Terminal management and control method and device based on artificial intelligence and platform equipment

Abstract

The invention discloses a terminal management and control method and device based on artificial intelligence and platform equipment, and the method applied to the platform equipment comprises the steps: receiving a strategy configuration operation through a strategy configuration interface; the strategy configuration operation comprises natural language data input into the strategy configuration interface; processing the natural language data by using a target model based on natural language processing to generate a corresponding standardized configuration instruction; the standardized configuration instruction is sent to gateway equipment, and the gateway equipment sends a corresponding target configuration instruction to a target terminal according to the standardized configuration instruction; and the target terminal executes the target configuration instruction.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a terminal management method, device, and platform equipment based on artificial intelligence. Background Technology

[0002] Currently, in scenarios involving the management and control of multiple terminals, operations and maintenance personnel need to master the syntax of management and control commands and memorize the parameter formats of these commands. This results in a reliance on manual intervention for terminal management and control, leading to low efficiency in terminal management. Summary of the Invention

[0003] In view of this, this application provides a terminal management method, apparatus, and platform device based on artificial intelligence, as follows:

[0004] An artificial intelligence-based terminal management method, applied to platform devices, the method comprising:

[0005] The policy configuration interface receives policy configuration operations; the policy configuration operations include natural language data input into the policy configuration interface.

[0006] The natural language data is processed using a target model based on natural language processing to generate corresponding standardized configuration instructions.

[0007] The standardized configuration instruction is sent to the gateway device, and the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction; the target terminal executes the target configuration instruction.

[0008] Optionally, the above method utilizes a target model based on natural language processing to process the natural language data and generate corresponding standardized configuration instructions, including:

[0009] Using a target model based on natural language processing, a terminal configuration strategy represented by the natural language data is obtained; the terminal configuration strategy includes at least configuration parameters for managing and controlling the terminal.

[0010] The terminal configuration policy is converted into standardized configuration instructions.

[0011] Optionally, the above method utilizes a target model based on natural language processing to obtain the terminal configuration strategy represented by the natural language data, including:

[0012] The natural language data is segmented to obtain the word sequence corresponding to the natural language data;

[0013] The word sequence is input into a target model based on natural language processing to obtain the terminal control intent output by the target model;

[0014] A terminal configuration policy is obtained based on the parameter entity words in the word sequence and the terminal control intent. The terminal configuration policy includes multiple key-value pairs. The multiple key-value pairs include at least the terminal control intent and the parameter entity words.

[0015] Optionally, the above method converts the terminal configuration policy into standardized configuration instructions, including:

[0016] Obtain the general instruction template corresponding to the terminal configuration policy;

[0017] The key-value pairs in the terminal configuration strategy are converted according to the general instruction template to obtain standardized configuration instructions;

[0018] The standardized configuration instructions include at least a number of configuration parameters.

[0019] Optionally, the above method may further include:

[0020] The terminal configuration policy is saved to the policy set corresponding to the terminal group to which the target terminal is located;

[0021] Wherein, terminals in the same terminal group have similar terminal attributes; the policy set includes at least one terminal configuration policy.

[0022] Optionally, in the above method, the terminal configuration policies in the policy set correspond to policy permission information;

[0023] The policy permission information is used to verify whether the requester has the permission to control the terminal according to the requested terminal configuration policy when the terminal configuration policy is requested.

[0024] The above methods may optionally include:

[0025] The gateway device receives the execution result of the target configuration instruction sent by the target terminal and the terminal performance data of the target terminal during the execution of the target configuration instruction.

[0026] Based on the execution result of the instruction and the terminal performance data, at least one configuration parameter of the terminal configuration strategy is adjusted.

[0027] The above methods may optionally include:

[0028] The gateway device receives terminal operation data collected by the target terminal and the execution results of the target configuration instructions.

[0029] By analyzing the terminal's operational data and the instruction execution results, risk prediction results are obtained.

[0030] Based on the risk prediction results, generate configuration repair instructions;

[0031] According to the configuration repair instruction, the corresponding configuration repair instruction is sent to the target terminal through the gateway device.

[0032] An artificial intelligence-based terminal control device, applied to platform equipment, the device comprising:

[0033] An operation receiving unit is used to receive policy configuration operations through a policy configuration interface; the policy configuration operations include natural language data input into the policy configuration interface.

[0034] The data processing unit is used to process the natural language data using a target model based on natural language processing, and generate corresponding standardized configuration instructions.

[0035] The instruction sending unit is used to send the standardized configuration instruction to the gateway device, the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction, and the target terminal executes the target configuration instruction.

[0036] A platform device, comprising:

[0037] A memory used to store computer programs and the data generated by the execution of said computer programs;

[0038] A processor for executing the computer program to achieve:

[0039] The policy configuration interface receives policy configuration operations; the policy configuration operations include natural language data input into the policy configuration interface.

[0040] The natural language data is processed using a target model based on natural language processing to generate corresponding standardized configuration instructions.

[0041] The standardized configuration instruction is sent to the gateway device, and the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction; the target terminal executes the target configuration instruction.

[0042] As can be seen from the above technical solutions, the terminal management method, apparatus, and platform device disclosed in this application, based on artificial intelligence, can provide a policy configuration interface where users can perform policy configuration operations, i.e., input natural language data. Therefore, in this embodiment, a target model based on natural language processing can be used to process the input natural language data, generating corresponding standardized configuration instructions. Based on these standardized configuration instructions, a gateway device can be used to send corresponding target configuration instructions to the target terminal, which then executes the instructions. It is evident that in this application, the user only needs to describe the desired terminal management method using natural language data on the policy configuration interface. Then, the target model based on natural language processing can be used to convert the user-input natural language data into standardized configuration instructions, which are then sent to the target terminal via the gateway device. Therefore, this application does not require users to master the instruction syntax of different terminals, nor does it require users to memorize the parameter formats of instructions; it is independent of manual intervention, thereby improving the efficiency of terminal management. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A flowchart illustrating an artificial intelligence-based terminal management method provided in this application embodiment;

[0045] Figure 2 This is a schematic diagram of the platform device architecture;

[0046] Figure 3 This is an example diagram of the policy configuration interface in an embodiment of this application;

[0047] Figure 4 A partial flowchart of an artificial intelligence-based terminal management method provided in an embodiment of this application;

[0048] Figure 5 Another flowchart of an artificial intelligence-based terminal management method provided in this application embodiment;

[0049] Figure 6 This is another part of a flowchart of an artificial intelligence-based terminal management method provided in an embodiment of this application;

[0050] Figure 7 A schematic diagram of the structure of an artificial intelligence-based terminal control device provided in an embodiment of this application;

[0051] Figure 8 Another structural schematic diagram of an artificial intelligence-based terminal control device provided in an embodiment of this application;

[0052] Figure 9 This application provides a schematic diagram of the structure of a platform device according to an embodiment of the present application.

