Security check method based on distributed experimental equipment and related equipment

Abstract

The invention belongs to the field of data processing, and particularly relates to a security check method based on distributed experimental equipment and related equipment.The method comprises the steps that in response to a receiving instruction for a target object in a distributed experimental equipment management and control system, to-be-checked information of the target object is obtained, setting the target object to be in a to-be-checked state; inputting the to-be-examined information of the target object into a security detection model, and performing security examination on the target object to obtain an examination result of the target object; setting the target object to be in a corresponding candidate inspection state based on an inspection result of the target object; and pushing to-be-examined information of the target object to an examination and approval channel matched with the candidate examination state, and executing safety examination and approval of the target object. According to the method, the operation efficiency and the management level of the equipment can be improved, and the safety and the reliability of the distributed experimental equipment are improved.

Description

Technical Field

[0001] The invention belongs to the field of data processing, and in particular relates to a safety inspection method based on distributed experimental equipment and related equipment. Background Art

[0002] In order to promote the popularization of intelligent applications in various industries and fields, distributed experimental equipment generates a large amount of data.

[0003] In related technologies, the data generated by distributed experimental equipment often contains various formats and types, and its structure may change dynamically. Traditional systems may be inefficient in processing such diverse and unstructured data, with limited processing capabilities and slow responses, and cannot meet the management of these data, causing difficulties in the management of distributed experimental equipment. For example, distributed experimental equipment may generate data in various formats and types, such as text, images, videos, sensor data, etc. These data may have different structures and semantics, and traditional systems are difficult to process and manage uniformly. In order to ensure the safe storage and security management of these data, it is necessary to conduct security checks on these data, but the current inspection methods all need to be manually set and evaluated, which is extremely inefficient and costly.

[0004] Therefore, it is urgent to design a safety inspection solution based on distributed experimental equipment to solve at least one of the above technical problems. Summary of the invention

[0005] The purpose of the present invention is to provide a security inspection method and related equipment based on distributed experimental equipment, which is used to realize an automated initial security inspection of equipment data through a security detection model, and then push the equipment data to different approval channels for further security inspection in combination with routing sorting, thereby greatly reducing the workload of security inspection, helping to improve the operating efficiency and management level of the equipment, and enhancing the safety and reliability of distributed experimental equipment.

[0006] The technical solution of the present invention:

[0007] In a first aspect, the present application provides a security inspection method based on distributed experimental equipment, which is applied to a distributed experimental equipment management and control system, wherein the distributed experimental equipment management and control system is used to operate and manage experimental equipment, and the experimental equipment includes different types of equipment; the security inspection method based on distributed experimental equipment includes:

[0008] In response to receiving instructions for a target object in the distributed experimental equipment management and control system, obtaining information to be reviewed of the target object and setting the target object to a state to be reviewed; the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters;

[0009] Inputting the to-be-reviewed information of the target object into the security detection model, and performing a security review on the target object to obtain a review result of the target object;

[0010] Based on the review result of the target object, the target object is set to a corresponding candidate inspection state; the candidate inspection state includes a security tag corresponding to the target object; the security tag is associated with the security state of the target object and the object type to which it belongs;

[0011] The information to be reviewed of the target object is pushed to the approval channel that matches the candidate inspection status, and the security approval of the target object is performed.

[0012] In a second aspect, the present application provides a distributed experimental equipment management and control system, the distributed experimental equipment management and control system is used to operate and manage experimental equipment, the experimental equipment includes different types of equipment, and the distributed experimental equipment management and control system includes:

[0013] The acquisition unit is configured to, in response to a receiving instruction for a target object in the distributed experimental equipment management and control system, acquire the information to be examined of the target object and set the target object to a state to be examined; the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters;

[0014] A checking unit, configured to input the to-be-examined information of the target object into a security detection model, and perform a security review on the target object to obtain a review result of the target object;

[0015] A setting unit is configured to set the target object to a corresponding candidate inspection state based on the review result of the target object; the candidate inspection state includes a security tag corresponding to the target object; the security tag is associated with the security state of the target object and the object type to which it belongs;

[0016] The pushing unit is configured to push the to-be-examined information of the target object to the approval channel that matches the candidate inspection status, and perform security approval on the target object.

[0017] In a third aspect, the present application provides an intelligent computing platform, the intelligent computing platform comprising:

[0018] at least one processor, memory, and input-output unit;

[0019] Among them, the memory is used to store computer programs, the processor is used to call the computer program stored in the memory to execute the first aspect of the security inspection method based on distributed experimental equipment, and the input and output unit is used to receive user input and display output information of the computer program stored in the memory.

[0020] In a fourth aspect, a computer-readable storage medium is provided, which includes instructions. When the instructions are executed on a computer, the computer executes the security inspection method based on distributed experimental equipment of the first aspect.

[0021] Beneficial effects of the present invention:

[0022] The technical solution provided by the present invention can be applied to a distributed experimental equipment management and control system. The distributed experimental equipment management and control system is used to operate and manage experimental equipment. These experimental equipment include different types of equipment. In this solution, first, in response to the receiving instruction of the target object in the distributed experimental equipment management and control system, the information to be reviewed of the target object is obtained, and the target object is set to a state to be inspected. Among them, the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters. Then, the information to be reviewed of the target object is input into the security detection model, and a security review is performed on the target object to obtain the review result of the target object. Then, based on the review result of the target object, the target object is set to a corresponding candidate inspection state. Finally, the information to be reviewed of the target object is pushed to the approval channel that matches the candidate inspection state, and the security approval of the target object is performed.

[0023] The technical solution of the present invention can respond to instructions in real time, quickly obtain information to be reviewed, and conduct security review. This can ensure that the control of experimental equipment is timely and effective, and reduce the security risks caused by delays. By reviewing the target object through the security detection model, it is also possible to comprehensively review the experimental equipment files, software and parameters, thereby improving the comprehensiveness and accuracy of the review. In addition, associating the review results with security tags can better manage the security status of the experimental equipment. The setting of security tags can help administrators or users quickly understand the security of the equipment and help to take necessary security measures in a timely manner. It can also automatically push the information to be reviewed to the matching approval channel and perform security approval. This can reduce the cost of manual intervention, improve approval efficiency, and reduce security vulnerabilities caused by human factors. Through the distributed experimental equipment management and control system, the review tasks can be assigned to different nodes for processing, thereby reducing the load of a single node and improving the concurrent processing capacity and overall performance of the system. The technical solution of this application realizes the automated security initial inspection of equipment data through the security detection model, and then pushes the equipment data to different approval channels for further security inspection in combination with routing sorting, which greatly reduces the workload of security detection, helps to improve the operating efficiency and management level of the equipment, and improves the security and reliability of distributed experimental equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0025] Figure 1 It is a flowchart of a safety inspection method based on distributed experimental equipment according to an embodiment of the present application;

[0026] Figure 2 It is a structural schematic diagram of a distributed experimental equipment control system according to an embodiment of the present application;

[0027] Figure 3 It is a structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0030] In order to promote the popularization of intelligent applications in various industries and fields, distributed experimental equipment generates a large amount of data.

