An intelligent computing center cloud platform computing power driven intelligent agent network security auditing method and device

Abstract

The application provides a computing power driven intelligent agent network security auditing method and device of an intelligent computing center cloud platform, relates to the technical fields of intelligent computing centers, intelligent computing centers and computing power infrastructure, and the method is executed by a first intelligent agent and includes the following steps: step S1: broadcast information is sent and broadcast information sent by other intelligent agents is received; step S2: a maintained routing table and an intelligent agent capability table are updated according to the received broadcast information; step S3: a user task sent by an edge intelligent agent is received, the user task is decomposed into multiple subtasks, a second intelligent agent executing the subtasks is selected according to the maintained intelligent agent capability table and the routing table, and the subtasks are distributed to the second intelligent agent; step S4: a subtask execution result is received, the subtask execution result is subjected to security auditing, and the subtask execution result is discarded or modified if the security auditing is not passed; and step S5: a task execution result is generated based on the subtask execution result that passes the security auditing and is sent to the edge intelligent agent.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent computing centers, wisdom computing centers and computing power infrastructure, in particular to a computing power driven intelligent agent network security auditing method and device of an intelligent computing center cloud platform. BACKGROUND

[0002] With the rapid development of artificial intelligence technology, "intelligent computing centers" and "wisdom computing centers" have emerged.

[0003] An "intelligent computing center" refers to a facility that provides required computing power, data and algorithms for artificial intelligence applications (such as artificial intelligence deep learning model development, model training and model inference scenarios) by using large-scale heterogeneous computing power resources, including general computing power and intelligent computing power. An intelligent computing center covers facilities, hardware and software, and can provide full-stack capabilities from bottom computing power to top application enablement.

[0004] An "intelligent computing center" includes but is not limited to a "wisdom computing center".

[0005] A "wisdom computing center", also known as an artificial intelligence computing center, is a type of computing power infrastructure based on artificial intelligence theory and using artificial intelligence computing architecture to provide computing power services, data services and algorithm services required for artificial intelligence applications.

[0006] "Computing power" is the core of "intelligent computing centers" and "wisdom computing centers". It is the ability of computer equipment or computing / data centers to process information. It is the ability of computer hardware and software to jointly execute certain computing needs. It is the computing power that achieves target result output by processing information data. It is a new type of productivity that integrates information computing power, network carrying capacity and data storage power, and mainly provides services to society through computing power infrastructure.

[0007] Agents based on large language models or multi-modal large models have become the core components of the intelligent computing center service ecosystem. Intelligent computing centers provide strong computing power support and resource scheduling capabilities for agents. Current agent technology is evolving from single agent to multi-agent system to cope with complex task scenarios and distributed task processing needs. In existing multi-agent systems, although a multi-agent collaboration mechanism is introduced, a centralized collaboration mechanism is generally used to manage the behavior and resource allocation of all agents. This multi-agent system requires the topology of the agent to be determined at the system design stage and known to all participating agents. When the number of agents increases significantly and needs to adapt to complex tasks, the existing solution causes the multi-agent system to lack further dynamic expansion capability, easily forms a performance bottleneck and increases the risk of failure, limiting the flexibility and robustness of the multi-agent system. In addition, how to audit the security of the agents in the multi-agent system is also a problem to be solved. SUMMARY

[0008] The application provides an intelligent computing center cloud platform computing power driven intelligent agent network security auditing method and device, which is used to solve the problems of the lack of further dynamic expansion capability of the existing multi-agent system adopting the centralized cooperation mechanism, the easy formation of performance bottleneck and the increase of failure risk, the limitation of flexibility and robustness of the multi-agent system, and how to audit the security of the intelligent agent in the multi-agent system.

[0009] In order to solve the above technical problems, the application is implemented as follows:

[0010] In a first aspect, the application provides an intelligent computing center cloud platform computing power driven intelligent agent network security auditing method, comprising:

[0011] Step S1: a first intelligent agent sends broadcast information and receives broadcast information sent by other intelligent agents, wherein the broadcast information contains the capability information of the intelligent agent sending the broadcast information, the routing table maintained by the intelligent agent sending the broadcast information, and the intelligent agent capability table;

[0012] Step S2: the first intelligent agent updates the routing table and the intelligent agent capability table maintained by itself according to the received broadcast information;

[0013] Step S3: when the first intelligent agent receives a user task sent by an edge intelligent agent, the user task is decomposed into a plurality of subtasks, and a second intelligent agent executing each subtask is selected according to the intelligent agent capability table maintained by itself, and the subtask is distributed to the second intelligent agent according to the routing table maintained by itself;

[0014] Step S4: the first intelligent agent receives the subtask execution result returned by the second intelligent agent, performs security auditing on the subtask execution result, and discards or modifies the subtask execution result if the subtask execution result fails the security auditing;

[0015] Step S5: the first intelligent agent generates a task execution result based on the subtask execution result passing the security auditing, and sends the task execution result to the edge intelligent agent;

[0016] Wherein, the routing table contains a third intelligent agent that can be reached by the intelligent agent maintaining the routing table, and a forwarding path between the third intelligent agent and the intelligent agent, and the intelligent agent capability table contains the capability information of the third intelligent agent.

[0017] Optionally, the step S4 comprises:

[0018] Step S41: at least one of the following security audits is performed on the subtask execution result:

[0019] detecting whether the subtask execution result contains inappropriate content, and if the inappropriate content is detected in the subtask execution result, determining that the subtask execution result fails the security audit;

[0020] detecting whether the subtask execution result contains an unsafe prompt word, and if the unsafe prompt word is detected in the subtask execution result, determining that the subtask execution result fails the security audit;

[0021] detecting whether the subtask execution result matches the intention of the subtask, and if the subtask execution result does not match the intention of the subtask, determining that the subtask execution result fails the security audit.

[0022] Optionally, the step S3 comprises:

[0023] The step S3 comprises:

[0024] Step S31: For a target subtask in the plurality of subtasks, the first agent selects a plurality of second agents to execute the target subtask, and sends the target subtask to the plurality of second agents according to a routing table maintained by the first agent.

[0025] The step S4 comprises:

[0026] Step S41': For the target subtask, the first agent performs a security audit on the subtask execution result returned by each second agent executing the target subtask, and if the subtask execution result fails the security audit, discards or modifies the subtask execution result; if there are a plurality of subtask execution results passing the security audit for the target subtask, selects one subtask execution result from the plurality of subtask execution results passing the security audit as the subtask execution result of the target subtask.