[0053] Figure 10 This is a flowchart illustrating the specific execution steps of this application. Detailed Implementation

[0054] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0055] refer to Figure 1 The diagram shown is a flowchart illustrating the implementation of an artificial intelligence-based terminal management method according to an embodiment of this application. This method can be applied to platform devices that connect multiple terminals via a gateway device. The platform device can be a computer or server, etc. Figure 2 The platform device shown has a gateway device connected to it. Simultaneously, the gateway device establishes an encrypted, bidirectional communication link with each terminal. The technical solution in this embodiment is primarily used to improve the efficiency of terminal management.

[0056] Specifically, the method in this embodiment may include the following steps:

[0057] Step 101: Receive policy configuration operations through the policy configuration interface.

[0058] For example, such as Figure 3 As shown, a policy configuration interface can be output on the platform device, and operation and maintenance personnel can perform policy configuration operations on the policy configuration interface.

[0059] The policy configuration operation can include natural language data entered into the policy configuration interface. For example, operations and maintenance personnel can enter "Set full-screen watermark for all terminals in the R&D department" or "Clean up redundant memory of terminals in the marketing department" in the policy configuration interface.

[0060] Step 102: Use the target model based on natural language processing to process natural language data and generate corresponding standardized configuration instructions.

[0061] The target model based on natural language processing can be an artificial intelligence model deployed with a parsing engine using Natural Language Processing (NLP), such as the Transformer-based Bidirectional Encoder Representations from Transformers (BERT) model. Based on this, this embodiment can use NLP to understand the natural language data input by operations and maintenance personnel and generate standardized configuration instructions accordingly. The configuration parameters in the standardized configuration instructions represent the control method (how to do it), the control content (what to do), and the control subject (who).

[0062] For example, taking the natural language data "set full-screen watermark for all terminals in the R&D department", the generated standardized configuration command can include: R&D department.watermark setting.full-screen display, which represents setting a watermark (control content) for the terminals (control subject) of the R&D department and setting it to full-screen display (control method).

[0063] Step 103: Send the standardized configuration command to the gateway device.

[0064] In this process, the gateway device sends corresponding target configuration instructions to the target terminal based on standardized configuration commands. The target terminal then executes these instructions, enabling control over the target terminal. For example, setting the R&D department's terminals to full-screen watermark, or clearing redundant memory from the marketing department's terminals.

[0065] Specifically, the gateway device can query the current terminal attributes, such as the operating system type, system version, and hardware configuration, from the dynamic rule base according to the standardized configuration instructions. It can also query the current conversion rule that matches the current terminal attributes from the dynamic rule base and convert the standardized configuration instructions into target configuration instructions that match the current terminal attributes according to the current conversion rule. After that, the gateway device transmits the target configuration instructions to the target terminal corresponding to the current terminal attributes.

[0066] The dynamic rule base records at least one mapping relationship between terminal attributes and conversion rules. For example, the dynamic rule base records the terminal attributes of all terminals that may connect to the gateway device. When a terminal goes online for the first time, it can upload its device attributes to the gateway device. Accordingly, each device attribute in the dynamic rule base is mapped to a corresponding conversion rule. The conversion rule represents the correspondence between standardized configuration instructions and target configuration instructions that match the corresponding terminal attributes.

[0067] As can be seen from the above technical solution, the terminal management method based on artificial intelligence provided in this application embodiment can provide a policy configuration interface, where users can perform policy configuration operations, i.e., input natural language data. Therefore, this embodiment can utilize a target model based on natural language processing to process the input natural language data and generate corresponding standardized configuration instructions. Based on this, according to the standardized configuration instructions, the corresponding target configuration instructions can be sent to the target terminal through a gateway device, and the target terminal can execute the target configuration instructions. It is evident that in this application, users only need to describe the desired terminal management method using natural language data on the policy configuration interface. Then, the target model based on natural language processing can convert the user-input natural language data into standardized configuration instructions, and then send the corresponding target configuration instructions to the target terminal through a gateway device. Therefore, this embodiment does not require users to master the instruction syntax of different terminals, nor does it require users to memorize the parameter formats of the instructions. It does not rely on manual intervention, thereby improving the efficiency of terminal management.

[0068] In one implementation, step 102, when processing the natural language data using a target model based on natural language processing to generate corresponding standardized configuration instructions, can be achieved in the following way: Figure 4 As shown:

[0069] Step 401: Use the target model based on natural language processing to obtain the terminal configuration strategy represented by natural language data.

[0070] The terminal configuration policy includes at least the configuration parameters for managing and controlling the terminal.

[0071] Specifically, the terminal configuration policy can be in JSON format.

[0072] In one implementation, step 401 can obtain the terminal configuration policy in the following way:

[0073] First, the natural language data is segmented to obtain the corresponding word sequence. Specifically, in this embodiment, a specific segmentation algorithm, such as the jieba segmentation algorithm, can be used to split the natural language data to obtain the corresponding word sequence. Additionally, in this embodiment, the dictionary used by the segmentation algorithm can be supplemented with vocabulary specific to the terminal management domain to ensure the accuracy of the segmentation. For example, in this embodiment, "set full-screen watermark for all terminals in the R&D department" can be split into "R&D department, terminal, settings, full-screen watermark".