[0031] In related technologies, the data generated by distributed experimental equipment often contains various formats and types, and its structure may change dynamically. Traditional systems may be inefficient in processing such diverse and unstructured data, with limited processing capabilities and slow responses, and cannot meet the management of these data, causing difficulties in the management of distributed experimental equipment. For example, distributed experimental equipment may generate data in various formats and types, such as text, images, videos, sensor data, etc. These data may have different structures and semantics, and traditional systems are difficult to process and manage uniformly. In order to ensure the safe storage and security management of these data, it is necessary to conduct security checks on these data, but the current inspection methods all need to be manually set and evaluated, which is extremely inefficient and costly.

[0032] Therefore, it is urgent to design a safety inspection solution based on distributed experimental equipment to solve at least one of the above technical problems.

[0033] The embodiments of the present application provide a security inspection method based on distributed experimental equipment and related equipment.

[0034] Specifically, in the security inspection scheme based on distributed experimental equipment, first, in response to the receiving instruction of the target object in the distributed experimental equipment management and control system, the information to be reviewed of the target object is obtained, and the target object is set to a state to be reviewed. Among them, the target object includes at least one of the following: experimental equipment files, experimental equipment software, experimental parameters. Then, the information to be reviewed of the target object is input into the security detection model, and a security review is performed on the target object to obtain the review result of the target object. Then, based on the review result of the target object, the target object is set to a corresponding candidate inspection state. Finally, the information to be reviewed of the target object is pushed to the approval channel that matches the candidate inspection state, and the security approval of the target object is performed.

[0035] In the security inspection scheme based on distributed experimental equipment, it is possible to respond to instructions in real time, quickly obtain information to be reviewed, and conduct security reviews. This ensures that the control of experimental equipment is timely and effective, and reduces security risks caused by delays. By reviewing the target object through the security detection model, it is also possible to comprehensively review the experimental equipment files, software, and parameters, thereby improving the comprehensiveness and accuracy of the review. In addition, associating the review results with security tags can better manage the security status of the experimental equipment. The setting of security tags can help administrators or users quickly understand the security of the equipment and help to take necessary security measures in a timely manner. It can also automatically push the information to be reviewed to the matching approval channel and perform security approval. This can reduce the cost of manual intervention, improve approval efficiency, and reduce security vulnerabilities caused by human factors. Through the distributed experimental equipment control system, review tasks can be assigned to different nodes for processing, thereby reducing the load of a single node and improving the concurrent processing capability and overall performance of the system.

[0036] In a word, in the safety inspection scheme based on distributed experimental equipment, the automated safety initial inspection of equipment data is realized through the safety detection model, and then the equipment data is pushed to different approval channels for further safety inspection in combination with routing sorting, which greatly reduces the workload of safety detection, helps to improve the operating efficiency and management level of the equipment, and improves the safety and reliability of distributed experimental equipment. The safety inspection scheme based on distributed experimental equipment provided by the embodiment of the present application can also be performed by an electronic device, which can be a server, a server cluster, or a cloud server. The electronic device can also be a terminal device such as a mobile phone, a computer, a tablet computer, a wearable device, or a special device (such as a special terminal device with a safety inspection system based on distributed experimental equipment, etc.). The chips introduced in the above embodiments can also be carried in these electronic devices. Alternatively, these electronic devices can also be installed with a service program for executing a safety inspection scheme based on distributed experimental equipment.

[0037] In the embodiment of the present application, the distributed experimental equipment management and control system is mainly responsible for the storage of various relevant data such as input data, calculation results, observation data, visualization data, etc. of the advanced computing platform. These data may come from different applications and need to be uniformly managed and stored for subsequent analysis and processing.

[0038] Figure 1 A schematic diagram of a security inspection method based on distributed experimental equipment provided in an embodiment of the present application is shown in FIG. Figure 1 , the method comprises the following steps:

[0039] 101. In response to receiving an instruction for a target object in the distributed experimental equipment management and control system, obtain the to-be-examined information of the target object and set the target object to a to-be-examined state.

[0040] In the embodiment of the present application, the distributed experimental equipment management and control system is used to operate and manage experimental equipment. These experimental equipment include different types of equipment.

[0041] In the embodiment of the present application, the security inspection method based on the distributed test equipment security management system involves a relatively complex system architecture, which needs to fully consider the communication, data transmission, security and other aspects between the server and the client. The following is an introduction to the technical background of the system and related equipment:

[0042] Further optionally, the distributed test equipment safety management and control system is designed based on a C / S architecture. C / S architecture refers to a client / server architecture, which separates the user interface and application logic. The user communicates with the server through the client, and the server processes the request and returns the result. This architecture is usually used for applications that require complex logic processing and large amounts of data storage.

[0043] Microservice architecture is an architectural pattern that splits an application into a series of small, independently deployed services. Each microservice can be independently developed, deployed, and expanded, and communicate through lightweight communication protocols, thus achieving a highly cohesive and loosely coupled system architecture.

[0044] QT is a cross-platform C++ application development framework that can be used to develop graphical user interface applications. It provides a wealth of GUI components and tools that can simplify the development process of cross-platform applications while having good performance and scalability.

[0045] Distributed test equipment refers to test equipment deployed at different locations or nodes. These devices can collaborate with each other to complete test tasks and communicate and exchange data through the network. Distributed test equipment is usually highly flexible and scalable, and can be dynamically configured and deployed as needed.

[0046] The server is the core of the entire system, responsible for the implementation of various functional modules and data processing. In this system, the server adopts a microservice architecture design, which includes multiple independent functional modules, such as operation status monitoring, audit log collection and analysis, ledger management, resource management, etc. Each functional module can be deployed and expanded independently, and interact through a lightweight communication mechanism.

[0047] The client is the interface for users to interact with the system. It is responsible for receiving user input, displaying information and sending requests to the server. In this system, the client is developed based on QT and includes functional modules such as user authentication, device management, information management, test parameter management, file management, software management, and remote control. The client realizes the interaction between users and the system and security checks through communication with the server.

[0048] In summary, the safety inspection method based on the distributed test equipment safety management and control system combines the C / S architecture, microservice architecture and QT development technology, realizes the real-time management and intelligent control of distributed test equipment, and provides strong support for the safety, stability and reliability of the test equipment.

[0049] Further optionally, the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters. When it comes to target objects in the experimental equipment management and control system, these objects are usually important components of the system that need to be managed and reviewed.

[0050] Specifically, experimental equipment files refer to various files related to experimental equipment, which may include experimental data files, configuration files, log files, etc. Experimental equipment files contain data and configuration information generated during the experiment, which is crucial to the accuracy and integrity of the experimental results. Reviewing experimental equipment files can ensure the integrity, confidentiality and credibility of data, prevent malicious tampering or leakage of data, and thus ensure the credibility and security of scientific research results.

