Metaapplication agent service system based on AI technology micro service

Abstract

The invention provides a meta-application agent service system, which is used for micro-service packaging, publishing, containerization operation and maintenance, meta-application combination, business verification, credible traceability and application integration of an AI technology algorithm, and realizes cooperative work of an AI technology publishing party, an application integration party and a system platform operation and maintenance party through the service system. The platform end supports a technology supplier to publish AI technology algorithm data into micro-services with specific functions through technical model description specifications, application technology multilayer classification and visual interactive publishing service; the meta-application construction module supports an application integration party to upload application demand data and select micro-services to be combined for meta-application simulation evaluation, meta-applications with a large model agent as a carrier are generated, and the meta-applications can interact with the micro-services, the meta-applications, an application system, a cloud platform and an operating system; the operation and maintenance module comprises a containerized micro-service management tool and supports an operation and maintenance management party to distribute containers to micro-services and maintain operation of the containers.

Description

Technical Field

[0001] The present invention relates to the field of AI technology, and in particular to a meta-application intelligent agent service system based on AI technology microservices. Background Art

[0002] In the context of AI empowering industries to carry out business innovation, application innovation and technological transformation, all walks of life need AI technology that can meet their targeted needs. As the application end of AI technology, the industry usually does not independently develop related AI technology, so it needs to obtain technical resources through technology transfer, industry-university-research cooperation, etc. On the supply side of AI technology, there is also a strong motivation to promote the transformation of developed AI technology, verify the effectiveness of the technology, and obtain revenue and R&D support.

[0003] The AI ​​technology resources currently on the market include mainstream cloud platforms such as Amazon and Alibaba Cloud, as well as some technology cloud platforms for specific industries, such as the Star Cloud Open Platform released by China Science Star Map on November 24, 2024. Although they have begun to be supplied to specific fields and provide field solutions, given that AI technology suppliers and application parties are in different fields, technology demanders have limited understanding of professional technology, and technology suppliers have limited understanding of business needs, there are still three major deficiencies and limitations, including:

[0004] (1) AI technology design and application are easily disconnected, resulting in both supply and demand sides needing to set up R&D projects or outsource AI technology realization and implementation, which is inefficient and costly.

[0005] (2) It is difficult for the application party to evaluate and select AI technology, which requires a lot of communication and consultation. Even for open technology cloud platforms that are oriented to specific fields, such as the China Science Star Map Open Cloud Platform, professional technical teams are still needed to implement their scenario solutions.

[0006] (3) The applicability of AI technology components is not strong, which is not conducive to the technology promotion of technology suppliers; the integration of AI technology components lacks flexibility, making integration difficult and costly for technology users.

[0007] It can be seen that in order to empower the industrial field with AI technology, there is an urgent need to innovate full-process facilitation services for technical collaboration between AI technology suppliers and application integrators, and to enhance AI technology supply resources, ensure technology quality, and improve application effectiveness and integration efficiency through the forefront of AI technology. Summary of the invention

[0008] The present invention is made to solve the above-mentioned problems, and its purpose is to provide a meta-application intelligent agent service system based on AI technology microservices.

[0009] The present invention provides a meta-application intelligent agent service system based on AI technology microservices, which is used for microservice encapsulation, publishing, containerized operation and maintenance, meta-application combination, business verification, trusted traceability, and application integration of AI technology algorithm data. The system has the following characteristics: a platform end and multiple user ends held by different users, wherein each user end is respectively connected to the platform end in communication, and the users include a technology supplier, an application integrator, and an operation and maintenance manager. The platform end includes: a microservice module, which publishes services through technical model description specifications, multi-layer classification of application technologies, and visual interaction, and uploads AI data from the technology supplier through the user end. I technology algorithm data publishes microservices with specific functions; the meta-application construction module, which includes a large language model based on knowledge enhancement, generates multiple microservice combinations and corresponding visualization maps according to the application demand data uploaded by the application integrator through the user end, and performs meta-application simulation evaluation according to the microservices to be combined selected by the application integrator, and publishes the meta-application integrated with the microservices selected by the application integrator and with the large model intelligent body as the software carrier; the operation and maintenance module, which includes a containerized microservice management tool, allocates containers to microservices and maintains the operation of containers according to the operation and maintenance instruction data uploaded by the operation and maintenance manager through the user end.

