13results about How to "Reduce update" patented technology

The invention discloses a federal learning model defense mechanism performance evaluation method under optimized backdoor attack, and belongs to the field of network security. The method comprises the following steps: an attacker normally participates in training in the early stage of federal learning training; in the middle stage of training, an attacker obtains the global model; when an attacker normally participates in pre-training, the obtained global model is utilized in advance, normal training and backdoor model training are simulated, and a trigger is optimized in advance by updating differences; in the later stage of training, performing backdoor attack by using the optimized trigger, and continuously optimizing the trigger by using the global model obtained again by using the same method, so as to improve the attack effect and efficiency; and performing backdoor attack test based on the finally optimized trigger, and evaluating the performance of the federal learning model defense mechanism. According to the federal learning model vulnerability evaluation method under the optimized backdoor attack, a set of new attack thought and an effective evaluation method are provided for existing defense, and vulnerabilities of the existing defense are revealed.

The present application belongs to the technical field of Internet of Vehicles and communication security, and discloses a federated learning training method based on vehicle state perception and applicable to dynamic Internet of Vehicles. The method collects vehicle running state information of vehicle clients, preliminarily screens candidate vehicle clients, calculates comprehensive scores of the screened vehicle clients, adopts a client selection strategy combining selection of the top-ranked clients with random supplementation to determine a set of vehicle clients participating in the current round of federated learning, configures local training parameters for the selected vehicle clients according to vehicle speed and computing capacity, performs local training on local data sets to obtain model updates, performs weighted aggregation on the model updates of the effective participating vehicle clients at the server end, introduces a momentum mechanism to smooth the model update process, and obtains new global model parameters. This method effectively improves the stability, robustness and training efficiency of federated learning in a high-dynamic Internet of Vehicles environment.

The present application discloses a method, device and equipment for rendering illumination information in a game scene, which relates to the field of 3D rendering technology and can reduce the update, transmission and storage costs of illumination information sampling in the game scene, and improve the rendering efficiency of illumination information. The method includes: arranging a probe grid for spatial voxels in the game scene, wherein the probe grid includes illumination probes for capturing illumination information in the game scene, updating illumination probes at corresponding viewpoint positions in the probe grid according to viewpoint positions in the game scene, and capturing illumination information at viewpoint positions; transmitting illumination information at viewpoint positions to a graphics processor after generating multiple probe data textures with the same data size through a central processing unit; in response to a rendering instruction called by the central processing unit, the graphics processor samples illumination information of objects in the game scene using resource textures and indirect textures according to the position information of sampling points on the surface of objects in the game scene, and renders the illumination information of the objects.

A fitting agent for a hearing device system is disclosed, the hearing device system comprising a hearing device worn by a hearing device user, where the fitting agent comprises one or more processors configured to: initialize a user model and an environment model, the user model comprising a plurality of user preference functions and associated user response distributions; acquiring environmental data indicating a current environment; determining a first initial environment probability of the first environment and a second initial environment probability of the second environment based on the environment data and the environment model, obtaining test settings including primary test settings and secondary test settings of the hearing device based on the first initial environment probability and the second initial environment probability; presenting the test settings to the hearing device user; obtaining a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting; and updating the user model based on the preferred test settings and the environmental data to provide an updated user model.

The invention provides an automatic navigation aided driving system, which relates to the technical field of aided driving and comprises a user interaction module, a navigation and positioning module, a road information processing module, a whole vehicle attitude calculation module, a map information processing and storage module, an aided driving road condition acquisition module and an aided driving control module. According to the invention, the camera, radar and other ADAS sensors on the vehicle are utilized to acquire images of common routes, the images are stored in the vehicle and are only used for an automatic navigation scheme of the vehicle, the scheme is high in flexibility, and the situation that the functions are difficult to use by customers in small places due to resource limitation of image acquirers is avoided; the map packet is small and fast to update, and almost has no cost; the technical problems that in the prior art, a high-precision map collection mode is likely to cause the small coverage of a high-precision map or most maps are not updated urgently and are not accurate, and use and accuracy of functions are limited are solved.

The utility model discloses a tree pit structure provided with a construction device for preventing soil balls of border trees from protruding, the border trees with the soil balls are planted in the tree pit, a planting soil layer is filled in the tree pit, the tree pit is located between a sidewalk paving area and a roadway, and curbstones are arranged on the side, close to the roadway, of the tree pit. A boomerang type construction device is arranged on the upper wall of the inner side of the plant pit, the boomerang type construction device comprises two unequal triangular bodies, each triangular body comprises a short bevel edge and a long bevel edge, the inner sides of the short bevel edges and the long bevel edges are connected, the short bevel edges are located above the soil ball, the long bevel edges are located on the side edge of the soil ball, and the size of the boomerang type construction device is determined according to the size of the plant pit. By means of the boomerang type construction device, damage to sidewalk paving materials and potential safety hazards brought to surrounding pedestrians due to the fact that soil balls protrude out of tree pits along with continuous growth of border trees are effectively avoided, and meanwhile influences on surrounding roadways in the growth process of the border trees are prevented. And the government maintenance cost is reduced.