[0051] Experimental equipment software is the software program run by the experimental equipment, including control software, analysis software, simulation software, etc. Experimental equipment software directly affects the function and performance of experimental equipment. Reviewing experimental equipment software can ensure the legality, security and stability of the software, timely discover and repair potential security vulnerabilities, prevent malicious attacks or illegal use, and ensure the normal operation of experimental equipment and the reliability of experimental data.

[0052] Experimental parameters refer to the parameters that need to be set and adjusted during the experiment, such as temperature, pressure, speed, etc. Experimental parameters directly affect the accuracy and repeatability of experimental results. Reviewing experimental parameters can ensure the rationality and safety of parameter settings, prevent experimental failures or accidents caused by incorrect parameter settings, improve the efficiency and success rate of the experiment, and ensure the accuracy of experimental data and the credibility of scientific research results.

[0053] In summary, experimental equipment files, experimental equipment software and experimental parameters are the target objects that need to be managed and reviewed in the experimental equipment control system. Their safety and stability directly affect the results and achievements of scientific research experiments. By conducting a comprehensive security review of these target objects, the safety and reliability of the experimental equipment control system can be guaranteed, providing reliable technical support for scientific research experiments.

[0054] As an optional embodiment, in 101, first, a system interface or communication protocol can be pre-set to receive instructions from users or other systems. This can be achieved through network communication, such as using HTTP requests or message queues. An instruction receiving module is set in the system to receive and parse the received instructions. This module needs to be able to understand the format and meaning of the instructions and convert them into operations within the system. Once the instructions are received, the system can perform corresponding information acquisition operations based on the target object information in the instructions. Different types of target objects such as experimental equipment files, experimental equipment software, and experimental parameters may require different acquisition methods. For example, for files, the file content can be read directly, for software, the software version and configuration information can be queried, and for parameters, they can be obtained from parameter configuration files or databases. After obtaining the information to be reviewed of the target object, the system marks the target object as a state to be checked for subsequent security review operations. This status mark can be achieved by setting a corresponding flag in the system database or memory, indicating that the target object needs to be reviewed.

[0055] In the embodiment of the present application, it is assumed that the target object is an experimental device file, and the file attributes include file name, size, creation time, modification time, etc. File content: if file content review is allowed, it may include the text content, image data, audio data, etc. File permissions: information such as the read and write permissions and execution permissions of the file.

[0056] Assuming the target object is the experimental equipment software, software version: the software version number and update history. Software permissions: the permissions and access scope required by the software. Software configuration: the configuration information of the software, such as parameter settings, plug-in list, etc.

[0057] Assuming the target object is the experimental parameter, experimental settings: basic information such as the name, description, and purpose of the experiment. Parameter configuration: parameter settings used in the experiment, such as experimental conditions, variable settings, etc. Experimental data: existing experimental data or results, which may include experimental records, log information, data files, etc.

[0058] The above information can be adjusted and expanded according to specific application scenarios and review requirements to ensure a comprehensive review and assessment of the target object.

[0059] In summary, to achieve this step, it is necessary to design and implement the system's instruction receiving module, information acquisition module, and status management module. Through good system design and reasonable functional division, it can ensure that the system can accurately and efficiently respond to instructions, obtain information to be reviewed, and set the target object to the status to be reviewed, laying the foundation for subsequent security review operations.

[0060] 102 , input the to-be-examined information of the target object into a security detection model, and perform a security review on the target object to obtain a review result of the target object.

[0061] Specifically, the security inspection model should have the ability to continuously learn and optimize to adapt to the ever-changing security threats and environment. For example, the continuous improvement of the model can be achieved by introducing online learning algorithms and adaptive model update strategies, thereby improving the effectiveness and accuracy of security inspections.

[0062] As an optional embodiment, the security detection model at least includes the following structures: a feature extraction layer, a group election layer, a random election layer, and an integrated output layer. The feature extraction layer is responsible for extracting key features from the information to be reviewed of the target object. These features can be attributes, parameters or statistical information related to security, which are used to describe the characteristics and features of the target object. Feature extraction can adopt various technologies, such as rule-based feature extraction, statistical feature extraction or deep learning-based feature extraction, to ensure that the extracted features are representative and discriminative. The group election layer aggregates the outputs of multiple security detection sub-models and makes decisions by voting or weighted averaging to improve the accuracy and robustness of the review results. In the group election layer, multiple different security detection sub-models can be included, and each sub-model may use different algorithms or technologies to increase the diversity and coverage of the model. The random election layer introduces random factors, and makes decisions by randomly selecting a part of the security detection sub-model to increase the diversity and robustness of the model. Random election can dynamically select different sub-models for decision-making during each review process, thereby reducing the risk of overfitting of the model and improving the generalization ability of the model. The integrated output layer integrates and combines the outputs of the group election layer and the random election layer to generate the final review result. The integrated output layer can perform weighted fusion on the outputs of each sub-model, or make decisions by voting, etc., to obtain the final comprehensive review result.

[0063] Through the above structure, the security detection model can make full use of the information and decisions of multiple sub-models to improve the accuracy and robustness of the review results, thereby more effectively discovering possible security issues and risks in the target object.

[0064] Based on the above model structure, in 102, the step of inputting the information to be reviewed of the target object into the security detection model and performing a security review on the target object to obtain the review result of the target object can be implemented as follows:

[0065] 201, extracting the to-be-examined features of the target object from the to-be-examined information through a feature extraction layer;

[0066] 202, through the group election layer, inputting the feature to be reviewed into a navigation module in the group election layer, and routing the feature to be reviewed to different election nodes by the navigation module, so that each election node processes the feature to be reviewed respectively, so as to obtain the approval prediction result of each election node;

[0067] 203, through the random election layer, a random matching selection mechanism is used to perform a security review on the feature to be reviewed to obtain a random approval prediction result;

[0068] 204 , dynamically integrating the approval prediction results of each election node and the random approval prediction results through the integrated output layer to obtain the review result of the target object.

[0069] In this way, through the group election layer and the random election layer, multiple review sub-models and random factors are introduced, which increases the diversity and robustness of the model. This can reduce the risk of overfitting of the model and improve the generalization ability of the model, so as to better adapt to the review of target objects of different types and scenarios. The integrated output layer dynamically integrates each election node and the random approval results, which can make full use of the decision information of different review sub-models and improve the accuracy and credibility of the review results. This can more comprehensively evaluate the security of the target object and provide more reasonable review suggestions. Combining the design of the navigation module and the random election layer, the review process and decision-making strategy can be dynamically adjusted according to the actual situation. This can better adapt to different review scenarios and needs and improve the flexibility and adaptability of the system. Due to the use of the integrated output layer, the review results are based on the comprehensive output of multiple review nodes and random approval results, so the review results are more convincing and explainable. This can help users better understand the review results and take corresponding actions and measures.

[0070] In summary, based on the proposed model structure, a comprehensive security review of the target object can be achieved, and more accurate and reliable review results can be provided, providing strong support for subsequent security management and risk control.

[0071] In the embodiment of the present application, the integration mechanism adopted in the dynamic integration processing process is obtained by dynamic feedback and evaluation based on the integration results in the previous cycle.