[0074] Then, the word sequence is input into the target model based on natural language processing to obtain the terminal control intent output by the target model. Specifically, in this embodiment, the word sequence can be processed based on a pre-trained and fine-tuned BERT model to obtain the terminal control intent output by the BERT model, such as the intent to "set a watermark".

[0075] Finally, based on the parameter entity words and terminal control intent in the vocabulary sequence, a terminal configuration strategy is obtained. The terminal configuration strategy may include multiple key-value pairs; these multiple key-value pairs include at least terminal control intent and parameter entity words. Specifically, in this embodiment, a Conditional Random Field (CRF) algorithm can be used to extract key parameter entity words from the vocabulary sequence, and then combined with entity word annotation rules in the terminal control domain to construct structured key-value pairs.

[0076] For example, based on the natural language data input by the operations and maintenance personnel, such as "Set full-screen watermark for all terminals in the R&D department", a terminal configuration policy can be constructed: {Terminal group: R&D department, Terminal type: Linux, Operation: Watermark setting, Parameter: Full-screen display}.

[0077] It should be noted that the key-value pairs in the terminal configuration policy are structured information.

[0078] Step 402: Convert the terminal configuration policy into standardized configuration instructions.

[0079] In one implementation, step 402 can be converted into standardized configuration instructions in the following way:

[0080] First, obtain the general instruction template corresponding to the terminal configuration policy; for example, in this embodiment, the system's built-in general instruction template library of the platform device can be called. In the general instruction template library, based on the key-value pairs in the terminal configuration policy, the corresponding general instruction template is found. The general instruction template is a template with a fixed format, such as: {terminal group}.{operation}.{parameter}.

[0081] Then, the multiple key-value pairs in the terminal configuration policy are converted according to the general instruction template to obtain standardized configuration instructions. These standardized configuration instructions include at least several configuration parameters.

[0082] It should be noted that standardized configuration commands do not depend on specific system syntax; they only describe "who, what, and how." For example, taking the general command template "{terminal group}.{operation}.{parameter}" as an example, the key-value pairs in the terminal configuration policy are converted according to the general command template to obtain the standardized configuration command: "R&D Department.Watermark Settings.Full Screen Display," where the configuration parameters are: terminal group is "R&D Department," operation is "watermark settings," and parameter is "full screen display."

[0083] As can be seen, in this embodiment, natural language data can be parsed through NLP, thereby converting the natural language data into executable standardized configuration instructions. After being converted by the gateway device, the native instructions that can be executed by the target terminal are obtained, namely the target configuration instructions.

[0084] In one implementation, this embodiment may further include the following processing:

[0085] Save the terminal configuration policy to the policy set corresponding to the terminal group to which the target terminal belongs.

[0086] Among them, terminals in the same terminal group have similar terminal attributes, and the policy set can include at least one terminal configuration policy.

[0087] It should be noted that in this embodiment, all terminals connected to the gateway device can be grouped on the platform device. Specifically, the terminals can be grouped according to similarity criteria. A similarity criterion could be that the similarity between the terminal states is greater than or equal to a similarity threshold.

[0088] The terminal attributes can include at least one of the following: terminal hardware attributes, terminal operating environment, and terminal service type. For example, terminal hardware attributes can include the model of the terminal's Central Processing Unit (CPU), the size of the terminal's memory, and the capacity of the terminal's disk; the terminal operating environment can include the operating system version and network environment such as wired / wireless; and the terminal service type can include R&D terminals, office terminals, production terminals, etc. Based on this, in this embodiment, terminals with similar attributes are grouped into the same terminal group. Each terminal group corresponds to a policy set, which includes terminal configuration policies applicable to the management of each terminal within the group.

[0089] As can be seen, in this embodiment, after obtaining the terminal configuration policy, the terminal configuration policy is saved according to the terminal group to which the terminal belongs. This allows maintenance personnel to directly select one of the terminal configuration policies from the policy set when they need to manage terminals again. Based on this, the terminal configuration policy selected by the maintenance personnel can be directly converted into standardized configuration instructions on the platform device and sent to the gateway device. After conversion by the gateway device, the corresponding target configuration instructions are sent to the corresponding target terminal, thereby further improving the efficiency of terminal management.

[0090] Based on the above implementation, each terminal configuration policy in the policy set corresponds to policy permission information. This policy permission information is used to verify whether the requester has the permission to control the terminal according to the requested terminal configuration policy when the terminal configuration policy is requested and selected.

[0091] In this context, the requester refers to the operations and maintenance personnel who select a terminal configuration policy on the platform device. When the requester selects a terminal configuration policy, this embodiment can verify whether the requester has the authority to manage the terminal according to the selected policy based on the policy permission information corresponding to the terminal configuration policy. If the verification passes, meaning the requester has the authority to manage the terminal according to the selected policy, the terminal configuration policy can be converted into standardized configuration instructions and sent to the gateway device. After conversion by the gateway device, the corresponding target configuration instructions are sent to the corresponding target terminal, thereby improving terminal management efficiency. If the verification fails, a policy adjustment prompt can be output, which prompts the requester to change the terminal configuration policy.

[0092] It should be noted that policy permission information can include: control permissions corresponding to the terminal, and control permissions corresponding to the terminal's configuration policy. Specifically, the verification can be considered successful if the requester's permissions match both of the following two control permissions: the control permissions corresponding to the terminal in the policy permission information, and the control permissions corresponding to the terminal's configuration policy in the policy permission information.

[0093] For example, if the operations and maintenance personnel have administrator privileges, but the selected terminal configuration policy requires root-level privileges, the verification will fail. Similarly, if the operations and maintenance personnel have regular user privileges, but the managed terminal requires administrator privileges, the verification will also fail.

[0094] Therefore, this embodiment can improve the security of terminal management.