The embodiment of the application discloses a dynamic phasor representation DC power system random transient analysis method, relates to the power electronic field of the DC power system, and can reflect system characteristics caused by source-load random excitation conditions and multi-rate simulation, and balance the contradiction between precision and efficiency. The application comprises the following steps: obtaining DC power system parameters and setting initial parameters; obtaining circuit models of two types of lumped elements in the DC power system; establishing a dynamic phasor model of a converter station in the DC power system and obtaining a dynamic companion circuit of the converter station; using the circuit models of the two types of lumped elements and the dynamic companion circuit of the converter station, establishing a model of the DC power system; obtaining node voltage equations of the DC power system; updating a node admittance matrix and an injected current vector according to changes in a switching function of the converter station; and converting dynamic phasors output by the model of the DC power system into instantaneous values.

The invention relates to a data updating method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps that in response to a received data updating instruction, a change data set and a state data set matched with the data updating instruction are determined, the change data set is used for storing data state change information, and the state data set is used for storing data state information; under the condition that the data size of the change data set is larger than a preset threshold value, data screening is conducted on the change data set according to field information corresponding to data in the state data set, and a screened change data set is obtained; and determining to-be-updated data according to the state data set and the screened change data set, and updating the to-be-updated data. By adopting the method, the data updating efficiency and accuracy can be improved.

The utility model relates to the field of tobacco tar injection equipment, in particular to leakage-proof oil injection equipment which comprises a needle head body and a connecting piece, the needle head body is communicated with a plurality of oil injection needles, one end of the connecting piece is connected with the needle head body, and the other end of the connecting piece is connected with a sealing plug. The plane where the bottom of the sealing plug is located is lower than the plane where the tail ends of needle heads of all the oil injection needles are located. The outer diameter of the sealing plug is larger than or equal to the diameter of the heating wire assembly hole. The sealing plug is arranged, the position of the sealing plug is lower than the positions of the tail ends of the needle heads of all the oil injection needles, and meanwhile it is ensured that the outer diameter of the sealing plug is larger than or equal to the hole diameter of the heating wire assembly hole, so that the heating wire assembly hole is effectively plugged, and oil liquid is prevented from leaking from the hole. By means of the design, the problem that oil liquid is prone to leakage when the heating core is not installed in a traditional oil injection mode is fundamentally solved.

The utility model relates to a gun head of a shoot nail fastening tool. The gun head comprises a gun nozzle support, a lower cover plate and a gun nozzle. A first nail channel is formed between the gun nozzle support and the lower cover plate, a material receiving part is arranged on the lower side of one end of the gun nozzle and extends into the channel, the arc-shaped concave face of the material receiving part and the gun nozzle support form a second nail channel, the diameter of the second nail channel is larger than the maximum diameter of nails and smaller than two times of the maximum diameter, multi-nail clamping can be effectively avoided, and a single nail can be accurately positioned. The gun nozzle is provided with a guide part matched with the guide notch of the gun nozzle support, and a receding groove is formed between the material receiving part and the guide part. The lower cover plate at the front end of the nail channel is provided with a receiving groove and a stop step. A third nail channel of the gun nozzle is provided with internal threads, the bottom of the lower cover plate is provided with an elastic clamping opening, a matched nail channel and a nail groove, and the nail groove comprises a nail point long groove, a universal clamping groove, a nail tail expansion groove and other special designs. Nail shooting stability, reliability and universality can be improved, nail clamping is reduced, and the service life of equipment is prolonged.

A fitting agent for a hearing device system is disclosed, the hearing device system comprising a hearing device worn by a hearing device user, where the fitting agent comprises one or more processors configured to: initialize a user model; initializing an environment model; obtaining environmental data indicating a current environment; determining an environmental state based on the environmental data using the environmental model; obtaining test settings for the hearing device based on the environmental state, the test settings including a test setting A and a test setting B; presenting the test settings to the hearing device user; obtaining a user input of a preferred test setting indicative of a preference for either the A test setting or the B test setting; updating the user model based on the preferred test settings and the environmental state to provide an updated user model; and updating the environmental model based on the preferred test settings and one of the user model and the updated user model to provide an updated environmental model.

This invention provides a method for dynamic urban renewal based on deep reinforcement learning, relating to the field of urban renewal. The method includes: step S1, data acquisition; step S2, state prediction; step S3, hierarchical decision-making; step S4, execution feedback; and step S5, iterative optimization. Hierarchical decision-making, based on the urban state data in step S1 and the predicted state in step S2, implements hierarchical update schemes for macro-level, meso-level, and micro-level intelligent agents. This method transforms urban renewal from a one-time static planning process to dynamic adaptive decision-making that evolves with the urban system, thereby achieving dynamic adaptation of infrastructure and public services and avoiding negative impacts such as environmental damage, facility idleness, resource waste, and increased costs caused by mismatch.

The invention provides a three-dimensional point cloud semantic segmentation method and device and a storage medium, and belongs to the technical field of computer vision and image processing, and the method comprises the steps: obtaining three-dimensional point cloud data in a target scene, constructing a semi-supervised average teacher network, and training the semi-supervised average teacher network through the point cloud data; performing dual-view uncertainty evaluation on each sample in the unlabeled point cloud data in the target scene by using the trained average teacher network to obtain a fusion uncertainty score of each sample, and screening out a to-be-labeled sample from the unlabeled samples by using a dual-drive selection strategy based on the score and a sample confusion degree score; and labeling the to-be-labeled sample, adding the labeled data set, and carrying out iterative training on the average teacher network until a preset condition is met to obtain a final point cloud semantic segmentation model suitable for the target scene. Based on the method, the invention further provides corresponding equipment and a storage medium. According to the invention, efficient point cloud sample selection is realized.