[0072] This means that each review cycle will use the integration results of the previous cycle to dynamically adjust the integration strategy of the current cycle. The system will record the integration output results of each cycle, including the approval prediction results of each election node and the random approval prediction results. These historical results will be used as a reference to evaluate the performance and effect of the model. According to the performance of historical results, the system will dynamically adjust the parameters in the integration process, such as the weight of each election node, the proportion of random elections, etc. In this way, the model's preferences and decision-making strategies can be adjusted according to actual conditions to improve the quality and stability of the integration results. After each review cycle, the system will perform a performance evaluation on the integration output results, such as accuracy, recall, F1 score and other indicators. Based on the evaluation results, the system will feedback and adjust the integration parameters to optimize the integration process of the next cycle. Based on the results and feedback information of the performance evaluation, the system will update the integration strategy and parameters to adapt to changes in the target object and changes in model performance. This can maintain the robustness and adaptability of the model and ensure that stable and efficient review results can be produced in different scenarios.

[0073] Through the above methods, the system can dynamically adjust the parameters and strategies in the integration process based on historical experience and real-time feedback to improve the quality and credibility of the review results. This can better respond to changes in target objects and security requirements, and provide continuous support and optimization for security reviews.

[0074] In the embodiment of the present application, the features to be reviewed include at least: device file features, device software features, and experimental parameter features.

[0075] In the embodiment of the present application, the features to be reviewed include device file features, device software features, and experimental parameter features. Device file features refer to file-related information on the target device, including but not limited to file type, file size, file permissions, file path, file creation time, file modification time, etc. This information can help determine whether the device has abnormal files or potential security risks, such as files with abnormal permissions, suspicious file paths, etc. Device software features cover relevant information about software or applications installed on the device, including software name, version number, installation path, running status, etc. By analyzing the software features on the device, it can be found whether there are unauthorized software, known vulnerability software, abnormal software behavior and other security issues. Experimental parameter features refer to parameter information generated by the device during the running of the experiment, such as sensor data, experimental records, experimental results, etc. These parameter information can reflect the running status of the device, changes and abnormal situations during the experiment, and help identify possible security risks and problems.

[0076] Based on the above characteristic information, the security detection model can comprehensively review the target object from different angles, identify potential security hazards and risks, and provide effective support and decision-making basis for security management and risk control.

[0077] As an optional embodiment, in 202, the feature to be reviewed is input into the navigation module in the group election layer through the group election layer, and the navigation module routes it to different election nodes, so that each election node processes the feature to be reviewed respectively to obtain the approval prediction result of each election node, which can be implemented as the following steps:

[0078] 301, identifying the attributes of the target object according to the features to be examined by a navigation module; and

[0079] 302, determining multiple routing paths corresponding to the target object based on the attribute recognition result, and routing the to-be-examined feature to different election nodes according to the multiple routing paths;

[0080] 303. Through each election node, the feature to be reviewed is security approved according to its own adaptive evaluation mechanism to obtain the approval prediction result of each election node.

[0081] In the embodiment of the present application, each election node is provided with its own corresponding adaptive evaluation mechanism. In this optional embodiment, each election node is provided with its own corresponding adaptive evaluation mechanism to dynamically adjust the review process and decision-making strategy according to the actual situation. For example, the election node can use historical review data for learning, and adjust its own review standards and decision-making rules based on previous review results and feedback information. By analyzing historical data, the node can identify common security issues and patterns, thereby improving the accuracy and efficiency of the review.

[0082] For example, the election node can monitor the changes of the features to be reviewed and the dynamic situation of the environment in real time, and adjust the review strategy in time according to the monitoring results. For example, if it is found that the changes of certain features are different from the past, the node can take stricter review measures to deal with potential security risks.

[0083] For example, the election node can dynamically adjust the review thresholds and rules based on the current environment and the characteristics of the target object. For example, for certain key features or high-risk scenarios, the node can lower the review threshold to improve security, while for low-risk scenarios, the threshold can be relaxed to reduce false positives.

[0084] For example, the election node can analyze the importance of the features to be reviewed and dynamically adjust the review strategy according to the impact of each feature on the review result. For example, for features with higher importance, the node can pay more attention to their review results, while for minor features, the review standards can be adjusted appropriately.

[0085] For example, election nodes can monitor the performance indicators of their own models, such as accuracy and recall, and adjust the review strategy and parameter settings in a timely manner according to changes in performance indicators. By continuously monitoring model performance, nodes can promptly identify problems and optimize them, thereby improving review effectiveness and stability.

[0086] Based on the above adaptive evaluation mechanism, each election node can dynamically adjust the review strategy and decision-making rules according to the actual situation to improve the accuracy, credibility and adaptability of the review results. This can better cope with different review scenarios and changes in target objects, thereby achieving a more effective security review process.

[0087] It is worth noting that each adaptive evaluation mechanism has specific dynamic evaluation parameters and predictive evaluation methods.

[0088] For example, in historical data learning, dynamic evaluation parameters may include learning rate and historical data weight, while predictive evaluation methods may involve statistical analysis or machine learning algorithms based on historical data, such as decision trees or neural networks. In addition, dynamic parameters for real-time monitoring and feedback may include monitoring frequency and anomaly detection threshold, while predictive evaluation methods may involve real-time data stream processing or anomaly detection algorithms, such as isolation forest or cluster analysis. Possible dynamic parameters for adaptive threshold setting include review pass threshold and false alarm rate control parameters, while predictive evaluation methods may include statistical analysis, ROC curve analysis, or cross-validation. Dynamic parameters for feature importance analysis may involve feature weights and feature selection thresholds, while predictive evaluation methods may include information gain, Gini coefficient, or model weight adjustment. Finally, dynamic parameters for model performance monitoring may include performance indicator thresholds and monitoring cycles, while predictive evaluation methods may include indicator trend analysis or anomaly detection algorithms, such as fluctuation detection or trend analysis. The most appropriate parameters and methods will be selected according to the specific situation during implementation to ensure the effectiveness and adaptability of the evaluation mechanism. The above are examples of possible dynamic evaluation parameters and predictive evaluation methods, and the most appropriate parameters and methods will be selected according to the specific situation during implementation. The selection of these parameters and methods can be adjusted and optimized according to actual scenarios and needs to ensure the effectiveness and adaptability of the evaluation mechanism.

[0089] In addition, each election node can use a deep learning model to learn node division rules when dividing to maximize the node's information gain or minimize the node's impurity. For example, the division rules can be based on the characteristics of the election node, such as file type, size, and other characteristics.

[0090] Furthermore, in order to increase the diversity and generalization ability of the model, randomness can be introduced when dividing the election nodes, for example, randomly selecting a feature subset for division when dividing the election nodes.

[0091] In the above step 303, each election node performs security approval on the to-be-examined feature according to its own adaptive evaluation mechanism to obtain the approval prediction result of each election node, which can be implemented as follows:

[0092] 3031, through the file analysis node, static analysis and dynamic analysis are performed on the received device file features to obtain a first predicted approval result.