[0010] The meta-application intelligent agent service system based on AI technology microservices provided by the present invention may also have the following features: when the meta-application is running, the meta-application intelligent agent generates scheduling instructions according to the user's demand instructions, calls the various microservices corresponding to the meta-application to run according to the scheduling instructions, transfers data between microservices, and interacts with microservices, meta-applications, application systems, cloud platforms, and operating systems.

[0011] In the meta-application agent service system based on AI technology microservices provided by the present invention, it can also have the following features: wherein, the microservice module includes: a technical model description specification, an application technology multi-layer classification unit, which is used by the technology supplier to set the type and technical performance indicators corresponding to the AI ​​technology algorithm data, a visualization unit, which includes a microservice publishing agent, which is used to parse each functional function in the AI ​​technology algorithm data and generate a corresponding visualized microservice map; a microservice draft generation unit, which includes a large language model based on knowledge enhancement and preset microservice templates of various types, which is used to generate a corresponding program draft for the functional functions to be released selected by the technology supplier from the visualized microservice map; a customization unit, which includes a preset microservice description specification, which is used by the technology supplier to customize the program draft according to the microservice description specification, debug and publish the microservice, and the types include field subtypes, industry subtypes, scenario subtypes, technology subtypes, algorithm subtypes and function subtypes, and the microservice description specification is used to set the description interface of the microservice.

[0012] In the meta-application agent service system based on AI technology microservice provided by the present invention, it can also have the following characteristics: wherein, the platform end also includes a meta-application retrieval and evaluation module and a traceability module, the meta-application retrieval and evaluation module and / or the client includes a meta-application recommendation unit, the meta-application recommendation unit includes a meta-application retrieval agent, which is used to generate at least one recommended meta-application according to the application requirement data, and the meta-application retrieval and evaluation module includes: a meta-application recommendation unit, which includes a large language model based on knowledge enhancement, which is used to generate at least one recommended meta-application according to the application requirement data; a business verification unit, which stores a meta-application test data set, which is used to perform business verification on the recommended meta-application selected by the application integrator according to the business data in the application requirement data and the meta-application test data set, and generate a business verification evaluation result, and the traceability module uses blockchain to store the business verification evaluation result of the recommended meta-application by the application integrator.

[0013] The meta-application intelligent agent service system based on AI technology microservices provided by the present invention may also have the following features: wherein, the platform end also includes a microservice evaluation module and a traceability module, the microservice evaluation module stores a microservice test data set, which is used by the user to evaluate the microservice selected by the user according to the uploaded own data set and the microservice test data set, and generate test results, the traceability module uses blockchain to store the user's evaluation results of the microservice, and the microservice evaluation module uses the professional AI technology evaluation program released by the platform end or the stored large language model to evaluate the microservice.

[0014] In the meta-application intelligent agent service system based on AI technology microservices provided by the present invention, it can also have the following features: wherein, the operation and maintenance module visually displays the status of the microservice container and meta-application to the operation and maintenance management party, and the operation and maintenance module also maintains the operation of the meta-application based on the operation and maintenance instruction data.

[0015] In the meta-application intelligent agent service system based on AI technology microservices provided by the present invention, it can also have the following features: wherein, the platform end also includes a traceability module, and the meta-application construction module includes: a combination recommendation unit, which includes a large language model based on knowledge enhancement, and generates multiple microservice combinations and corresponding visualization maps according to application demand data; a simulation verification unit, which stores a preset simulation verification data set, and simulates and verifies the microservice combination selected by the application integrator according to the simulation verification data set and the business data in the application demand data, and generates a simulation verification evaluation result; the traceability module is used to use blockchain to store the user's simulation verification evaluation results of the published meta-application.

[0016] The meta-application intelligent agent service system based on AI technology microservices provided by the present invention may also have the following features: wherein, the platform end also includes a traceability module, which is used to store the evidence materials uploaded by users for microservices or meta-applications, and the evidence materials include at least one of qualification certificates, intellectual property rights, and application cases.

