8results about How to "Reduce deployment complexity" patented technology

This application relates to a knowledge blind spot perception distillation method, system, and readable storage medium for target language models. The method includes: segmenting domain documents into semantic fragments; generating a candidate question-answer pair set containing knowledge category labels using a first language model; obtaining the target language model's answers to each question without context; calculating a knowledge mastery score through fact-level comparison; filtering out an unknown question-answer pair set that the model does not yet possess; performing hash bucketing clustering on the unknown question-answer pair set according to category labels to obtain multiple topic clusters; and allocating character sub-budgets to each topic cluster based on the proportion of question-answer pairs; generating structured text based on the sub-budgets and unknown question-answer pairs; and finally assembling the text into distilled prompt words for use by the target model. This application effectively filters out knowledge already possessed by the model, increases the information density of the prompt words, and reduces the computational and storage overhead of edge deployment.

The invention relates to the technical field of signal processing, and discloses a simultaneous same-frequency full-duplex self-interference elimination method and system, and the method comprises the steps: carrying out the data preprocessing of a generated self-interference signal, generating a self-interference signal containing a nonlinear factor, and constructing a training data set according to the self-interference signal; establishing a plurality of CNN-LSTM attention models for extracting nonlinear features and time sequence features of the self-interference signals; an auxiliary sub-module is introduced in the training stage of the model, so that the model has internal knowledge migration, and knowledge distillation of a feature level is realized; training the model, and performing parameter optimization by adopting a loss function which considers the amplitude and the phase of the complex signal at the same time; based on the trained model, linear and nonlinear two-stage self-interference elimination is carried out on the received signal, self-interference suppression in a full duplex system is realized, and the problems that the existing full duplex nonlinear self-interference elimination method based on deep learning is insufficient in feature extraction capability, limited in supervision information and low in training speed are effectively solved through the method.

The application discloses a kind of end side closed-loop full-automatic electrochemiluminescence analysis system, it is related to electrochemiluminescence detection technology, including man-machine interaction and analysis subsystem and electrochemiluminescence signal acquisition subsystem;Man-machine interaction and analysis subsystem and electrochemiluminescence signal acquisition subsystem between establishment has communication connection to execute data interaction;Wherein, man-machine interaction and analysis subsystem in preset automatic detection closed-loop process, for via communication connection to electrochemiluminescence signal acquisition subsystem sends electrochemical excitation control instruction, and the image data generated in the process of electric excitation is collected, then image data is automatically detected and handled, to obtain target luminescence signal intensity.The application further discloses a kind of detection method.The application is through full-automatic detection process control mechanism, realizes the automatic collaborative operation of luminescence process acquisition, electric excitation application, image processing and signal analysis, reduces manual participation, greatly improves processing efficiency.

The application discloses a ship traffic flow prediction method and system based on a neural network and an attention mechanism, and the method comprises the following steps: performing data elimination and data preprocessing on an obtained historical ship traffic data set to obtain a pretreated traffic data set; performing data discretization on the pretreated traffic data set to obtain a discrete ship traffic data set of a plurality of voyage units; calculating the heading entropy of the discrete ship traffic data set; constructing ship traffic multidimensional features, performing feature extraction on the ship traffic multidimensional features, obtaining short-term correlation features and long-term correlation features; and obtaining fusion features through splicing and fusion; obtaining spatial attention weights and time attention weights, and obtaining a broadcast tensor through tensor broadcasting; performing multi-scale prediction on the broadcast tensor to obtain ship traffic flow prediction data. The application realizes data screening by combining ship features, realizes space-time feature fusion to improve model interpretability, and reduces model deployment complexity through multi-step prediction.

This invention relates to a latency guarantee method for cloud-based PLC services under a 5G-TSN architecture. It adds service priority mapping, priority queue management, and wireless resource scheduling functions to the 5G-TSN network. The wireless resource priority allocation method for cloud-based PLC industrial control services includes statistical queue data volume, feedback of wireless channel quality, calculation of required wireless resources, and allocation of wireless resources. This reduces 5G network latency, minimizes intermediate cables and equipment, expands the mobile range of equipment terminals, and ensures priority resource allocation for cloud-based PLC services over the wireless interface. In multi-service mixed transmission scenarios, the proposed algorithm achieves lower latency than algorithms without guarantees. The virtual communication interface can flexibly connect with multiple protocols, supports interconnection of devices from multiple manufacturers, enables cloud-based collaborative control, improves production efficiency, and allows the 5G-TSN network to simultaneously carry multiple services while providing latency deterministic guarantees for industrial control services, reducing the deployment complexity of industrial field networks.

This invention discloses a fake image detection method based on artifact decomposition. The method includes acquiring the image to be detected and inputting it into a trained three-branch artifact-aware encoder (scene consistency branch, imaging realism branch, and signal naturalness branch) for feature extraction, resulting in three artifact reflection maps. A trained cross-dimensional gated collaborative fusion module is used to fuse the features of the three artifact reflection maps, generating a unified artifact embedding representation. This unified artifact embedding representation is then input into a trained classification head, and the label classification head of the classification head outputs the true / false prediction probability of the image to be detected. This fake image detection method solves the problems of existing technologies, such as inability to uniformly detect multi-source heterogeneous fake images, poor generalization ability, and susceptibility to overfitting.

The invention discloses an application program performance detection method and system and a nonvolatile storage medium. The method comprises the following steps: receiving original stack data of a target application transmitted by a proxy end; based on the original stack data, symbolized data corresponding to the target application program is determined, and the symbolized data comprises the name of a calling function, the file path of the calling function and the source code line number of the calling function; and based on the symbolized data, determining a performance detection result of the target application program, the performance detection result including a function execution path and a resource consumption condition. According to the method and the device, the technical problems of high complexity and high overhead in deployment and operation of an application program performance detection technology in a production environment at present are solved.

The embodiment of the invention relates to the technical field of automatic testing, and discloses an automatic testing system and method and a vehicle, and the system comprises an upper computer main service module which is used for obtaining a testing script, carrying out the abstract conversion of the testing script, and obtaining a target abstract instruction; the target abstract instruction is a standardized target abstract instruction for describing a test operation type; the access and access control module is used for determining a corresponding virtual device node when the target abstract instruction is received; converting the target abstract instruction into an executable instruction corresponding to the operation equipment through the virtual equipment node; controlling the operation equipment to execute the executable instruction, and receiving a returned execution result; the virtual device node is created based on a device capability description model of the operating device. According to the technical scheme, the test script is converted into the standardized abstract instruction, so that the test script does not need to be modified for different devices or scenes, and cross-device and cross-scene efficient multiplexing is realized.