[0095] Based on the above implementation scheme, the following processing can also be performed in this embodiment, such as... Figure 5 As shown:

[0096] Step 501: Receive the execution result of the target configuration command sent by the target terminal and the terminal performance data of the target terminal during the execution of the target configuration command through the gateway device.

[0097] The instruction execution result indicates whether the target configuration instruction was executed successfully. If the target configuration instruction fails to execute, the instruction execution result may also include the reason for the failure. Terminal performance data includes the terminal's status data across multiple performance metrics. For example, resource utilization may increase or decrease during the execution of the target configuration instruction by the target terminal.

[0098] Step 502: Adjust at least one configuration parameter of the terminal configuration policy based on the instruction execution result and terminal performance data.

[0099] Specifically, based on the command execution results and terminal performance parameters, the parameter values ​​of the configuration parameters are adjusted up or down. For example, the transparency of the full-screen display is adjusted from 80% to 50%.

[0100] In one implementation, the terminal configuration policy can be adjusted at specific time intervals. For example, in this embodiment, the configuration parameters of the corresponding terminal configuration policy are adjusted weekly based on the instruction execution results and terminal performance data of the previous week. The adjusted terminal configuration policy is then synchronously updated in its policy set.

[0101] As can be seen, the terminal configuration policy can be dynamically adjusted in this embodiment, thereby improving the accuracy of terminal control.

[0102] In one implementation, this embodiment may also include the following processing, such as... Figure 6 As shown:

[0103] Step 601: Receive terminal operation data collected by the target terminal and the execution results of the target configuration instructions through the gateway device.

[0104] For example, terminal operation data may include CPU load, memory usage, network bandwidth, disk usage, etc. The instruction execution result refers to the state of the target terminal after the target configuration instruction is executed. It can be understood as the effect of executing the target configuration instruction on the target terminal, such as whether the watermark is displayed correctly, whether the process starts successfully, and whether the policy takes effect.

[0105] Step 602: Analyze the terminal's running data and instruction execution results to obtain the risk prediction results.

[0106] In this embodiment, the terminal's operating data can be compared with the normal data thresholds based on a historical normal data threshold library (which stores the thresholds corresponding to the terminal's operating data during normal historical operation). Combined with the instruction execution results, anomaly results are obtained. Then, based on the anomaly results, the risk level of the target terminal is assessed to obtain a risk prediction result, characterizing the risk level of the target terminal.

[0107] Specifically, abnormal results include static and dynamic anomalies. For example, static anomalies may include CPU load consistently above 90% or watermark status as "not enabled"; while dynamic anomalies may include memory usage surging from 30% to 90% within 10 minutes.

[0108] Step 603: Generate configuration repair instructions based on the risk prediction results.

[0109] In this embodiment, a corresponding repair plan can be determined based on the risk prediction results, such as no risk, low risk, medium risk, or high risk, and a configuration repair instruction can be generated based on the repair plan.

[0110] It should be noted that different risk levels correspond to different remediation solutions, and the corresponding remediation instructions generated will be different.

[0111] Specifically, in low-risk or no-risk situations, no repair is required, but the terminal configuration policy can be dynamically adjusted based on the command execution results; in medium-risk or high-risk situations, configuration repair commands are generated based on the abnormal reasons corresponding to the abnormal results.

[0112] For example, when the cause of the error is "insufficient permissions", the configuration repair command can be "elevate execution privileges and re-execute the watermark configuration command as the root user"; another example is that when the cause of the error is "high CPU load", the configuration repair command can be "terminate redundant processes (process ID: 12345) and release CPU resources".

[0113] Step 604: Based on the configuration repair instruction, send the corresponding configuration repair instruction to the target terminal through the gateway device.

[0114] Specifically, in this embodiment, the configuration repair command can be sent to the gateway device. The gateway device converts the configuration repair command into an command that can be executed by the target terminal according to the corresponding conversion rules, so that configuration repair can be performed on the target terminal.

[0115] As can be seen, this application can improve the accuracy of terminal management and enhance the operational reliability of the terminal by predicting the risks of the terminal and performing control and repair based on the prediction results.

[0116] refer to Figure 7 This is a schematic diagram of the structure of an artificial intelligence-based terminal control device provided in an embodiment of this application. This device can be applied to platform equipment, such as... Figure 2 The platform device shown may include the following structures:

[0117] The operation receiving unit 701 is used to receive policy configuration operations through the policy configuration interface; the policy configuration operations include natural language data input into the policy configuration interface.

[0118] Data processing unit 702 is used to process the natural language data using a target model based on natural language processing, and generate corresponding standardized configuration instructions;

[0119] The instruction sending unit 703 is used to send the standardized configuration instruction to the gateway device, the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction, and the target terminal executes the target configuration instruction.

[0120] As can be seen from the above technical solution, the artificial intelligence-based terminal management device provided in this application embodiment can provide a policy configuration interface, where users can perform policy configuration operations, i.e., input natural language data. Therefore, this embodiment can utilize a target model based on natural language processing to process the input natural language data, generating corresponding standardized configuration instructions. Based on these standardized configuration instructions, a gateway device can send corresponding target configuration instructions to the target terminal, which then executes the instructions. It is evident that in this embodiment, the user only needs to describe the desired terminal management method using natural language data on the policy configuration interface. Then, the target model based on natural language processing can convert the user-input natural language data into standardized configuration instructions, which are then sent to the target terminal via the gateway device. Therefore, this embodiment does not require users to master the instruction syntax of different terminals, nor does it require users to memorize the parameter formats of instructions; it is not reliant on manual intervention, thereby improving the efficiency of terminal management.

[0121] In one implementation, the data processing unit 702 is specifically used to: obtain a terminal configuration strategy represented by the natural language data using a target model based on natural language processing; the terminal configuration strategy includes at least configuration parameters for controlling the terminal; and convert the terminal configuration strategy into standardized configuration instructions.