[0027] Optionally, the step S4 comprises:

[0028] Step S41'': For a target subtask in the plurality of subtasks, the first agent performs a security audit on the subtask execution result returned by the second agent, and if the subtask execution result fails the security audit, discards the subtask execution result failing the security audit, and reselects a second agent for the target subtask according to an agent capability table maintained by the first agent, and distributes the target subtask to the reselected second agent according to a routing table maintained by the first agent.

[0029] Optionally, the step S3 comprises:

[0030] Step S31': when the first intelligent agent contains multiple candidate intelligent agents with the same or similar capabilities in the intelligent agent capability table maintained by the first intelligent agent, the first intelligent agent obtains a comprehensive value score of the multiple candidate intelligent agents, and selects a second intelligent agent for executing the subtask from the multiple candidate intelligent agents according to the comprehensive value scores of the multiple candidate intelligent agents;

[0031] The comprehensive value score is determined based on at least one of the following target indicators: a computing node performance indicator, a network performance indicator, a historical performance indicator, a load state indicator, a cost benefit indicator, and a security indicator, wherein the security indicator is determined according to whether the subtask execution result returned by the candidate intelligent agent in the past passes the security audit.

[0032] Optionally, the method further comprises:

[0033] Step S6: If it is detected that the subtask execution result returned by the second intelligent agent does not pass the security audit, the first intelligent agent reduces the score of the security indicator in the comprehensive value score of the second intelligent agent;

[0034] Step S7: When the score of the security indicator in the comprehensive value score of the second intelligent agent is lower than a preset threshold, the first intelligent agent deletes the second intelligent agent from the intelligent agent capability table maintained by the first intelligent agent or adds the second intelligent agent to a blacklist.

[0035] In a second aspect, the present application provides an intelligent computing center cloud platform computing power driven intelligent agent network security audit device, comprising:

[0036] A transceiving module is configured to send broadcast information and receive broadcast information sent by other intelligent agents, wherein the broadcast information contains capability information of an intelligent agent sending the broadcast information, a routing table maintained by the intelligent agent sending the broadcast information, and an intelligent agent capability table;

[0037] An updating module is configured to update the routing table and the intelligent agent capability table maintained by itself according to the received broadcast information;

[0038] A scheduling module is configured to, in the case of receiving a user task sent by an edge intelligent agent, decompose the user task into multiple subtasks, select a second intelligent agent for executing each of the subtasks according to the intelligent agent capability table maintained by itself, and distribute the subtasks to the second intelligent agent according to the routing table maintained by itself;

[0039] A security audit module is configured to receive a subtask execution result returned by the second intelligent agent, perform a security audit on the subtask execution result, and discard or modify the subtask execution result if the subtask execution result does not pass the security audit;

[0040] The feedback module is used to generate a task execution result based on the sub-task execution result that has passed the security audit, and send the task execution result to the edge agent;

[0041] The routing table contains the third agents that the agent maintaining the routing table can reach, as well as the forwarding paths between the agent and the third agent. The agent capability table contains the capability information of the third agent.

[0042] Thirdly, the present invention provides an electronic device, comprising: a processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, it implements the steps of the computing power-driven intelligent agent network security auditing method of the intelligent computing center cloud platform as described in the first aspect above.

[0043] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the computing power-driven intelligent agent network security auditing method of the intelligent computing center cloud platform as described in the first aspect above.

[0044] Fifthly, the present invention provides a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the computing power-driven intelligent agent network security auditing method of the intelligent computing center cloud platform as described in the first aspect above.

[0045] In this invention, a dynamic routing mechanism enables agents to discover and connect with each other, thereby dynamically constructing a multi-agent system with a decentralized network. Agents can join or leave the network at any time without needing to know the global topology, giving the multi-agent system dynamic scalability, reducing performance bottlenecks and failure risks, and significantly improving the system's flexibility and robustness. Furthermore, through a self-organizing collaborative mode among agents, complex tasks are decomposed and computational resources are allocated, enabling efficient processing of complex tasks. Moreover, by performing security audits on the sub-task execution results returned by agents within the multi-agent system, the security and user experience of the multi-agent system are improved. Attached Figure Description

[0046] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0047] Figure 1A flowchart of an intelligent computing center cloud platform computing power driven intelligent agent network security auditing method of the present application is shown in Figure 1.

[0048] Figure 2 A structural diagram of a multi-agent system of the present application is shown in Figure 2.

[0049] Figure 3 A structural diagram of an intelligent computing center cloud platform computing power driven intelligent agent network security auditing device of the present application is shown in Figure 3.

[0050] Figure 4 A structural diagram of an electronic device of the present application is shown in Figure 4. DETAILED DESCRIPTION

[0051] The technical solutions in the present application will be described clearly and completely below in conjunction with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0052] The technical terms involved in the present application will be described simply below.

[0053] The "computing power" referred to in the present application refers to the ability of a computer device or a computing / data center to process information, the ability of computer hardware and software to jointly perform certain computing requirements, the computing power to achieve target result output by processing information data, and a new type of productivity integrating information computing power, network carrying capacity, and data storage power, which mainly provides services to society through computing power infrastructure.

[0054] The "computing power" (CP) referred to in the present application refers to the ability of a data center server to process data and achieve result output, which is a comprehensive index for measuring the computing power of a data center, including general computing power, supercomputing power, and intelligent computing power. The commonly used unit of measurement is the number of floating point operations per second (FLOPS, 1 EFLOPS = 10^18 FLOPS), and the larger the value, the stronger the comprehensive computing power. According to estimates, 1 EFLOPS is about the computing power output of 5 Tianhe 2A, 500,000 mainstream server CPUs, or 2 million mainstream notebook computers. The calculation formula is: CP = CP 通用 + CP 智能 + CP 超级 .

[0055] The "network power" (NP) of the present application refers to the performance of the computing power facility data transmission capability, including the comprehensive capability of network architecture, network bandwidth, transmission delay, intelligent management and scheduling, etc., involving the network transmission inside and between data centers, and is a comprehensive index for measuring the network transmission scheduling capability.

[0056] The "storage power" (SP) of the present application refers to the comprehensive capability of the data center in four aspects of data storage capacity, performance, safety and reliability, and green low carbon, which is a comprehensive index for measuring the data storage capability of the data center, including external storage devices such as storage arrays and built-in storage devices of servers. The commonly used unit of storage capacity is exabyte (EB, 1EB=2^60bytes), the commonly used unit of performance is the number of read-write times per second per unit capacity (IOPS / TB, Input / Output Operations Per Second / TB), and the disaster recovery ratio is an important performance of safety and reliability.

[0057] The "computing power infrastructure" of the present application refers to a new type of information infrastructure integrating information computing power, network carrying capacity and data storage power, which can realize centralized computing, storage, transmission and application of information.