[0093] In an embodiment of the present application, the first predicted approval result is used to indicate whether the experimental equipment file of the target object contains potential security risks.

[0094] 3032, through the software evaluation node, adaptively predict the version information and permission settings in the received device software features to obtain a second predicted approval result.

[0095] In the embodiment of the present application, the second predicted approval result is used to indicate the security of the experimental equipment software of the target object.

[0096] 3033, through the parameter review node, based on the received experimental parameter characteristics, the legality and integrity of the parameters are adaptively verified and analyzed to obtain the third prediction approval result.

[0097] In the embodiment of the present application, the third prediction approval result is used to indicate whether the experimental equipment parameters of the target object meet the safety requirements.

[0098] Through multiple nodes for approval, the security of multiple aspects such as device files, software and parameters can be comprehensively considered, thereby improving the comprehensiveness and accuracy of the approval results. Each node reviews a specific type of feature, which can more focusedly identify and evaluate the possible security risks of that type of feature, improving the accuracy and efficiency of the review. Each node adopts an adaptive evaluation mechanism to dynamically adjust the review strategy and parameter settings according to real-time conditions, which can better adapt to the changes and complexity of different types of features and improve the adaptability and accuracy of the review. Covering the review content of multiple aspects such as device files, software and parameters, it can comprehensively check the security of the target object and reduce the possibility of security vulnerabilities and risks. Each node can quickly generate predicted approval results and provide decision makers with information about device security in a timely manner, helping them make more accurate decisions and take necessary security measures in a timely manner.

[0099] Taking all the above beneficial effects into consideration, this multi-node approval can improve the comprehensiveness, accuracy and efficiency of the approval process, and provide more reliable protection for equipment safety.

[0100] In an embodiment of the present application, the approval prediction results of each election node are in at least one of the following forms: binary form, multivariate form, and fractional form. Further optionally, the first predicted approval result is in binary form, the second predicted approval result is in multivariate form, and the third predicted approval result is in fractional form. The first predicted approval result is expressed in binary form, indicating whether there is a security risk. For example, the result may be "passed" or "failed", indicating whether the file contains potential security risks. The second predicted approval result adopts a multivariate form, which can provide richer information. This form may include multiple categories, each category representing a different degree of security or risk level. For example, the result may include multiple levels such as "safe", "suspicious", and "dangerous". The third predicted approval result is expressed in fractional form, usually referring to the security score of the device parameters. This form can provide more refined quantitative information to help users better understand the security of the device parameters. The score is usually in a continuous numerical range, such as 0 to 100, and the higher the score, the higher the security.

[0101] In practical applications, further optionally, the third prediction approval result is a safety confidence score of the device parameters. This means that the result not only provides a safety assessment of the device parameters, but also provides a confidence level for the assessment. The confidence score can reflect the reliability of the assessment result and help users better understand the accuracy and credibility of the assessment result.

[0102] In summary, the use of different forms of approval prediction results can more comprehensively reflect the safety of equipment and provide users with richer information, which helps to make more accurate decisions and take necessary safety measures.

[0103] As an optional embodiment, in 303, after each election node performs security review on the features to be reviewed according to its own adaptive evaluation mechanism to obtain the approval prediction results of each election node, the adaptive parameters in each election node can be optimized and adjusted through the parameter adjustment layer to improve the approval prediction performance of each election node.

[0104] Specifically, in the optional embodiment, in the above steps, the adaptive parameters in each election node are optimized and adjusted through the parameter adjustment layer to improve the approval prediction performance of each election node, which can be implemented as follows:

[0105] Method 1: For file analysis nodes, dynamically adjust static analysis parameters according to the type information, file size information, and file data structure characteristics contained in the device file characteristics to improve the static analysis effect of different types of files; and dynamically adjust dynamic analysis parameters according to the real-time operation characteristics of the experimental equipment and the file behavior change characteristics.

[0106] Optimization of static analysis parameters: According to the type information, size information and data structure characteristics of the device file, the static analysis parameters can be adjusted dynamically, such as adjusting the flexibility and sensitivity of the file parsing algorithm to improve the static analysis effect of different types of files. For example, for files of different formats, the parameters of the parsing algorithm can be adjusted to better identify potential security risks.

[0107] Optimization of dynamic analysis parameters: According to the real-time operation characteristics of the experimental equipment and the characteristics of file behavior changes, dynamic analysis parameters can be dynamically adjusted, such as adjusting the threshold and sensitivity of the monitoring algorithm to more accurately capture the abnormal behavior of the file. For example, according to the dynamic behavior of different files, the parameters of the monitoring algorithm can be adjusted to improve the detection rate and accuracy of abnormal behavior.

[0108] By optimizing the adaptive parameters of the file analysis nodes, the approval prediction performance of device file security can be improved, and the static and dynamic analysis effects of different types of files can be enhanced, thereby more effectively identifying and assessing potential security risks.

[0109] Method 2: For software evaluation nodes, adjust the parameters of the version information prediction model according to the type, release cycle, and update frequency of the software to adapt to the version change characteristics of different software; and adjust the parameters of the permission setting prediction model according to the permission structure and security policy of the software to improve the prediction accuracy of different permission settings.

[0110] Parameters of the version information prediction model: According to the type of software, release cycle and update frequency, the parameters of the version information prediction model can be adjusted, such as adjusting the learning rate and weight of the model to adapt to the characteristics of different software version changes. For example, for frequently updated software, the learning rate of the model can be adjusted to adapt to the characteristics of the new version more quickly.

[0111] Parameters of the permission setting prediction model: According to the permission structure and security policy of the software, the parameters of the permission setting prediction model can be adjusted, such as adjusting the model's feature weights and thresholds, to improve the prediction accuracy for different permission settings. For example, for software with different permission settings, the model's feature weights can be adjusted to more accurately assess its security.

[0112] By optimizing the adaptive parameters of software evaluation nodes, the approval prediction performance of device software security can be improved, and the security under different software versions and permission settings can be more accurately predicted, thereby effectively reducing security risks.

[0113] Method 3: For parameter review nodes, adjust the parameters of the legitimacy verification model according to the type, range, and constraints of the experimental parameters to ensure the accuracy of the legitimacy check of different types of parameters; and adjust the parameters of the integrity verification model according to the changing rules and historical data of the experimental parameters to adapt to the changing needs of parameter integrity analysis.

[0114] Parameters of the legality verification model: According to the type, range and constraints of the experimental parameters, the parameters of the legality verification model can be adjusted, such as adjusting the rules and constraints of the model to ensure the accuracy of the legality check of different types of parameters. For example, for different types of parameters, the constraints of the model can be adjusted to better identify illegal parameter values.

[0115] Parameters of the integrity verification model: According to the variation patterns of experimental parameters and historical data, the parameters of the integrity verification model can be adjusted, such as adjusting the threshold and sensitivity of the model, to adapt to the changes in the needs of parameter integrity analysis. For example, according to the variation patterns of different parameters, the threshold of the model can be adjusted to more accurately identify abnormal changes in parameters.