[0122] Specifically, when the data processing unit 702 obtains the terminal configuration strategy represented by the natural language data using a target model based on natural language processing, it is used to: perform word segmentation on the natural language data to obtain a word sequence corresponding to the natural language data; input the word sequence into the target model based on natural language processing to obtain the terminal control intent output by the target model; and obtain the terminal configuration strategy based on the parameter entity words in the word sequence and the terminal control intent, wherein the terminal configuration strategy includes multiple key-value pairs; the multiple key-value pairs include at least the terminal control intent and the parameter entity words.

[0123] Specifically, when the data processing unit 702 converts the terminal configuration policy into standardized configuration instructions, it is used to: obtain a general instruction template corresponding to the terminal configuration policy; convert multiple key-value pairs in the terminal configuration policy according to the general instruction template to obtain standardized configuration instructions; wherein, the standardized configuration instructions include at least multiple configuration parameters.

[0124] In one implementation, the data processing unit 702 is further configured to: save the terminal configuration policy to a policy set corresponding to the terminal group to which the target terminal is located; wherein, terminals in the same terminal group have similar terminal attributes; and the policy set includes at least one terminal configuration policy.

[0125] The terminal configuration policies in the policy set correspond to policy permission information; the policy permission information is used to verify whether the requester has the permission to control the terminal according to the requested terminal configuration policy when the terminal configuration policy is requested.

[0126] In one implementation, the data processing unit 702 is further configured to: receive, through the gateway device, the instruction execution result of the target configuration instruction sent by the target terminal and the terminal performance data of the target terminal during the execution of the target configuration instruction; and adjust at least one configuration parameter of the terminal configuration strategy according to the instruction execution result and the terminal performance data.

[0127] In one implementation, the apparatus in this embodiment may further include the following structures, such as... Figure 8 As shown:

[0128] The configuration repair unit 704 is configured to: receive terminal operation data collected by the target terminal and the instruction execution result of the target configuration instruction through the gateway device; parse the terminal operation data and the instruction execution result to obtain a risk prediction result; generate a configuration repair instruction based on the risk prediction result; and send the corresponding configuration repair instruction to the target terminal through the gateway device according to the configuration repair instruction.

[0129] It should be noted that the specific implementation of each unit in this embodiment can be referred to the corresponding content above, and will not be described in detail here.

[0130] refer to Figure 9 This is a schematic diagram of the structure of a platform device provided in an embodiment of this application. The platform device may include the following structures:

[0131] Memory 901 is used to store computer programs and data generated by the execution of said computer programs;

[0132] Processor 902 is configured to execute the computer program to achieve:

[0133] The policy configuration interface receives policy configuration operations; the policy configuration operations include natural language data input into the policy configuration interface.

[0134] The natural language data is processed using a target model based on natural language processing to generate corresponding standardized configuration instructions.

[0135] The standardized configuration instruction is sent to the gateway device, and the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction; the target terminal executes the target configuration instruction.

[0136] In practice, the platform equipment can be deployed as edge servers and cloud servers. The cloud server can provide a policy configuration interface, on which maintenance personnel can perform policy configuration operations to achieve control over the terminals.

[0137] As can be seen from the above technical solution, the platform device provided in this application embodiment can provide a policy configuration interface, where users can perform policy configuration operations, i.e., input natural language data. Therefore, in this embodiment, a target model based on natural language processing can be used to process the input natural language data, generating corresponding standardized configuration instructions. Based on these standardized configuration instructions, a gateway device can send corresponding target configuration instructions to the target terminal, which then executes the instructions. It is evident that in this embodiment, the user only needs to describe the desired terminal management method using natural language data on the policy configuration interface. Then, the target model based on natural language processing can convert the user-input natural language data into standardized configuration instructions, which are then sent to the target terminal via the gateway device. Thus, this embodiment does not require the user to master the instruction syntax of different terminals, nor does it require the user to memorize the parameter formats of the instructions. It is not reliant on manual intervention, thereby improving the efficiency of terminal management.

[0138] It should be noted that the specific implementation of the processor in this embodiment can be referred to the corresponding content above, and will not be described in detail here.

[0139] by Figure 2 Taking the system architecture shown below as an example, the technical solution of this application is illustrated in the following example:

[0140] First, this application proposes an AI-based intelligent interactive system for terminal management and control and its corresponding implementation method. The core innovation lies in solving the shortcomings of existing technologies in the dimensions of interaction, strategy, security and decision-making through three major technical paths: "AI-driven natural language interaction, terminal clustering and grouping + reinforcement learning strategy optimization, and deep binding of permissions and policies", so as to achieve intelligent, integrated and secure management and control of multiple system terminals.

[0141] This application is based on the technical concept of "AI-driven + cross-protocol integration + deep permission binding". It solves the defects of existing technologies in the dimensions of interaction, decision-making, strategy and security by dynamically scheduling hardware resources, multi-protocol gateway communication, and integrated design of AI algorithm model training and permission policy.

[0142] In its specific implementation, this application supports multiple architectures including x86 and ARM; high-performance edge servers are deployed at the edge layer to handle preliminary analysis of real-time terminal data and lightweight AI inference; the cloud layer uses a cloud server cluster composed of Elastic Cloud Servers (ECS), equipped with high-performance graphics processing units (GPUs) for training complex AI models and storing core resources such as full terminal runtime data, policy templates, and permission configurations; the edge servers and cloud servers form a platform device; the network layer uses a multi-protocol gateway based on the Message Queuing Telemetry Transport (MQTT) protocol to achieve cross-protocol communication between the terminal, edge, and cloud, supporting private network penetration and multi-link load balancing to ensure stable data transmission in weak network environments.

[0143] Based on this, maintenance personnel initiate requests through the natural language interface provided by the platform equipment. The platform equipment's NLP parsing engine performs word segmentation, intent recognition, and entity extraction on the natural language data input by the maintenance personnel, breaking down the requirements into structured information (i.e., terminal configuration strategies) such as "terminal group = R&D department, operation = watermark setting, parameter = full-screen display". The instruction standardization module maps the structured information into a cross-system universal instruction template (i.e., standardized configuration instructions, such as abstracted as "{terminal group}.{operation}.{parameter}" format), and then the multi-protocol adaptation gateway automatically generates executable instructions based on the terminal type and issues them.

