cathode region on the substrate.

Abstract: A method for reinforcing soft ground by post-grouting combined with pressurized vacuum preloading is proposed, by pre-burying prefabricated vertical drains and air-boosted pipes in granular material piles, and the air-boosted pipes are used as grouting pipes to reduce the number of times of piling, which not only improves the construction efficiency, but also reduces the structural disturbance of the soil and the influence of smear effect, thus reducing the impact on the radial permeability and the radial consolidation coefficients. The method does not use geotextile bags for granular materials, which can avoid the problem of forming a localized clogging area around the geotextile bags, and the method not only improves the efficiency of vacuum transfer in a pre-consolidation stage, but also improves the grouting effect in the later stage, effectively enhances the strength of soft soil and makes granular material piles and the surrounding soil form composite ground.

Abstract: The embodiments of the present disclosure provide a method for predicting a gate station filter replacement at a gate station for smart gas and an Internet of Things system. The method includes: obtaining, by a smart gas data center, usage information of a filter element through a smart gas sensor network platform, the usage information at least including at least one of ventilation efficiency of the filter element, filtered impurity information, and a blockage degree; obtaining, by a smart gas pipeline network device management sub-platform, the usage information from the smart gas data center, determining a filter element maintenance plan based on the usage information, and sending the filter element maintenance plan to the smart gas data center; and sending, by the smart gas data center, the filter element maintenance plan to a smart gas user platform through a smart gas service platform.

Abstract: A jet flow adjustment method for rotary steerable downhole release includes steps of: defining a diversion ratio p of a jet nozzle by a ground control system; and sending the diversion ratio p of the jet nozzle to a control unit of a downhole rotary steerable system; and determining an opening areas of the jet nozzle by the control unit according to the diversion ratio p of the jet nozzle and an equivalent channel area S of a drilling tool; controlling an adjustable bypass valve by the control unit of the downhole rotary steerable system according to the calculated opening area s of the jet nozzle, so as to control an overlapping area between sector holes of a rotor end and a stator end, thereby opening the jet nozzle for release. The present invention has simple control and strong release ability.

Abstract: Method and an IoT system for gas gate station monitoring are provided. The method includes obtaining operation data of a gas gate station; obtaining a downstream user feature corresponding to the gas gate station and historical warning data corresponding to the gas gate station; determining, based on the downstream user feature and the historical warning data and in combination with the operation data, the associated node in a preset time period, the associated node including at least one of a gate station internal node and an external pipeline network node; obtaining monitoring data of the associated nodes, the monitoring data at least including node monitoring data of the gate station internal node and node monitoring data of the external pipeline network node; and issuing a warning notification in response to the monitoring data not satisfying a preset condition.

Abstract: A model training method and a related apparatus to help improve a convergence speed of model training and improve end-to-end communication quality. The method includes: a first communication apparatus sends first data to a second communication apparatus through a channel, where the first data is an output result of the first machine learning model. The second communication apparatus receives second data through a channel, inputs the second data into a second machine learning model to obtain third data; determines a first loss function based on the third data and the first training data; and sends the first loss function to the first communication apparatus through a feedback channel.

Abstract: A drug coated balloon includes a balloon body and a drug loaded coating layer. The drug loaded coating layer includes a first drug loaded coating component having at least two controllable sustained release drugs with different drug release kinetics, and a second drug loaded coating component including an active drug. The second drug loaded coating component is dispersed among the controllable sustained release drugs and couples the controllable sustained release drugs on the balloon body. The first and second drug loaded coating components are both non-hydrophilic compositions. The drug coated balloon provides sufficient initial drug loading dosage to the lesion site, and has adjustable drug release rate, thereby allowing the drug coated balloon to provide bioactive drugs to the treatment site at different stages of the cascade reaction of vascular restenosis, providing the controllable long-acting therapeutic drugs, and improving the overall vascular drug release efficacy.

Abstract: Disclosed are a resolution method for intent-based wireless network resource conflicts and an apparatus thereof. The method includes the following steps: periodically acquiring configured resource information about a current wireless network; pre-allocating resources required for a newly-added intent; and determining whether a pre-allocated resource configuration policy is incompatible with a resource configuration policy being executed in a current network. The apparatus includes: an interface module; a solution module; a feedback module; and a storage module. Through the relative priority of intent requirements, a bias of an implementation satisfaction degree of a network configuration policy is determined, and network resource conflicts of intent requirements with different priorities are solved.

Abstract: This specification provides a training method of a hybrid graph neural network model. The hybrid graph neural network model includes an encoding function and a decoding function. The method includes the following: using instances corresponding to all targets in training samples and several nearest neighbors of the instances as nodes in a graph, a graph representation vector of each instance is generated by using the encoding function based on graph data of all the instances. t rounds of training are performed on a decoding parameter; and in each round, bs targets are extracted from training samples, a predicted quantity of each target is generated by using the decoding function based on the graph representation vector of the instance corresponding to each target and non-graph data corresponding to each target, and the decoding parameter is optimized based on a loss quantity of the current round that is determined by the predicted quantities and label quantities of the bs targets in the current round.

Abstract: A device and method for glaucoma auxiliary diagnosis, and a non-transitory storage medium are provided. The device includes an obtaining unit and a processing unit. The obtaining unit is configured to obtain a color fundus image of a patient. The processing unit is configured to perform feature extraction on the color fundus image to obtain a first feature map. The processing unit is further configured to perform image segmentation on the color fundus image according to the first feature map to obtain an optic disc image in the color fundus image, where the optic disc image corresponds to an optic disc area in the color fundus image. The processing unit is further configured to perform feature extraction on the optic disc image and the color fundus image according to the first feature map to obtain a probability that the patient has glaucoma.

Abstract: A server device which is self-adaptive to different placements and locations includes a chassis, a plurality of storage elements, and a storage cabinet. First and third slide rails are positioned at opposite sides of the storage cabinet along a first direction. Second and fourth slide rails are positioned at other two opposing sides of the storage cabinet. The storage cabinet is movably positioned in the chassis along a third direction. When the chassis is horizontal, the first slide rail and the third slide rail provide support for the storage cabinet and when the chassis is vertical, support is provided by the second and fourth slide rails. The problem of the storage cabinet being difficult to slide when the placement of the chassis is changed is reduced, and the server device is more flexible in its placement.

Abstract: A method including obtaining image data and attribute information of a first item in an item catalog. The method also can include generating candidate variant items from the item catalog for the first item using a combination of (a) a k-nearest neighbors approach to search for first candidate variant items based on text embeddings for the attribute information of the first item, and (b) an elastic search approach to search for second candidate variant items based on image embeddings for the image data of the first item. The method additionally can include performing respective classifications based on respective pairs comprising the first item and each of the candidate variant items to filter the candidate variant items. The method further can include determining a respective distance between the first item and each of the candidate variant items, as filtered.

Abstract: This application discloses a model training method, which may be applied to the field of artificial intelligence. The method includes: when training a first neural network model based on a training sample, determining N parameters from M parameters of the first neural network model based on a capability of affecting data processing precision by each parameter; and updating the N parameters. In this application, on a premise that it is ensured that the data processing precision of the model meets a precision requirement, because only N parameters in M parameters in an updated first neural network model are updated, an amount of data transmitted from a training device to a terminal device can be reduced.