[0116] By optimizing the adaptive parameters of the parameter review node, the approval prediction performance of the experimental equipment parameter safety can be improved, and the legitimacy and integrity of the parameters can be verified more accurately, thereby effectively ensuring the safety of the equipment parameters.

[0117] In summary, by optimizing and adjusting the adaptive parameters in each election node through these methods, the performance and accuracy of the entire approval system can be improved, thereby more effectively identifying and assessing the security risks of equipment.

[0118] 103. Based on the review result of the target object, set the target object to a corresponding candidate inspection state.

[0119] In an embodiment of the present application, the candidate inspection state includes a security label corresponding to the target object. Specifically, setting the target object to the corresponding candidate inspection state usually involves marking the object as a different state for subsequent review and processing. These candidate inspection states may include security labels for the target object to better identify and track the security status of the object. For example, a security label is a mark used to identify the current security status or risk level of the target object. These labels can be generated based on review results or other security assessment indicators, and reflect the security status of the target object.

[0120] Candidate review status is a classification of different states or stages that a target object may be in during the review process. Each state is usually associated with a specific security label to indicate the security status of the object. For example, a target object is marked as pending review status, indicating that it requires a security review. At this stage, the security label may be unknown or pending, and further review is required to determine the security of the object. If the target object is determined to be safe after review, it can be set to a safe state. In this case, the security label may be "safe" or "passed", indicating that the object does not contain obvious security risks. If the target object is determined to have security risks after review, it can be set to a risky state. In this case, the security label may be "risky", "warning" or other relevant risk level labels to indicate that the object has potential security issues. In some cases, the target object may be set to an abnormal state because its safety cannot be determined by review or abnormal conditions have occurred. In this case, the security label may be "abnormal" or "unknown", indicating that the security of the object requires further investigation or processing.

[0121] By setting the target object to different candidate inspection states and attaching corresponding security tags, the security status of the object can be better managed and tracked, and necessary security measures can be taken in time to deal with potential security risks.

[0122] In the embodiment of the present application, the security tag is associated with the security status of the target object and the object type to which it belongs.

[0123] For different security states, corresponding security labels can be defined to indicate the security of the object. For example, for an object in a safe state, the security label can be "safe" or "passed" to indicate that the object has been determined to be safe after review; for an object in a risky state, the security label can be "risky", "warning", etc. to indicate that the object has potential security issues. Different types of target objects may have different security characteristics and review requirements, so different security labels can be defined according to the object type. For example, for a file object, the security label can reflect the data type, file structure, and possible security risks it contains; for a software object, the security label can reflect its version information, release cycle, and known vulnerabilities. For different security states and object types, security labels can be further subdivided and customized to more accurately describe the security situation of the object. For example, different security labels can be defined for different levels of risk states, such as "low risk", "medium risk", "high risk", etc.; specific security labels can be defined for different types of files, such as "suspicious file", "malicious file", etc.

[0124] By associating with security status and object type, security labels can more comprehensively reflect the security characteristics and review results of target objects, provide more targeted and actionable security information, and help users more accurately understand and handle security issues of objects.

[0125] It is worth noting that the security labels associated with the candidate inspection status can be further expanded to consider more dimensional information, such as security risk level, threat type, impact scope, etc., to provide richer security information and decision support.

[0126] 104 , pushing the to-be-examined information of the target object to the approval channel that matches the candidate inspection status, and performing security approval on the target object.

[0127] In this application, the division of approval channels can be based on factors such as the nature, urgency, complexity, and degree of automation of the approval task. For example, the approval channel can be divided into a manual approval channel and an automated approval channel. The manual approval channel is suitable for complex, high-risk approval tasks that require professional knowledge and judgment. These tasks may involve situations where a large amount of details, subjective judgment, or expertise in a specific field need to be reviewed, and cannot be completely handled by an automated system. The manual approval channel is usually handled by professional reviewers or approval teams, who will carefully review the information to be reviewed, conduct a comprehensive assessment, and make approval decisions based on their professional judgment. The automated approval channel is suitable for approval tasks that are relatively simple, low-risk, or can be automatically judged by rules or algorithms. These tasks may include some routine and repetitive approval processes that can be automated through pre-set rules, models, or algorithms. The automated approval channel is usually handled by computer systems or software, which can quickly and efficiently process the information to be reviewed and generate approval results based on the set rules or algorithms. In practical applications, the division of approval channels can be flexibly adjusted and combined according to specific circumstances. For example, for some complex approval tasks, the manual approval channel can be used for preliminary approval, and then the automated approval channel can be used for quick confirmation or further processing; for some simple approval tasks, they can be directly processed through the automated approval channel to improve approval efficiency and speed.

[0128] In short, manual approval channels and automated approval channels each have their own advantages in the approval process. Reasonable division and combination can improve approval efficiency, accuracy and flexibility, and meet the needs of different approval tasks.

[0129] As an optional embodiment, in 104, further, when pushing the information to be reviewed to the approval channel that matches the candidate inspection status, an intelligent algorithm can be introduced to select the most appropriate approval channel. This can make decisions based on factors such as the characteristics of the target object, the review results, and the load of the approval channel, thereby improving the approval efficiency and accuracy.

[0130] Further optionally, in the process of security review and approval, a real-time review and feedback mechanism can be introduced to timely discover and deal with potential security threats. This can be achieved through integration with real-time monitoring systems, and through a rapid response mechanism to promptly respond to security incidents.

[0131] In addition, further, during the review process, a visual interface for security review results can be provided to intuitively display the security status and related information of the target object, helping users better understand the review results and make decisions.

[0132] In an embodiment of the present application, it can be applied to a distributed experimental equipment management and control system. The distributed experimental equipment management and control system is used to operate and manage experimental equipment. These experimental equipment include different types of equipment. In this scheme, first, in response to the receiving instruction of the target object in the distributed experimental equipment management and control system, the information to be reviewed of the target object is obtained, and the target object is set to a state to be inspected. Among them, the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters. Then, the information to be reviewed of the target object is input into the security detection model, and a security review is performed on the target object to obtain the review result of the target object. Then, based on the review result of the target object, the target object is set to a corresponding candidate inspection state. Finally, the information to be reviewed of the target object is pushed to the approval channel that matches the candidate inspection state, and the security approval of the target object is performed.

[0133] In the embodiment of the present application, it is possible to respond to instructions in real time, quickly obtain information to be reviewed, and conduct security review. This ensures that the control of experimental equipment is timely and effective, and reduces the security risks caused by delays. By reviewing the target object through the security detection model, it is also possible to comprehensively review the experimental equipment files, software and parameters, thereby improving the comprehensiveness and accuracy of the review. In addition, associating the review results with security tags can better manage the security status of the experimental equipment. The setting of security tags can help administrators or users quickly understand the security of the equipment and help to take necessary security measures in a timely manner. It is also possible to automatically push the information to be reviewed to the matching approval channel and perform security approval. This can reduce the cost of manual intervention, improve approval efficiency, and reduce security vulnerabilities caused by human factors. Through the distributed experimental equipment management and control system, the review task can be assigned to different nodes for processing, thereby reducing the load of a single node and improving the concurrent processing capacity and overall performance of the system. In the embodiment of the present application, the automated security initial inspection of the equipment data is realized through the security detection model, and then the equipment data is pushed to different approval channels for further security inspection in combination with routing sorting, which greatly reduces the workload of security detection, helps to improve the operating efficiency and management level of the equipment, and improves the security and reliability of distributed experimental equipment.