[0144] In addition, within the platform equipment, the real-time data acquisition module collects second-level data such as CPU load, memory usage, network bandwidth, and process status through the terminal agent module (a lightweight program deployed on each terminal). After preliminary cleaning at the edge layer, this data is uploaded to the cloud layer. The anomaly identification model performs multi-dimensional analysis of historical and real-time data and pushes risk warnings to operations and maintenance personnel. The automatic repair engine, in conjunction with multi-protocol gateways, generates and issues repair instructions based on a preset repair strategy library. After execution by the terminal, the results are returned, forming a closed loop of "perception-warning-repair-verification".

[0145] refer to Figure 10 The specific execution steps of this application are as follows:

[0146] Step 1001: The operations and maintenance personnel initiate a natural language request:

[0147] Operations and maintenance personnel can directly input natural language data representing terminal management needs through the natural language interactive interface (i.e., policy configuration interface) provided by the platform equipment, such as "set full-screen watermark for all terminals in the R&D department" or "clean up redundant memory of terminals in the marketing department", without distinguishing terminal system type or memorizing command syntax.

[0148] Step 1002: The NLP parsing engine performs intent recognition and entity extraction on the natural language request (i.e., natural language data):

[0149] Based on a pre-trained BERT model, and combined with a terminal-specific dictionary, three layers of processing are completed:

[0150] (1) Word segmentation: The jieba word segmentation algorithm is used to split the natural language request to obtain a word sequence. In addition, this application supplements the dictionary with special words for the control domain to ensure the accuracy of word segmentation. For example, "set full screen watermark for all terminals of the R&D department" is split into "R&D department, terminal, settings, full screen watermark";

[0151] (2) Intent recognition: Based on the fine-tuned BERT model, the core control intent (such as "set watermark", "clean memory", "terminate abnormal process") is identified. The input of the BERT model is the word sequence after word segmentation, and the output is the intent category. Invalid semantic words (such as the modal word "give") are excluded.

[0152] (3) Entity extraction: The Conditional Random Field (CRF) algorithm is used to extract key parameter entity words in the vocabulary sequence. Combined with the entity annotation rules in the terminal management field, structured key-value pairs are formed, i.e. structured information. For example, the above requirement is extracted as: {Terminal group: R&D department, Terminal type: Linux, Operation: Watermark setting, Parameter: Full screen display}.

[0153] Step 1003: Instruction Standardization Module: Generate general instructions (i.e., standardized standard instructions):

[0154] (1) Call the system's built-in general instruction template library, and map the structured information parsed by NLP to a fixed format template: {terminal group}.{operation}.{parameter};

[0155] (2) Transform the above structured information into general instructions, such as: R&D Department. Watermark Settings. Full Screen Display. This template does not depend on specific system syntax, but only describes "who, what, and how".

[0156] (3) Template verification: If there are missing parameters (such as "watermark transparency" not specified), the system will automatically pop up a window to ask the maintenance personnel for supplementation, or fill in the default value based on the terminal's historical configuration (such as default transparency of 50%).

[0157] Step 1004: Multi-protocol Adaptation Gateway: Select the protocol and issue the target configuration command:

[0158] (1) Terminal type identification: By deploying Agent programs (lightweight terminal agents) on each terminal, basic terminal information (system type, hardware model, network address) is obtained and reported in advance, and the gateway queries the target terminal's attributes from the cloud database.

[0159] (2) Protocol and instruction adaptation: Select the corresponding communication protocol according to the terminal type, and convert the general template into the target configuration instructions that can be executed by the terminal, as follows:

[0160] ① (Linux Terminal): General Template R&D Department. Watermark Settings. Convert Full-Screen Display to Protocol Executable Commands:

[0161] sudo apt-get install watermark-tool && watermark-tool -modefullscreen -group R&D Department;

[0162] ② (Windows Terminal): Convert to protocol commands:

[0163] reg add "HKCU\Software\WatermarkConfig" / v DisplayMode / t REG_SZ / dFullScreen && taskkill / f / im explorer.exe && start explorer.exe.

[0164] (3) Stable delivery: Select the optimal transmission link through load balancing of multiple links (multiple communication links between the gateway device and the terminal, such as bidirectional and encrypted communication links), and adopt data fragmentation transmission and breakpoint resume technology in weak network environment; carry the terminal's unique identification ID when delivering, to ensure that the instruction is accurately delivered to the target terminal.

[0165] Step 1005: The terminal executes the instruction and returns the execution result:

[0166] (1) After receiving the target configuration instruction, the terminal agent first performs a preliminary legality check;

[0167] (2) After the verification is successful, execute the command operation (such as watermark installation and configuration); if the execution fails, record the reason for the failure.

[0168] (3) After execution, Agent sends the results back through the original communication protocol.

[0169] Step 1006: (Platform Equipment) Real-time Data Acquisition Module: Acquires terminal operation data:

[0170] (1) Collection content: collect basic terminal operation data (CPU load, memory usage, network bandwidth, disk usage) and post-instruction status data (such as whether the watermark is displayed normally, whether the process has started successfully, and whether the policy has taken effect) at the second level.

[0171] (2) Data preprocessing: The Agent performs preliminary filtering on the collected data (removing noisy data, such as CPU instantaneous 100% but duration <1 second is considered abnormal fluctuation and is directly filtered), and then uploads it to the edge server;

[0172] (3) Data flow: The edge server performs lightweight analysis on the data (such as calculating the average CPU load) and then synchronizes it to the distributed database of the cloud ECS cluster (storing historical and real-time data, supporting high-concurrency queries).

[0173] Step 1007: (Platform Equipment) Anomaly Identification Model: Analyze data to identify potential risks.

[0174] (1) The model calls the "historical normal data threshold library" stored in the cloud (e.g., the normal CPU load of the R&D department's Linux terminal is ≤80% and the watermark is "on").