[0058] The "new type of information infrastructure" of the present application refers to network infrastructure mainly including 5G network, fiber broadband network, backbone network, international communication network, satellite internet, etc., computing power infrastructure including data center, general computing power center, intelligent computing center, supercomputing center, etc., and new technology facilities including artificial intelligence, blockchain, quantum computing, etc.

[0059] The "computing power" of the present application includes general computing power, intelligent computing power and supercomputing power.

[0060] The "general computing power" of the present application refers to the computing capability provided by the server based on CPU (Central Processing Unit) chip, which is used to support basic general computing such as cloud computing and edge computing.

[0061] The "intelligent computing power" of the present application refers to the large-scale deployment of computing platforms based on GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), ASIC (Application Specific Integrated Circuit) and other special chips for various artificial intelligence innovative applications, such as natural language processing, machine vision, etc.

[0062] The "super computing power" described in the application refers to: the computing power provided by a supercomputer and other high-performance computing clusters, which utilizes the centralized computing resources of multiple computer systems working in parallel and processes extremely complex or data-intensive problems through a dedicated operating system, mainly for computing in cutting-edge scientific fields, such as planet simulation, drug molecule design, gene analysis, etc.

[0063] The "intelligent computing center" described in the application refers to: a facility that provides the required computing power, data and algorithms for artificial intelligence applications (such as artificial intelligence deep learning model development, model training and model inference scenarios) by using large-scale heterogeneous computing resources, including general-purpose computing power (CPU) and intelligent computing power (GPU, FPGA, ASIC, etc.). The intelligent computing center covers facilities, hardware and software, and can provide full-stack capabilities from bottom computing power to top application enablement.

[0064] The "intelligent computing center cloud platform" described in the application refers to: a cloud computing platform based on the comprehensive services of hardware resources and software resources of the intelligent computing center.

[0065] The "intelligent computing center" described in the application includes but is not limited to "intelligent computing center".

[0066] The "intelligent computing center" described in the application, i.e. artificial intelligence computing center, is based on artificial intelligence theory, adopts artificial intelligence computing architecture, and provides computing power services, data services and algorithm services required by artificial intelligence applications.

[0067] The "computing power center" described in the application refers to: a facility with computing power, carrying capacity and storage capacity, mainly composed of infrastructure such as wind, fire, water and electricity, and IT hardware and software devices, including general-purpose data centers, intelligent computing centers, supercomputing centers, etc.

[0068] The "supercomputing center" described in the application refers to: a supercomputing data center, which is based on a supercomputer or a large-scale computing cluster, and can provide large-scale computing, storage and network services, etc. It is widely used in aerospace, national defense, oil exploration, climate modeling and genome sequencing, etc.

[0069] The "computing power resource" described in the application refers to: the technology and facilities required for the development of digital society, including but not limited to computing resources such as CPU and GPU, network resources such as switches and routers, storage resources such as storage arrays and distributed storage, security resources such as firewalls and intrusion detection systems, and supporting resources such as wind, fire, water and electricity.

[0070] The "computing power node" described in the application refers to: the computing resources of servers / containers that can handle computing tasks.

[0071] The "large language model" described in the present application refers to a large language model (LLM), which is a language model with a large number of parameters, aiming to understand and generate human language, and is trained through a large amount of text data, and can perform a wide range of tasks including text summarization, translation, sentiment analysis, etc.

[0072] The "multimodal large model" described in the present application refers to a model that jointly trains text, image, video, audio and other multimodal information, including but not limited to multimodal large language models.

[0073] The "agent" described in the present application refers to an agent that can perceive the environment and take actions to achieve specific goals, and is an important part of the intelligent computing center cloud platform. It can be software, hardware or a system, with autonomy, adaptability and interaction ability. The agent perceives changes in the environment (such as through sensors or data input), makes judgments and decisions based on its learned knowledge and algorithms, and then performs actions to influence the environment or achieve predetermined goals. Agents are widely used in the field of artificial intelligence, commonly found in automated systems, robots, virtual assistants and game characters, and their core is to autonomously learn and continuously evolve to better complete tasks and adapt to complex environments.

[0074] To solve the problem of the existing multi-agent system adopting a centralized collaboration mechanism, which lacks further dynamic expansion capability, is easy to form a performance bottleneck and increase the risk of failure, limits the flexibility and robustness of the multi-agent system, and how to audit the security of the agents in the multi-agent system, please refer to Figure 1 The present application provides a computing power driven agent network security auditing method of an intelligent computing center cloud platform, which comprises:

[0075] Step S1: The first agent sends broadcast information and receives broadcast information sent by other agents, wherein the broadcast information contains the capability information of the agent sending the broadcast information, the routing table maintained by the agent sending the broadcast information, and the agent capability table;

[0076] In the present application, optionally, the first agent comprises at least one of the following: an agent deployed inside the intelligent computing center; an agent deployed outside the intelligent computing center. That is, the first agent can be an agent deployed inside the intelligent computing center, or an agent deployed outside the intelligent center.

[0077] The first agent sends broadcast information to other agents, and the broadcast information sent by the first agent comprises: capability information of the first agent, a routing table maintained by the first agent, and an agent capability table. The routing table sent by the first agent comprises a third agent reachable by the first agent and a forwarding path between the first agent and the third agent, and the agent capability table sent by the first agent comprises capability information of the third agent.

[0078] In the application, the capability information of the agent is stored in a preset data table in the form of a structured description statement. The capability range of the agent covers two levels of basic function capability and business application capability. The basic function capability includes but is not limited to: document parsing processing capability, code generation and analysis capability, image recognition and processing capability, and multi-modal information processing capability. The business application capability includes but is not limited to: specialized processing capability of specific business scenarios such as life scenarios and financial data scenarios.

[0079] In the application, the third agent refers to a destination node in the routing table, and the forwarding path between the first agent and the third agent refers to a forwarding path of information sent by the first agent to the third agent.

[0080] Optionally, the first agent periodically sends the broadcast information. The sending period can be set as required.

[0081] The first agent can also receive broadcast information sent by other agents, and the broadcast information sent by the other agents comprises: capability information of the other agents, a routing table maintained by the other agents, and an agent capability table. The routing table sent by the other agents comprises a third agent reachable by the other agents and a forwarding path between the other agents and the third agent, and the agent capability table sent by the other agents comprises capability information of the third agent.

[0082] In the application, the third agent refers to a destination node in the routing table, and the forwarding path between the other agents and the third agent refers to a forwarding path of information sent by the other agents to the third agent.

[0083] In the application, the first agent receives broadcast information sent by adjacent other agents. It should be noted that the agents adjacent to each other can be connected by a wired mode or by a wireless mode.