[0017] Functions and Effects of the Invention

[0018] According to the meta-application intelligent agent service system based on AI technology microservices involved in the present invention, because the AI ​​technology algorithm data is split and customized to generate microservices through the microservice module, multiple microservices are collaboratively called to generate meta-applications according to user needs in combination with the large model intelligent agent through the meta-application construction module, and microservices and meta-applications are operated and managed through the operation and maintenance module. Therefore, the meta-application intelligent agent service system based on AI technology microservices of the present invention can generate corresponding meta-applications flexibly integrated by multiple microservices according to any task, solve the problems of weak applicability and composability of technical components in the current technical service system, low flexibility of technical combination integration, and little consideration of the technical level and evaluation requirements of the technology release and integration parties, and realize the effective coordination of technical resource release, application integration and operation management. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Schematic diagram of the working principle of the meta-application intelligent agent service system based on AI technology microservices in an embodiment of the present invention;

[0020] Figure 2 is a block diagram of a meta-application agent service system based on AI technology microservices in an embodiment of the present invention;

[0021] Figure 3 is a block diagram of a platform end in an embodiment of the present invention;

[0022] Figure 4 It is a schematic diagram of the principle of cross-border payment anomaly analysis in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the following embodiments and the accompanying drawings specifically illustrate the meta-application intelligent agent service system based on AI technology microservices of the present invention.

[0024] This embodiment provides a meta-application intelligent agent service system based on AI technology microservices. It focuses on improving the efficiency, benefits, convenience and accessibility of AI technology resource supply and application, and is oriented to the whole process of AI technology supply, application integration, and operation and maintenance management. It provides microservice packaging, publishing, containerized operation and maintenance, meta-application combination, business verification, trusted traceability, application integration, and full-process services until the application system, covering AI technology algorithms. It can be used to build a flexible supply enterprise technology middle platform, a technology transaction public service platform, and an industry-university-research collaborative technology research and development and application verification platform.

[0025] Figure 1 It is a schematic diagram of the working principle of the meta-application intelligent agent service system based on AI technology microservices in an embodiment of the present invention.

[0026] like Figure 1 As shown, the meta-application intelligent agent service system based on AI technology microservices in this embodiment provides meta-applications for specific application needs to designated objects through an application integration interaction interface.

[0027] The meta-application is constructed by calling the corresponding technical microservices through the system platform knowledge-enhanced big model. The meta-application agent is a big model agent that serves as a meta-application carrier. The big model agent is constructed by an arbitrary big language model based on the meta-application-related knowledge enhancement. The content of knowledge enhancement includes but is not limited to the user-defined meta-application description, the platform service knowledge of service traceability, and the application knowledge of meta-application traceability.

[0028] Microservices are packaged based on AI algorithms or software and tested using specified data sets. Data sets include but are not limited to platform technology test data sets, user uploaded data, and platform business data.

[0029] Specified objects include but are not limited to other meta-application agents, application platforms, cloud platforms, and operating systems.

[0030] The meta-application agent service system based on AI technology microservices includes a platform end and multiple user ends owned by different users. Each user end is connected to the platform end in communication. Users include technology suppliers, application integrators, and operation and maintenance managers.

[0031] Figure 2 It is a block diagram of the platform end in an embodiment of the present invention.

[0032] like Figure 2 As shown, the platform end 10 includes a microservice module 11, a microservice evaluation module 12, a meta-application construction module 13, a meta-application retrieval and evaluation module 14, an operation and maintenance module 15, a traceability module 16 and a general control module 17 for controlling the operation of the above modules.

[0033] The microservice module 11 publishes services through technical model description specifications, multi-layer classification of application technologies, and visual interaction, and publishes microservices with specific functions based on the AI ​​technology algorithm data uploaded by the technology supplier through the user end.

[0034] The microservice module 11 includes an application technology multi-layer classification unit 111 , a visualization unit 112 , a microservice draft generation unit 113 and a customization unit 114 .

[0035] The applied technology multi-layer classification unit 111 is used by the technology supplier to set the type and technical performance indicators corresponding to the AI ​​technology algorithm data.

[0036] Among them, the types include domain subtype, industry subtype, scenario subtype, technology subtype, algorithm subtype and function subtype, that is, microservices are classified according to the six levels of domain, industry, scenario, technology, algorithm and function, so as to select the corresponding microservice template.