[0175] (2) Compare the real-time operating data of the terminal with the historical normal data threshold to obtain the anomaly identification result (i.e., the anomaly result):

[0176] Static anomalies: such as CPU load consistently above 90% or watermark status being "not enabled" (instruction executed successfully but not taking effect);

[0177] Dynamic anomalies: such as memory usage surging from 30% to 90% within 10 minutes (abnormal trend);

[0178] (3) The model assesses the risk level of anomalies (high / medium / low), and the assessment rules can be configured by the operation and maintenance personnel (e.g., "if it affects the operation of core business, it is high risk").

[0179] Step 1008: Risk Level Assessment:

[0180] Operations and maintenance personnel can configure risk thresholds (e.g., high risk ≥ 80 points, medium risk 60-79 points, low risk < 60 points). The default thresholds are based on industry best practices.

[0181] (1) If the risk level is high (e.g., medium risk / high risk, such as "CPU load remains at 90%+, affecting R&D compilation business"): proceed to the automatic repair process in step 1009;

[0182] (2) If the risk level is low (e.g., low risk or no risk, such as "memory usage occasionally reaches 85%, and then drops back after 10 seconds"): proceed to the strategy optimization process in step 1010.

[0183] Step 1009: Automatic Repair Engine: Generate and issue configuration repair commands:

[0184] (1) The automatic repair engine matches a repair solution based on the cause of the anomaly:

[0185] For example, if the error is due to "insufficient permissions", generate the configuration repair command "elevate execution privileges and re-execute the watermark configuration command as the root user";

[0186] For example, if the cause of the error is "CPU load is too high", generate the configuration repair command "Terminate redundant processes (process ID: 12345) and release CPU resources".

[0187] (2) The repair command is sent to the terminal agent through the multi-protocol gateway. After the terminal executes the command, it sends back the repair result (e.g., "Repair successful: Watermark has been enabled with root privileges and CPU load has dropped to 60%").

[0188] (3) When the repair fails, the engine automatically generates a manual intervention prompt (such as "Repair failed: Dependency components need to be installed manually") and pushes it to the operation and maintenance personnel interface.

[0189] Step 1010: Terminal Clustering Engine: Grouping by Terminal Attributes

[0190] (1) Clustering dimensions: terminal hardware attributes (CPU model, memory size, disk capacity), operating environment (operating system version, network environment: wired / wireless), business type (R&D terminal, office terminal, production terminal);

[0191] (2) Grouping results: Terminals with similar attributes are grouped together, such as "R&D Department - Linux - High-performance terminal" and "Marketing Department - Windows - Ordinary terminal";

[0192] (3) Group storage: The grouping results are synchronized to the cloud terminal group library and support dynamic updates (such as automatically matching the group to which a new terminal belongs).

[0193] Step 1011: Reinforcement Learning Optimization Module: Iterative Optimization of Policy Parameters:

[0194] A reinforcement learning optimization module (in the platform device) is constructed using DQN (Deep Q-Network) to achieve automatic iteration of policy parameters. The core logic is as follows:

[0195] (1) Model definition:

[0196] ①State space S: terminal grouping, current policy parameters (policy parameters of terminal configuration policy), terminal execution feedback (i.e. instruction execution result);

[0197] ②Action Space A: Policy parameter adjustment actions (such as transparency ±10%, CPU threshold ±5%).

[0198] (2) Reward Mechanism Setting:

[0199] Positive rewards: Strategy execution success rate ≥ 95%, terminal resource utilization rate increased by more than 10%;

[0200] Negative rewards: The strategy causes compatibility issues and degrades terminal performance;

[0201] (3) Parameter optimization: For example, for the "Watermarking Strategy for Linux High-Performance Terminal in R&D Department", the initial parameter was "Full-screen display + 80% transparency". After iteration, it was found that "80% transparency affects R&D personnel's viewing of code" (negative reward), so it was automatically adjusted to "Full-screen display + 50% transparency".

[0202] (4) Optimization cycle: The existing technology’s “monthly manual adjustment” has been shortened to “weekly automatic iteration”. The strategy parameters are updated weekly based on the execution data of the previous week and synchronized to the cloud strategy template library.

[0203] Step 1012: Policy Permission Verification Engine: Matching Terminal / User Permissions:

[0204] (1) Permission Query: The policy permission verification engine obtains two types of permission information from the cloud permission configuration library:

[0205] Terminal permissions: The permission level of the target terminal (e.g., regular user, administrator, root);

[0206] Operation user permissions: The scope of permissions of the operations and maintenance personnel who initiated the request (e.g., whether they can issue administrator-level policies);

[0207] (2) Compatibility verification: For example, "Watermark settings require administrator privileges". If the target terminal has ordinary user privileges, the verification will fail and adjustment suggestions will be generated (such as "This terminal does not have administrator privileges. It is recommended to modify the policy to a version that can be executed by ordinary users").

[0208] (3) Verification result: If the verification passes, proceed to the next step; if it fails, return the adjustment suggestion, which will be confirmed by the operation and maintenance personnel and the policy will be regenerated.

[0209] Step 1013: Generate and distribute personalized policies:

[0210] (1) Policy generation: Generate personalized policies (i.e., terminal configuration policies) for each group of terminals.

[0211] (2) Batch distribution: Policies are distributed in batches through a multi-protocol gateway. The terminal Agent receives the corresponding policies according to its own group and they take effect.

[0212] (3) Policy update: If the reinforcement learning module optimizes the parameters in the future, the policy will be automatically updated and reissued without manual intervention.

[0213] Step 1014: (Platform Equipment) Audit Log Module: Records the entire process log.

[0214] Log content: Records key information throughout the entire process, including:

[0215] (1) Operator, operation time, and natural language request text;

[0216] (2) NLP parsing results (i.e., terminal configuration strategy), general instruction template (i.e., standardized configuration instructions), and specific instructions issued (i.e., target configuration instructions).

[0217] (3) Terminal instruction execution results, anomaly identification results, and repair / optimization process;

[0218] (4) Permission verification records, personalized policy content, and issuance time;

[0219] The logs are encrypted and stored in a compliant database within the cloud-based ECS cluster, accessible only to authorized personnel.