[0084] Optionally, the other agents periodically send the broadcast information. The sending period can be set as required. Further optionally, the sending period of each agent (including the first agent and the other agents) can be the same.

[0085] Optionally, the other intelligent agents comprise at least one of the following:

[0086] The intelligent agent deployed in the first intelligent computing center, wherein the first intelligent computing center is the intelligent computing center to which the first intelligent agent belongs;

[0087] The intelligent agent deployed in the second intelligent computing center;

[0088] The intelligent agent deployed outside the intelligent computing center.

[0089] The routing table in the present application can refer to Table 1, which is the routing table maintained by the intelligent agent A.

[0090] Table 1

[0091]

[0092] The intelligent agent capability table in the present application can refer to Table 2, which is the intelligent agent capability table maintained by the intelligent agent A.

[0093] Table 2

[0094]

[0095] The intelligent agent A can be the first intelligent agent or the other intelligent agent.

[0096] In the present application, as shown in Table 3, the routing table and the intelligent agent capability table can also be integrated into one table.

[0097] Table 3

[0098]

[0099] It should be noted that the format of the above routing table and intelligent agent capability table is only an example, and other formats can also be used.

[0100] Step S2: The first intelligent agent updates the routing table and the intelligent agent capability table maintained by itself according to the received broadcast information;

[0101] In the present application, after the first agent receives the broadcast information sent by other agents, it can query whether the identifier of the other agent is contained in the routing table and the agent capability table maintained by itself. If not, the identifier of the other agent needs to be added to the routing table and the agent capability table, and at the same time, the forwarding path between the first agent and the other agent is generated in the routing table, and the capability information of the other agent is recorded in the agent capability table. In addition, it is also necessary to determine the agent that can be reached by the first agent through the relay of the other agent according to the routing table of all the other agents, and the forwarding path between the first agent and the agent that can be reached through the relay of the other agent, and update it to the routing table maintained by the first agent. At the same time, the capability information of the agent that can be reached through the relay of the other agent is updated to the agent capability table maintained by the first agent.

[0102] If the identifier of the other agent is contained in the routing table and the agent capability table maintained by itself, the first agent also needs to update the forwarding path between the other agent and the agent that can be reached through the relay of the other agent in the routing table maintained by itself, and at the same time, it is also necessary to update the capability information of the other agent and the agent that can be reached through the relay of the other agent recorded in the agent capability table maintained by itself.

[0103] So far, through the above dynamic routing mechanism, the mutual discovery and connection between agents can be completed, thereby dynamically constructing a multi-agent system with a decentralized network.

[0104] Please refer to Figure 2 , Figure 2 The structure diagram of the multi-agent system of the present application is shown in Figure 2 In the present application, each agent (A, B, C,...) maintains its own routing table and agent capability table, and periodically sends broadcast information. The agent receiving the broadcast information needs to update the routing table and the agent capability table maintained by itself based on the received broadcast information to complete the mutual discovery and connection between agents. The first agent can be any agent in Figure 2 .

[0105] Figure 2 In the present application, the intelligent computing center 1 and the intelligent computing center 2 each include a plurality of internal agents, and can also include aggregatable external agents. In the present application, the "aggregatable external agent" refers to an agent deployed outside the intelligent computing center and capable of cooperating with the agents inside the intelligent computing center to complete tasks.

[0106] The above intelligent agent includes three main modules, one of which is a routing module for sending and receiving broadcast information and maintaining a routing table; a capability information module for determining self-capability information and maintaining an intelligent agent capability table; and a task execution module for executing tasks. In addition, some intelligent agents can also have a task decomposition module for executing task decomposition. Some intelligent agents can also have a security audit module for performing security audits.

[0107] Step S3: In the case where the first intelligent agent receives a user task sent by the edge intelligent agent, the user task is decomposed into multiple sub-tasks, and according to the intelligent agent capability table maintained by itself, a second intelligent agent for executing each sub-task is selected, and according to the routing table maintained by itself, the sub-tasks are distributed to the second intelligent agent;

[0108] It should be noted that the first intelligent agent in the present application is an intelligent agent with task decomposition and computing resource allocation capabilities, which can be any intelligent agent in the above-mentioned decentralized network. Preferably, based on security and computing power considerations, it can be an intelligent agent inside the intelligent computing center. Further preferably, it is an intelligent agent inside the intelligent computing center that serves as a connection with external intelligent agents, such as intelligent agent A and intelligent agent F in Figure 2 In the present application, optionally, an intelligent agent for task decomposition and computing resource allocation can be provided for each intelligent computing center.

[0109] In the present application, the edge intelligent agent can be deployed at the network edge or at the user side, responsible for the access and preliminary processing of user tasks.

[0110] In an optional way, when the edge intelligent agent receives a user task sent by the user, it can directly forward the user task to the first intelligent agent. In this case, the edge intelligent agent can select the first intelligent agent in various ways, for example, it can select the intelligent agent inside the intelligent computing center closest to it as the first intelligent agent, or it can select the intelligent agent inside the intelligent computing center with the lowest cost as the first intelligent agent, etc.

[0111] In another optional way, when the edge intelligent agent receives a user task sent by the user, it can preprocess the user task, such as evaluating the required capabilities of the user task, so as to select a first intelligent agent that is adapted to the capabilities of the user task, or select an intelligent computing center that is adapted to the capabilities of the user task. Alternatively, the edge intelligent agent can also perform preliminary decomposition of the user task.

[0112] In the present application, the edge intelligent agent needs to query the routing table maintained by itself, and according to the forwarding path between the selected first intelligent agent queried from the routing table, the user task is sent to the first intelligent agent.

[0113] In the present application, the first agent can decompose the user task into a plurality of sub-tasks which can be executed independently, and the plurality of sub-tasks can include a plurality of sub-tasks which can be executed simultaneously, or sub-tasks which need to wait for the execution of other sub-tasks. The first agent determines the agent capability required for each sub-task, and selects a second agent for executing each sub-task from the maintained agent capability table. Each second agent can execute one sub-task, or a plurality of sub-tasks. Alternatively, one sub-task can also be executed simultaneously by a plurality of second agents. The first agent distributes the sub-tasks to the corresponding second agents according to the maintained routing table.

[0114] Step S4: The first agent receives the sub-task execution result returned by the second agent, performs security audit on the sub-task execution result, and discards or modifies the sub-task execution result if the sub-task execution result fails the security audit.

[0115] Step S5: The first agent generates a task execution result based on the sub-task execution result that passes the security audit, and sends the task execution result to the edge agent.

[0116] In the present application, if the sub-task execution result is modified, it can be considered that the modified sub-task execution result passes the security audit, or the sub-task execution result is modified and then the security audit is performed again on the sub-task execution result.