[0134] In another embodiment of the present application, a distributed experimental equipment management and control system is also provided, wherein the distributed experimental equipment management and control system is used to operate and manage experimental equipment, and the experimental equipment includes different types of equipment, see Figure 3 The distributed experimental equipment management and control system includes the following units:

[0135] The acquisition unit is configured to, in response to a receiving instruction for a target object in the distributed experimental equipment management and control system, acquire the information to be examined of the target object and set the target object to a state to be examined; the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters;

[0136] A checking unit, configured to input the to-be-examined information of the target object into a security detection model, and perform a security review on the target object to obtain a review result of the target object;

[0137] A setting unit is configured to set the target object to a corresponding candidate inspection state based on the review result of the target object; the candidate inspection state includes a security tag corresponding to the target object; the security tag is associated with the security state of the target object and the object type to which it belongs;

[0138] The pushing unit is configured to push the to-be-examined information of the target object to the approval channel that matches the candidate inspection status, and perform security approval on the target object.

[0139] Further optionally, the security detection model includes at least the following structures: a feature extraction layer, a group election layer, a random election layer, and an integrated output layer; the inspection unit inputs the information to be reviewed of the target object into the security detection model, performs a security review on the target object, and obtains the review result of the target object, and is configured as follows:

[0140] Extracting the features of the target object to be reviewed from the information to be reviewed through the feature extraction layer; the features to be reviewed include at least: device file features, device software features, and experimental parameter features;

[0141] Through the group election layer, the features to be reviewed are input into the navigation module in the group election layer, and the navigation module routes them to different election nodes, so that each election node processes the features to be reviewed respectively, so as to obtain the approval prediction results of each election node;

[0142] Through the random election layer, the features to be reviewed are subjected to security review using a random matching selection mechanism to obtain random approval prediction results;

[0143] Through the integrated output layer, the approval prediction results of each election node and the random approval prediction results are dynamically integrated to obtain the review results of the target object;

[0144] The integration mechanism used in the dynamic integration process is obtained through dynamic feedback and evaluation based on the integration results in the previous cycle.

[0145] Further optionally, the inspection unit inputs the feature to be reviewed into a navigation module in the group election layer through the group election layer, and the navigation module routes the feature to be reviewed to different election nodes, so that each election node processes the feature to be reviewed respectively to obtain the approval prediction result of each election node, and is configured as follows:

[0146] Using a navigation module, identifying attributes of the target object according to the features to be examined; and

[0147] Determine multiple routing paths corresponding to the target object based on the attribute recognition result, and route the to-be-examined features to different election nodes according to the multiple routing paths;

[0148] Through each election node, the features to be reviewed are security approved according to their respective adaptive evaluation mechanisms to obtain the approval prediction results of each election node;

[0149] Among them, each election node is provided with its own corresponding adaptive evaluation mechanism; each adaptive evaluation mechanism has specific dynamic evaluation parameters and predictive evaluation methods.

[0150] Further optionally, the inspection unit performs security approval on the to-be-examined feature through each election node according to the respective adaptive evaluation mechanism to obtain the approval prediction result of each election node, and is configured as follows:

[0151] Through the file analysis node, static analysis and dynamic analysis are performed on the received device file features to obtain a first prediction approval result; the first prediction approval result is used to indicate whether the experimental device file of the target object contains potential security risks;

[0152] By means of the software evaluation node, adaptively predicting the version information and permission settings in the received device software features to obtain a second prediction approval result; the second prediction approval result is used to indicate the security of the experimental device software of the target object;

[0153] Through the parameter review node, based on the received experimental parameter characteristics, the legality and integrity of the parameters are adaptively verified and analyzed to obtain a third predicted approval result; the third predicted approval result is used to indicate whether the experimental equipment parameters of the target object meet the safety requirements.

[0154] Further optionally, the approval prediction results of each election node have at least one of the following forms: binary form, multivariate form, and fractional form.

[0155] Further optionally, the first prediction approval result is in binary form, the second prediction approval result is in multivariate form, and the third prediction approval result is in fractional form; the third prediction approval result is a safety confidence score of the equipment parameters.

[0156] Further optionally, the inspection unit is also configured to: through each election node, perform security review of the features to be reviewed according to their respective adaptive evaluation mechanisms, so as to obtain the approval prediction results of each election node, and then, through the parameter adjustment layer, optimize and adjust the adaptive parameters in each election node to improve the approval prediction performance of each election node.

[0157] Further optionally, the inspection unit optimizes and adjusts the adaptive parameters in each election node through the parameter adjustment layer to improve the approval prediction performance of each election node, and is configured as follows:

[0158] For the file analysis node, dynamically adjust the static analysis parameters according to the type information, file size information, and file data structure characteristics contained in the device file characteristics to improve the static analysis effect of different types of files; and,

[0159] Dynamically adjust dynamic analysis parameters according to the real-time operation characteristics of the experimental equipment and the characteristics of file behavior changes;

[0160] For software evaluation nodes, adjust the parameters of the version information prediction model according to the software type, release cycle, and update frequency to adapt to the version change characteristics of different software; and

[0161] According to the software's permission structure and security policy, adjust the parameters of the permission setting prediction model to improve the prediction accuracy of different permission settings;

[0162] For parameter review nodes, adjust the parameters of the legality verification model according to the type, range, and constraints of the experimental parameters to ensure the accuracy of the legality check of different types of parameters;

[0163] According to the changing rules of experimental parameters and historical data, the parameters of the integrity verification model are adjusted to adapt to the changing needs of parameter integrity analysis.

[0164] In the embodiment of the present application, an automated initial security check of device data is implemented through a security detection model, and then combined with routing sorting, the device data is pushed to different approval channels for further security checks, greatly reducing the workload of security detection.

[0165] In another embodiment of the present application, an intelligent computing platform is further provided, comprising: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0166] Memory, used to store computer programs;

[0167] The processor is used to implement the security inspection method based on distributed experimental equipment described in the method embodiment when executing the program stored in the memory.

[0168] The communication bus 1140 mentioned in the above electronic device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 1140 may be divided into an address bus, a data bus, a control bus, etc.

[0169] The embodiment of the present application provides a safety inspection method based on distributed experimental equipment for constructing a low-energy computing unit.

[0170] For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0171] The communication interface 1120 is used for communication between the above electronic device and other devices.