[0220] It is evident that this application has the following advantages:

[0221] 1. This application simplifies operation and maintenance and improves management efficiency. By coupling NLP with a multi-protocol gateway, operation and maintenance personnel do not need to master professional instructions or differentiated operations. The time consumption for batch management of multiple terminals is reduced by more than 60% compared with existing solutions, adapting to the needs of multi-system terminal cluster management. Based on this, this application can solve the problem of "high interaction threshold" in existing technologies: existing technologies require operation and maintenance personnel to master professional instructions and parameter formats of multiple systems, which is difficult to operate and prone to errors.

[0222] 2. This application enhances risk prediction capabilities and significantly reduces the risk of business interruption. By using AI models to analyze terminal operation data in real time, it proactively identifies potential risks and triggers automatic repair processes, reducing manual intervention and mitigating the risk of business interruption. Based on this, this application addresses the problem of "fragmented security control" in existing technologies: existing technologies feature dispersed permission configurations and lack verification between policies and terminal permissions.

[0223] 3. This application adapts to terminal performance optimization strategies, reducing the compatibility failure rate from over 25% to below 5%. Through reinforcement learning, it dynamically iterates strategy parameters, shortening the optimization cycle from monthly to weekly, thus improving terminal resource utilization and preventing strategy misoperation. Specifically, this application combines clustering algorithms and reinforcement learning to accurately group terminals according to their attributes and dynamically optimize strategy parameters, achieving self-adaptation of strategies and self-optimization of effects. Based on this, this application solves the problems of "poor strategy adaptability and lack of dynamic optimization" in existing technologies: existing technologies use standardized strategy templates that do not adapt to terminal differences and require manual adjustment.

[0224] 4. This application deeply integrates policy generation with access control, reducing the failure rate of control caused by chaotic access. Simultaneously, the end-to-end audit logs support traceability and meet compliance requirements, eliminating security risks such as data leakage and malware injection. Based on this, this application solves the problem of "fragmented security control" in existing technologies: existing technologies have scattered access control configurations and lack verification between policies and terminal permissions.

[0225] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0226] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0227] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0228] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A terminal control method based on artificial intelligence, characterized in that, Applied to platform devices, the method includes: The policy configuration interface receives policy configuration operations; the policy configuration operations include natural language data input into the policy configuration interface. The natural language data is processed using a target model based on natural language processing to generate corresponding standardized configuration instructions. The standardized configuration instruction is sent to the gateway device, and the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction; the target terminal executes the target configuration instruction.

2. The method according to claim 1, characterized in that, The natural language data is processed using a target model based on natural language processing to generate corresponding standardized configuration instructions, including: Using a target model based on natural language processing, a terminal configuration strategy represented by the natural language data is obtained; the terminal configuration strategy includes at least configuration parameters for managing and controlling the terminal. The terminal configuration policy is converted into standardized configuration instructions.

3. The method according to claim 2, characterized in that, Using a target model based on natural language processing, the terminal configuration strategy represented by the natural language data is obtained, including: The natural language data is segmented to obtain the word sequence corresponding to the natural language data; The word sequence is input into a target model based on natural language processing to obtain the terminal control intent output by the target model; A terminal configuration policy is obtained based on the parameter entity words in the word sequence and the terminal control intent. The terminal configuration policy includes multiple key-value pairs. The multiple key-value pairs include at least the terminal control intent and the parameter entity words.

4. The method according to claim 2, characterized in that, Converting the terminal configuration policy into standardized configuration instructions includes: Obtain the general instruction template corresponding to the terminal configuration policy; The key-value pairs in the terminal configuration strategy are converted according to the general instruction template to obtain standardized configuration instructions; The standardized configuration instructions include at least a number of configuration parameters.

5. The method according to claim 2, characterized in that, The method further includes: The terminal configuration policy is saved to the policy set corresponding to the terminal group to which the target terminal is located; Wherein, terminals in the same terminal group have similar terminal attributes; the policy set includes at least one terminal configuration policy.

6. The method according to claim 5, characterized in that, The terminal configuration policies in the policy set correspond to policy permission information; The policy permission information is used to verify whether the requester has the permission to control the terminal according to the requested terminal configuration policy when the terminal configuration policy is requested.

7. The method according to claim 5, characterized in that, Also includes: The gateway device receives the execution result of the target configuration instruction sent by the target terminal and the terminal performance data of the target terminal during the execution of the target configuration instruction. Based on the execution result of the instruction and the terminal performance data, at least one configuration parameter of the terminal configuration strategy is adjusted.

8. The method according to claim 1, characterized in that, Also includes: The gateway device receives terminal operation data collected by the target terminal and the execution results of the target configuration instructions. By analyzing the terminal's operational data and the instruction execution results, risk prediction results are obtained. Based on the risk prediction results, generate configuration repair instructions; According to the configuration repair instruction, the corresponding configuration repair instruction is sent to the target terminal through the gateway device.

9. A terminal control device based on artificial intelligence, characterized in that, Applied to platform equipment, the device includes: An operation receiving unit is used to receive policy configuration operations through a policy configuration interface; the policy configuration operations include natural language data input into the policy configuration interface. The data processing unit is used to process the natural language data using a target model based on natural language processing, and generate corresponding standardized configuration instructions. The instruction sending unit is used to send the standardized configuration instruction to the gateway device, the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction, and the target terminal executes the target configuration instruction.

10. A platform device, characterized in that, include: A memory used to store computer programs and the data generated by the execution of said computer programs; A processor for executing the computer program to achieve: The policy configuration interface receives policy configuration operations; the policy configuration operations include natural language data input into the policy configuration interface. The natural language data is processed using a target model based on natural language processing to generate corresponding standardized configuration instructions. The standardized configuration instruction is sent to the gateway device, and the gateway device sends a corresponding target configuration instruction to the target terminal according to the standardized configuration instruction; the target terminal executes the target configuration instruction.

Images