[0117] The first agent can combine the sub-task execution results returned by a plurality of second agents to obtain the task execution result, or analyze the sub-task execution results returned by a plurality of second agents to obtain the task execution result, according to the specific situation.

[0118] The first agent sends the task execution result to the edge agent according to the query routing table and the forwarding path between the first agent and the edge agent in the routing table.

[0119] The routing table (including the routing table of the above-mentioned first agent and the routing table of other agents) contains a third agent that can be reached by the agent that maintains the routing table, and the forwarding path between the third agent and the agent, and the agent capability table contains the capability information of the third agent.

[0120] In the present application, through the dynamic routing mechanism, the mutual discovery and connection between agents are completed, thereby dynamically constructing a multi-agent system with a decentralized network. The agents can join or exit the network at any time without predicting the global topology structure, so that the multi-agent system can have dynamic expansion capability, reduce performance bottlenecks and failure risks, and greatly improve the flexibility and robustness of the multi-agent system. Furthermore, through the self-organizing cooperation mode between agents, the decomposition of complex tasks and the allocation of computing resources are realized, thereby efficiently completing the processing of complex tasks. Further, by performing security audit on the sub-task execution results returned by the agents in the multi-agent system, the security and user experience of the multi-agent system are improved.

[0121] It should be noted that the agent in the present application is an agent based on a large language model or a multi-modal large model. The "based on a large language model or a multi-modal large model" means that the large language model or multi-modal large model is called to perform reasoning, thinking, design, dialogue with the user, generate scheme design and / or code generation, etc. tasks by organizing input information and designing prompt words.

[0122] In the present application, optionally, the step S4 comprises:

[0123] Step S41: performing at least one of the following security audits on the sub-task execution result:

[0124] (1) detecting whether the sub-task execution result contains inappropriate content, and if the sub-task execution result is detected to contain the inappropriate content, determining that the sub-task execution result fails the security audit;

[0125] The inappropriate content can include content that does not match the user's age, etc. For example, if it is identified that the user's task is to develop an English learning course for children under the age of 6, and the returned sub-task execution result contains game advertisement type content, it is considered that there is inappropriate content.

[0126] When the sub-task execution result contains inappropriate content, the sub-task execution result can be discarded directly, or the sub-task execution result can be modified, such as filtering out inappropriate content.

[0127] (2) detecting whether the sub-task execution result contains an unsafe prompt word, and if the sub-task execution result is detected to contain the unsafe prompt word, determining that the sub-task execution result fails the security audit;

[0128] This clause is mainly used to avoid the behavior of the second agent attacking the first agent by sending an unsafe prompt word to the first agent. For example, the second agent obtains the privacy data of the first agent by sending a prompt word, or modifies the task command of the first agent, or replaces the role of other agents, etc.

[0129] When the unsafe prompt word exists in the subtask execution result, the subtask execution result can be discarded directly.

[0130] (3) detecting whether the subtask execution result matches the intention of the subtask, and if it is detected that the subtask execution result does not match the intention of the subtask, determining that the subtask execution result fails to pass the security audit.

[0131] In the present application, optionally, the first agent can identify the intention of the subtask when the user task is decomposed into a plurality of subtasks, and after the second agent returns the subtask execution result, detect whether the subtask execution result matches the intention of the subtask. If it is detected that the subtask execution result does not match the intention of the subtask, it means that the subtask execution result deviates from the intention of the subtask to a high degree, and it is determined that the subtask execution result fails to pass the security audit, and the subtask execution result can be discarded directly.

[0132] In the present application, optionally, the step S3 comprises:

[0133] Step S31: For a target subtask in the plurality of subtasks, the first agent selects a plurality of second agents for executing the target subtask, and sends the target subtask to the plurality of second agents according to the routing table maintained by itself.

[0134] That is to say, for a subtask, a plurality of second agents can be selected for execution.

[0135] The target subtask can be part of the plurality of subtasks, and each subtask in the part of subtasks corresponds to a plurality of second agents, and the other subtasks can correspond to one second agent.

[0136] The target subtask can also be all subtasks, that is, each subtask in the plurality of subtasks corresponds to a plurality of second agents.

[0137] The step S4 comprises:

[0138] Step S41': For the target subtask, the first agent performs security audit on the subtask execution result returned by each second agent executing the target subtask, and if the subtask execution result fails to pass the security audit, discards or modifies the subtask execution result; if there are a plurality of subtask execution results passing the security audit for the target subtask, one subtask execution result is selected from the plurality of subtask execution results passing the security audit as the subtask execution result of the target subtask.

[0139] In the present application, one subtask execution result can be selected from a plurality of subtask execution results passing the security audit as the subtask execution result of the target subtask according to the matching degree of the subtask execution result and the intention. For example, the subtask execution result with the highest matching degree of intention of the target subtask is selected.

[0140] In the present application, the step S4 comprises:

[0141] Step S41'': for the target subtask in the plurality of subtasks, the first agent performs security audit on the subtask execution result returned by the second agent, if the subtask execution result does not pass the security audit, discarding the subtask execution result not passing the security audit, and reselecting a second agent for the target subtask according to the agent capability table maintained by itself, and distributing the target subtask to the reselected second agent according to the routing table maintained by itself.

[0142] It should be noted that the steps S41, S41' and S41'' in the present application are only used to illustrate different steps, and do not represent the order between the steps.

[0143] In the present application, the step S3 comprises:

[0144] Step S31'': when the agent capability table maintained by the first agent contains a plurality of candidate agents with the same or similar capabilities, obtaining a comprehensive value score of the plurality of candidate agents, and selecting a second agent for executing the subtask from the plurality of candidate agents according to the comprehensive value score of the plurality of candidate agents; wherein the comprehensive value score is determined based on at least one of the following target indicators: computing power node performance, network performance indicator, historical performance indicator, load state indicator, cost benefit indicator, and security indicator, wherein the security indicator is determined according to whether the subtask execution result returned by the candidate agent historically passes the security audit.

[0145] The computing power node performance can include at least one of the following: CPU performance, GPU performance, memory capacity, storage speed, and other hardware indicators.

[0146] The network performance indicator can include at least one of the following: access delay, bandwidth capacity, network stability, and other connection quality indicators.

[0147] The historical performance indicator can include at least one of the following: historical task completion success rate, historical task processing quality, and the like.

[0148] The load state indicator can include at least one of the following: load condition, available resource amount, and the like.

[0149] The cost-effectiveness indicator can include at least one of the following: use cost, cost performance of expected benefits, etc.

[0150] In the present application, optionally, the comprehensive value score is the weighted sum of at least two target indicators, i.e., the comprehensive value of the agent is evaluated from multiple dimensions.