[0172] The memory 1130 may include a random access memory (RAM) or a non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0173] The above-mentioned processor 1110 can be a general-purpose processor, including an artificial intelligence processor, a graphics processing unit (GPU), an artificial intelligence processor card (Machine Learning Unit, MLU), a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0174] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed, can implement each step that can be executed by an electronic device in the above method embodiment.

Claims

1. A safety inspection method based on distributed experimental equipment, characterized in that: Applied to a distributed experimental equipment management and control system, the distributed experimental equipment management and control system is used to operate and manage experimental equipment, the experimental equipment includes different types of equipment; the security inspection method based on distributed experimental equipment includes: In response to receiving instructions for a target object in the distributed experimental equipment management and control system, obtaining information to be reviewed of the target object and setting the target object to a state to be reviewed; the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters; Inputting the to-be-reviewed information of the target object into the security detection model, and performing a security review on the target object to obtain a review result of the target object; Based on the review result of the target object, the target object is set to a corresponding candidate inspection state; the candidate inspection state includes a security tag corresponding to the target object; the security tag is associated with the security state of the target object and the object type to which it belongs; The information to be reviewed of the target object is pushed to the approval channel that matches the candidate inspection status, and the security approval of the target object is performed.

2. The safety inspection method based on distributed experimental equipment according to claim 1 is characterized in that: The security detection model at least includes the following structures: feature extraction layer, group election layer, random election layer, and integrated output layer; The step of inputting the information to be reviewed of the target object into the security detection model and performing a security review on the target object to obtain a review result of the target object includes: Extracting the features of the target object to be reviewed from the information to be reviewed through the feature extraction layer; the features to be reviewed include at least: device file features, device software features, and experimental parameter features; Through the group election layer, the features to be reviewed are input into the navigation module in the group election layer, and the navigation module routes them to different election nodes, so that each election node processes the features to be reviewed respectively, so as to obtain the approval prediction results of each election node; Through the random election layer, the features to be reviewed are subjected to security review using a random matching selection mechanism to obtain random approval prediction results; Through the integrated output layer, the approval prediction results of each election node and the random approval prediction results are dynamically integrated to obtain the review results of the target object; The integration mechanism used in the dynamic integration process is obtained through dynamic feedback and evaluation based on the integration results in the previous cycle.

3. The safety inspection method based on distributed experimental equipment according to claim 2 is characterized in that: The feature to be reviewed is input into the navigation module in the group election layer through the group election layer, and the navigation module routes the feature to be reviewed to different election nodes, so that each election node processes the feature to be reviewed respectively to obtain the approval prediction result of each election node, including: By means of a navigation module, attributes of the target object are identified according to the features to be examined; and multiple routing paths corresponding to the target object are determined based on the attribute identification results, and the features to be examined are respectively routed to different election nodes according to the multiple routing paths; Through each election node, the features to be reviewed are security approved according to their respective adaptive evaluation mechanisms to obtain the approval prediction results of each election node; Among them, each election node is provided with its own corresponding adaptive evaluation mechanism; each adaptive evaluation mechanism has specific dynamic evaluation parameters and predictive evaluation methods.

4. The safety inspection method based on distributed experimental equipment according to claim 3 is characterized in that: The security approval of the to-be-examined features is performed by each election node according to its own adaptive evaluation mechanism to obtain the approval prediction results of each election node, including: Through the file analysis node, static analysis and dynamic analysis are performed on the received device file features to obtain a first prediction approval result; the first prediction approval result is used to indicate whether the experimental device file of the target object contains potential security risks; By means of the software evaluation node, adaptively predicting the version information and permission settings in the received device software features to obtain a second prediction approval result; the second prediction approval result is used to indicate the security of the experimental device software of the target object; Through the parameter review node, based on the received experimental parameter characteristics, the legality and integrity of the parameters are adaptively verified and analyzed to obtain a third predicted approval result; the third predicted approval result is used to indicate whether the experimental equipment parameters of the target object meet the safety requirements.

5. The safety inspection method based on distributed experimental equipment according to claim 4 is characterized in that: The approval prediction results of each election node have at least one of the following forms: binary form, multivariate form, and fractional form.

6. The safety inspection method based on distributed experimental equipment according to claim 5 is characterized in that: The first predicted approval result is in binary form, the second predicted approval result is in multivariate form, and the third predicted approval result is in fractional form; The third prediction approval result is a safety confidence score of the equipment parameters.

7. The safety inspection method based on distributed experimental equipment according to claim 4 is characterized in that: After the security approval of the to-be-examined features by each election node according to the respective adaptive evaluation mechanisms to obtain the approval prediction results of each election node, the method further includes: Through the parameter adjustment layer, the adaptive parameters in each election node are optimized and adjusted to improve the approval prediction performance of each election node.

8. The safety inspection method based on distributed experimental equipment according to claim 7 is characterized in that: The parameter adjustment layer optimizes and adjusts the adaptive parameters in each election node to improve the approval prediction performance of each election node, including: For file analysis nodes, dynamically adjust static analysis parameters according to the type information, file size information, and file data structure characteristics contained in the device file characteristics to improve the static analysis effect of different types of files; and dynamically adjust dynamic analysis parameters according to the real-time operation characteristics of the experimental equipment and the characteristics of file behavior changes; For software evaluation nodes, adjust the parameters of the version information prediction model according to the software type, release cycle, and update frequency to adapt to the version change characteristics of different software; and According to the software's permission structure and security policy, adjust the parameters of the permission setting prediction model to improve the prediction accuracy of different permission settings; For parameter review nodes, adjust the parameters of the legality verification model according to the type, range, and constraints of the experimental parameters to ensure the accuracy of the legality check of different types of parameters; According to the changing rules of experimental parameters and historical data, the parameters of the integrity verification model are adjusted to adapt to the changing needs of parameter integrity analysis.

9. A distributed experimental equipment management and control system, characterized in that: The distributed experimental equipment management and control system is used to operate and manage experimental equipment, and the experimental equipment includes different types of equipment. The distributed experimental equipment management and control system includes: The acquisition unit is configured to, in response to a receiving instruction for a target object in the distributed experimental equipment management and control system, acquire the information to be examined of the target object and set the target object to a state to be examined; the target object includes at least one of the following: experimental equipment files, experimental equipment software, and experimental parameters; A checking unit, configured to input the to-be-examined information of the target object into a security detection model, and perform a security review on the target object to obtain a review result of the target object; A setting unit is configured to set the target object to a corresponding candidate inspection state based on the review result of the target object; the candidate inspection state includes a security tag corresponding to the target object; the security tag is associated with the security state of the target object and the object type to which it belongs; The pushing unit is configured to push the to-be-examined information of the target object to the approval channel that matches the candidate inspection status, and perform security approval on the target object.

10. An intelligent computing platform, characterized in that: The intelligent computing platform includes: at least one processor, memory, and input-output unit; In which, the memory is used to store computer programs, the processor is used to call the computer program stored in the memory to execute the safety inspection method based on distributed experimental equipment as described in any one of claims 1 to 8, and the input and output unit is used to receive user input and display output information of the computer program stored in the memory.

Images