[0151] In the present application, optionally, the agent with the highest comprehensive value score is selected as the second agent.

[0152] In the present application, optionally, the process of selecting the second agent can be performed for a certain subtask, for example, for a certain subtask, a plurality of candidate agents with the same or similar capabilities are queried from the agent capability table, at this time, the agent with the highest comprehensive value score is selected as the second agent.

[0153] The process of selecting the second agent can also be performed for multiple subtasks at the same time, for example, multiple subtasks need the same or similar agents to execute, and a plurality of candidate agents with the same or similar capabilities are queried from the agent capability table, the number of candidate agents is greater than the number of subtasks, at this time, the plurality of agents with the highest comprehensive value score can be selected from the queried candidate agents as the second agent for processing the plurality of subtasks.

[0154] In the present application, optionally, the step S3 further comprises:

[0155] Step S31'': the first agent collects the target indicators of each agent in the agent capability table maintained by the first agent, and calculates the comprehensive value score of each agent in the agent capability table maintained by the first agent according to the collected target indicators;

[0156] In the present application, optionally, the first agent can periodically collect the target indicators of each agent in the agent capability table maintained by the first agent, and calculate the comprehensive value score of each agent in the agent capability table maintained by the first agent, and record the comprehensive value score of each agent in the agent capability table maintained by the first agent, thereby facilitating the query.

[0157] The first agent can also collect the target indicators of each agent in the agent capability table maintained by the first agent in real time when the user task needs to be executed, and calculate the comprehensive value score of each agent in the agent capability table maintained by the first agent, thereby making the collected target indicators more real-time.

[0158] It should be noted that the steps S31, S31' and S31'' in the present application are only used to illustrate different steps, and do not represent the order between the steps.

[0159] Optionally, the method further comprises:

[0160] Step S6: If the subtask execution result returned by the second agent fails the security audit, the first agent reduces the score of the security index in the comprehensive value score of the second agent;

[0161] For example, the score of the security index in the comprehensive value score of the second agent is reduced by a preset score each time the subtask execution result returned by the second agent fails the security audit.

[0162] Step S7: When the score of the security index in the comprehensive value score of the second agent is lower than a preset threshold, the first agent deletes the second agent from the agent capability table maintained by the first agent or adds the second agent to a blacklist.

[0163] Adding the second agent to the blacklist means that the second agent will not be added to the agent capability table maintained by the first agent in the future. Deleting the second agent from the agent capability table maintained by the first agent means that the second agent can be added to the agent capability table maintained by the first agent again in the future.

[0164] Optionally, before the step S1, the method further comprises:

[0165] Step S0: The first agent determines capability information; wherein when the first agent is an agent in a first intelligent computing center for performing task decomposition and distribution, the capability information determined by the first agent comprises capability information of all agents aggregated by the first intelligent computing center.

[0166] For example, please refer to Figure 2 , it is assumed that the first agent is agent A, and the capability information determined by agent A comprises capability information of all agents aggregated by intelligent computing center 1, i.e. the capability information of intelligent computing center 1 as a whole, including the capability information of agents A, B, C and K. It is assumed that the first agent is agent F, and the capability information determined by agent F comprises capability information of all agents aggregated by intelligent computing center 2, i.e. the capability information of intelligent computing center 2 as a whole, including the capability information of agents F, G, H, I and J.

[0167] In addition to the agent for performing task decomposition and distribution, the capability determined by other agents is the capability of the agent itself, such as Figure 2 the capability information determined by agent B in

[0168] Please refer toFigure 3 The application also provides an intelligent computing center cloud platform computing power driven intelligent agent network security auditing device 10, comprising:

[0169] A transceiving module 11 is configured to send broadcast information and receive broadcast information sent by other intelligent agents, wherein the broadcast information comprises the capability information of the intelligent agent sending the broadcast information, a routing table maintained by the intelligent agent sending the broadcast information and an intelligent agent capability table;

[0170] An updating module 12 is configured to update the routing table and the intelligent agent capability table maintained by itself according to the received broadcast information;

[0171] A scheduling module 13 is configured to, in the case of receiving a user task sent by an edge intelligent agent, decompose the user task into a plurality of subtasks, select a second intelligent agent for executing each of the subtasks according to the intelligent agent capability table maintained by itself, and distribute the subtasks to the second intelligent agent according to the routing table maintained by itself;

[0172] A security auditing module 14 is configured to receive the subtask execution result returned by the second intelligent agent, perform security auditing on the subtask execution result, and discard or modify the subtask execution result if the subtask execution result fails to pass the security auditing;

[0173] A feedback module 15 is configured to generate a task execution result based on the subtask execution result that passes the security auditing, and send the task execution result to the edge intelligent agent;

[0174] The routing table comprises a third intelligent agent that can be reached by the intelligent agent maintaining the routing table and a forwarding path between the intelligent agent and the third intelligent agent, and the intelligent agent capability table comprises the capability information of the third intelligent agent.

[0175] In the application, the dynamic routing mechanism is used to complete the mutual discovery and connection between intelligent agents, thereby dynamically constructing a multi-intelligent agent system with a decentralized network, the intelligent agents can join or quit the network at any time without knowing the global topology structure in advance, so that the multi-intelligent agent system can have a dynamic expansion capability, reduce performance bottlenecks and failure risks, and greatly improve the flexibility and robustness of the multi-intelligent agent system. Furthermore, the self-organizing cooperation mode between the intelligent agents is used to realize the decomposition of complex tasks and the allocation of computing power resources, thereby efficiently completing the processing of complex tasks. Furthermore, the security auditing of the subtask execution result returned by the intelligent agent in the multi-intelligent agent system is performed, thereby improving the security and user experience of the multi-intelligent agent system.

[0176] Optionally, the first intelligent agent comprises at least one of the following:

[0177] The intelligent agent is deployed inside the intelligent computing center.

[0178] the agent deployed outside the intelligent computing center;

[0179] and / or, the other agent comprises at least one of:

[0180] the agent deployed in the first intelligent computing center, the first intelligent computing center being the intelligent computing center to which the first agent belongs;

[0181] the agent deployed in the second intelligent computing center;

[0182] the agent deployed outside the intelligent computing center.

[0183] Optionally, the security auditing module 14 is configured to perform at least one of the following security audits on the subtask execution result:

[0184] detecting whether the subtask execution result contains inappropriate content, and determining that the subtask execution result fails the security audit if the inappropriate content is detected in the subtask execution result;

[0185] detecting whether the subtask execution result contains an unsafe prompt word, and determining that the subtask execution result fails the security audit if the unsafe prompt word is detected in the subtask execution result;

[0186] detecting whether the subtask execution result matches the intention of the subtask, and determining that the subtask execution result fails the security audit if the subtask execution result does not match the intention of the subtask.

[0187] Optionally, the scheduling module 13 is configured to, for a target subtask in the plurality of subtasks, select a plurality of second agents that execute the target subtask, and send the target subtask to the plurality of second agents according to a routing table maintained by the scheduling module 13.

[0188] The security auditing module 14 is configured to, for the target subtask, perform a security audit on a subtask execution result returned by each second agent that executes the target subtask, discard or modify the subtask execution result if the subtask execution result fails the security audit, and select one subtask execution result from a plurality of subtask execution results that pass the security audit as a subtask execution result of the target subtask if there are a plurality of subtask execution results that pass the security audit for the target subtask.

[0189] Optionally, the security auditing module 14 is configured to perform security auditing on the subtask execution result returned by the second agent for a target subtask in the plurality of subtasks, discard the subtask execution result that fails the security auditing, and reselect a second agent for the target subtask according to the agent capability table maintained by the first agent, and distribute the target subtask to the reselected second agent according to the routing table maintained by the first agent.

[0190] Optionally, the scheduling module 13 comprises:

[0191] The selection sub-module is configured to, when the agent capability table maintained by the first agent contains a plurality of candidate agents with the same or similar capabilities, obtain a comprehensive value score of the plurality of candidate agents, and select a second agent for executing the subtask from the plurality of candidate agents according to the comprehensive value scores of the plurality of candidate agents; wherein the comprehensive value score is determined based on at least one of the following target indicators: a computing node performance indicator, a network performance indicator, a historical performance indicator, a load state indicator, a cost-benefit indicator, and a security indicator, wherein the security indicator is determined according to whether the subtask execution result returned by the candidate agent in the past passes the security auditing.

[0192] Optionally, the comprehensive value score is a weighted sum of at least two of the target indicators.

[0193] Optionally, the scheduling module 13 comprises:

[0194] The calculation sub-module is configured to collect the target indicators of each agent in the agent capability table maintained thereby, and calculate the comprehensive value scores of each agent in the agent capability table maintained thereby according to the collected target indicators.

[0195] Optionally, the apparatus further comprises:

[0196] The comprehensive value score updating module is configured to, if it is detected that the subtask execution result returned by the second agent fails the security auditing, reduce the score of the security indicator in the comprehensive value score of the second agent.

[0197] The processing module is configured to, when the score of the security indicator in the comprehensive value score of the second agent is lower than a preset threshold, delete the second agent from the agent capability table maintained by the first agent or add the second agent to a blacklist.

[0198] Optionally, the apparatus further comprises:

[0199] The determining module is configured to determine the capability information; when the first intelligent agent is an intelligent agent in a first intelligent computing center for performing task decomposition and distribution, the capability information determined by the first intelligent agent includes capability information of all intelligent agents that can be aggregated by the first intelligent computing center to which the first intelligent agent belongs.

[0200] Reference is made to Figure 4 The present application also provides an electronic device 20, comprising a processor 21, a memory 22, a computer program stored in the memory 22 and executable on the processor 21, which, when executed by the processor 21, implements each process of the above-mentioned computing center cloud platform's computing power-driven intelligent agent network security auditing method embodiment and achieves the same technical effects. To avoid repetition, details are not repeated here.

[0201] The present application also provides a computer readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements each process of the above-mentioned computing center cloud platform's computing power-driven intelligent agent network security auditing method embodiment and achieves the same technical effects. To avoid repetition, details are not repeated here. The computer readable storage medium includes a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0202] The present application also provides a computer program product, which includes computer instructions, and the computer instructions, when executed by a processor, implement each process of the above-mentioned computing center cloud platform's computing power-driven intelligent agent network security auditing method embodiment and achieve the same technical effects. To avoid repetition, details are not repeated here. Figure 1 The present application also provides a computer program product, which includes computer instructions, and the computer instructions, when executed by a processor, implement each process of the above-mentioned computing center cloud platform's computing power-driven intelligent agent network security auditing method embodiment and achieve the same technical effects. To avoid repetition, details are not repeated here.

[0203] It should be noted that in this document, the terms "comprise", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or device. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of additional identical elements in the process, method, article or device that includes the element.

[0204] Those skilled in the art can clearly understand the above-mentioned embodiment method can be realized by means of software and necessary general hardware platform, of course, also can be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application essentially or say the part of contribution to the prior art can be embodied in the form of software product, the computer software product is stored in a storage medium (such as ROM / RAM, disk, optical disk), including a plurality of instructions to make a terminal (may be a mobile phone, computer, server, air conditioner, or network equipment, etc.) executes the method described in various embodiments of the present application.

[0205] The embodiments of the present application are described above in conjunction with the drawings, but the present application is not limited to the above-mentioned specific embodiments, the above-mentioned specific embodiments are only illustrative, but not restrictive, those skilled in the art can make many forms under the inspiration of the present application without departing from the purpose of the present application and the scope protected by the claims, which all belong to the protection of the present application.

Claims

1. A computing power driven intelligent agent network security auditing method of an intelligent computing center cloud platform, characterized in that, Comprise: Step S1: the first intelligent agent sends broadcast information, and receives the broadcast information sent by other intelligent agents, wherein the broadcast information contains the capability information of the intelligent agent sending the broadcast information, the routing table and the intelligent agent capability table maintained by the intelligent agent sending the broadcast information; Step S2: the first intelligent agent updates the routing table and the intelligent agent capability table maintained by itself according to the received broadcast information; Step S3: in the case that the first intelligent agent receives a user task sent by an edge intelligent agent, the first intelligent agent decomposes the user task into multiple sub-tasks, selects a second intelligent agent for executing each sub-task according to the intelligent agent capability table maintained by itself, and distributes the sub-tasks to the second intelligent agent according to the routing table maintained by itself; Step S4: the first intelligent agent receives the sub-task execution result returned by the second intelligent agent, performs security audit on the sub-task execution result, and discards or modifies the sub-task execution result if the sub-task execution result fails the security audit; Step S5: the first intelligent agent generates a task execution result based on the sub-task execution result that passes the security audit, and sends the task execution result to the edge intelligent agent; Wherein, the routing table contains a third intelligent agent that the intelligent agent maintaining the routing table can reach and the forwarding path between the third intelligent agent and the intelligent agent, and the intelligent agent capability table contains the capability information of the third intelligent agent; The first intelligent agent comprises at least one of the following: an intelligent agent deployed inside an intelligent computing center; an intelligent agent deployed outside an intelligent computing center; The other intelligent agents comprise at least one of the following: An intelligent agent deployed in a first intelligent computing center, wherein the first intelligent computing center is the intelligent computing center to which the first intelligent agent belongs; An intelligent agent deployed in a second intelligent computing center; An intelligent agent deployed outside an intelligent computing center; Wherein, the edge intelligent agent selects the first intelligent agent by one of the following ways: selecting the intelligent agent inside the intelligent computing center closest to it as the first intelligent agent, selecting the intelligent agent inside the intelligent computing center with the lowest cost as the first intelligent agent; selecting the first intelligent agent that is suitable for the user task; selecting the intelligent agent in the intelligent computing center that is suitable for the user task as the first intelligent agent; Wherein, the step S3 comprises: Step S31': when the intelligent agent capability table maintained by the first intelligent agent contains multiple candidate intelligent agents with the same or similar capabilities, the first intelligent agent obtains the comprehensive value scores of the multiple candidate intelligent agents, and selects a second intelligent agent for executing the sub-task from the multiple candidate intelligent agents according to the comprehensive value scores of the multiple candidate intelligent agents; Wherein, the comprehensive value score is determined based on at least one of the following target indicators: computing node performance indicator, network performance indicator, historical performance indicator, load state indicator, cost benefit indicator, and security indicator, wherein the security indicator is determined according to whether the sub-task execution result returned by the candidate intelligent agent in the past passes the security audit.

2. The method of claim 1, wherein, The step S4 comprises: Step S41: performing at least one of the following security audits on the subtask execution result: detecting whether the subtask execution result contains inappropriate content, and determining that the subtask execution result fails the security audit if the inappropriate content is detected in the subtask execution result; detecting whether the subtask execution result contains an unsafe prompt word, and determining that the subtask execution result fails the security audit if the unsafe prompt word is detected in the subtask execution result; detecting whether the subtask execution result matches the intention of the subtask, and determining that the subtask execution result fails the security audit if the subtask execution result does not match the intention of the subtask.

3. The method of claim 1, wherein: the step S3 comprises: Step S31: for a target subtask in the plurality of subtasks, the first agent selects a plurality of second agents to execute the target subtask, and sends the target subtask to the plurality of second agents according to the routing table maintained by the first agent; the step S4 comprises: Step S41': for the target subtask, the first agent performs a security audit on the subtask execution result returned by each second agent that executes the target subtask, discards or modifies the subtask execution result if the subtask execution result fails the security audit, and selects one subtask execution result from the plurality of subtask execution results that pass the security audit as the subtask execution result of the target subtask if there are a plurality of subtask execution results that pass the security audit for the target subtask.

4. The method of claim 1, wherein, the step S4 comprises: Step S41'': for a target subtask in the plurality of subtasks, the first agent performs a security audit on the subtask execution result returned by the second agent, discards the subtask execution result that fails the security audit, and reselects a second agent for the target subtask according to the agent capability table maintained by the first agent, and distributes the target subtask to the reselected second agent according to the routing table maintained by the first agent.

5. The method of claim 1, wherein, Further comprising: Step S6: if the subtask execution result returned by the second agent is detected to fail the security audit, the first agent reduces the score of the security indicator in the comprehensive value score of the second agent; Step S7: when the score of the security indicator in the comprehensive value score of the second agent is lower than a preset threshold, the first agent deletes the second agent from the agent capability table maintained by the first agent or adds the second agent to a blacklist.

6. A computing power driven intelligent agent network security auditing device of an intelligent computing center cloud platform, characterized in that, The device is applied to a first agent, and the device comprises: a transceiver module configured to send broadcast information and receive broadcast information sent by other agents, wherein the broadcast information contains capability information of an agent sending the broadcast information, a routing table maintained by the agent sending the broadcast information, and an agent capability table; an updating module configured to update the routing table and the agent capability table maintained by the first agent according to the received broadcast information; and an updating module configured to update the routing table and the agent capability table maintained by the first agent according to the received broadcast information. The scheduling module is configured to, in the case of receiving a user task sent by an edge intelligent agent, decompose the user task into a plurality of subtasks, select a second intelligent agent for executing each of the subtasks according to an intelligent agent capability table maintained by the scheduling module, and distribute the subtasks to the second intelligent agents according to a routing table maintained by the scheduling module; The security auditing module is configured to receive a subtask execution result returned by the second intelligent agent, perform security auditing on the subtask execution result, and discard or modify the subtask execution result if the subtask execution result fails to pass the security auditing. The feedback module is configured to generate a task execution result based on the subtask execution result that passes the security auditing, and send the task execution result to the edge intelligent agent. The routing table contains third intelligent agents reachable by the intelligent agent that maintains the routing table and forwarding paths between the third intelligent agents, and the intelligent agent capability table contains capability information of the third intelligent agents. The first intelligent agent includes at least one of the following: An intelligent agent deployed inside an intelligent computing center; An intelligent agent deployed outside an intelligent computing center. The other intelligent agents include at least one of the following: An intelligent agent deployed in a first intelligent computing center to which the first intelligent agent belongs; An intelligent agent deployed in a second intelligent computing center; An intelligent agent deployed outside an intelligent computing center. The edge intelligent agent selects the first intelligent agent in one of the following ways:

7. An electronic device, comprising: Selecting an intelligent agent inside the nearest intelligent computing center as the first intelligent agent, or selecting an intelligent agent inside the intelligent computing center with the lowest cost as the first intelligent agent; Selecting a first intelligent agent that is capable of executing the user task; 8. A computer-readable storage medium, characterized in that, The scheduling module includes: A selection submodule configured to, in the case that the intelligent agent capability table maintained by the first intelligent agent contains a plurality of candidate intelligent agents with the same or similar capabilities, obtain a comprehensive value score of the plurality of candidate intelligent agents, and select a second intelligent agent for executing the subtask from the plurality of candidate intelligent agents according to the comprehensive value scores of the plurality of candidate intelligent agents; wherein the comprehensive value score is determined based on at least one of the following target indicators: a computing power node performance indicator, a network performance indicator, a historical performance indicator, a load state indicator, a cost-benefit indicator, and a security indicator, wherein the security indicator is determined according to whether a subtask execution result historically returned by the candidate intelligent agent passes security auditing. A processor, a memory, and a program stored on the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the intelligent computing center cloud platform computing power-driven intelligent agent network security auditing method according to any one of claims 1 to 5. The computer program is stored on the computer-readable storage medium and, when executed by the processor, implements the steps of the intelligent computing center cloud platform computing power-driven intelligent agent network security auditing method according to any one of claims 1 to 5.

9. A computer program product, characterised in that, Computer program product including computer instructions, which, when executed by a processor, implement the steps of the method for auditing the security of a network of intelligent agents driven by computing power of an intelligent computing center cloud platform according to any one of claims 1 to 5.

Images