[0037] The visualization unit 112 includes a microservice publishing agent, which is used to parse each function in the AI ​​technology algorithm data and generate a corresponding visualized microservice map. In this embodiment, the large language model interprets the AI ​​technology algorithm data, i.e., the algorithm code, to parse out each function. In this embodiment, the microservice publishing agent is a large model agent used to parse function functions and generate a visualized microservice map, which is also constructed through a corresponding large language model based on knowledge enhancement.

[0038] The microservice draft generation unit 113 includes a large language model based on knowledge enhancement and preset microservice templates of different types, which is used to generate the corresponding program draft for the function to be released selected by the technology supplier from the visual microservice map. In this embodiment, knowledge enhancement is to strengthen the large language model through microservice knowledge derived from algorithm knowledge, historical knowledge derived from historical data of domain microservice combinations, and microservice descriptions derived from users. In this embodiment, each microservice template corresponds to a different preset microservice test code, so that the large language model can test the program drafts corresponding to different microservice templates.

[0039] The customization unit 114 includes a preset microservice description specification, which is used by the technology supplier to customize the draft of the program according to the microservice description specification, debug and publish the microservice. Among them, the microservice description specification is used to set the description interface of the microservice. In this embodiment, the microservice is a microservice monomer component based on AI technology. The component granularity and interface design of the microservice are set according to the application combination requirements of a specific field and the operation requirements based on the container. The published microservice includes microservice software and interface description files, i.e. configuration files. The interface of the microservice includes "input parameters", "output parameters", "conditional parameters", "effect parameters", etc. In this embodiment, the technology supplier performs visual publishing and debugging through an interactive operation cycle until the requirements of the technology supplier are met, and then the final version of the microservice is published.

[0040] The microservice evaluation module 12 stores a microservice test data set, which is used for the user to evaluate the microservice selected by the user based on the uploaded own data set and the microservice test data set, and generate test results.

[0041] Among them, the microservice evaluation module 12 uses the professional AI technology evaluation program released by the platform end 20 or the stored large language model to evaluate the microservice.

[0042] In this embodiment, the technology supplier, application integrator and operation and maintenance manager can all use means such as personal workspace, AI technology microservice management, traceability, microservice lists at the test site, or search to discover and select microservices for evaluation.

[0043] Figure 3 It is a schematic diagram of the working principle of the meta-application building module in an embodiment of the present invention.

[0044] like Figure 3 As shown, the meta-application construction module 13 includes a large language model based on knowledge enhancement, namely RAG, and the application integrator or the operation and maintenance manager, namely the operation and maintenance administrator, uploads the application demand data through the user end, and defines the meta-application name, type, demand and prompt, etc., and generates multiple microservice combinations and corresponding visual maps through the large language model based on knowledge enhancement. Subsequently, the user selects the microservices to be combined and the simulation verification data set. The meta-application construction module 13 performs a meta-application simulation evaluation based on the microservices to be combined selected by the user through the simulation verification data set, thereby providing the execution results, technical indicators and suggestions as the simulation evaluation results. The user determines whether to further adjust the microservices to be combined and re-perform the simulation evaluation according to the simulation evaluation results until the preset standards are met. The user publishes the meta-application integrated with the microservices selected by the application integrator and with the large model intelligent body as the software carrier, namely the meta-application release, through the meta-application construction module 13. At the same time, the traceability module 16 stores the simulation evaluation results through the blockchain, namely the blockchain storage.

[0045] In this embodiment, the large language model of the meta-application construction module 13 is enhanced according to system knowledge, and is used to generate a large model agent according to application requirement data and microservices to be combined, and further construct a meta-application.

[0046] When the meta-application is running, the meta-application agent generates a scheduling instruction according to the user's demand instruction, calls each microservice corresponding to the meta-application to run according to the scheduling instruction, and transmits data between microservices. In this embodiment, the meta-application interacts with other meta-applications, application platforms, cloud platforms, and operating systems through the meta-application agent to realize meta-application calling and integration.

[0047] In this embodiment, the meta-application construction module 13 displays a meta-application construction interface to the user, which can assist the user in defining the name of the meta-application to be constructed, and select the application field, industry, scenario, and technical field of the meta-application in turn through the drop-down box, and customize the role, functional requirements, and technical requirements of the meta-application, such as recall rate, precision rate, computational efficiency, etc., as well as related AI algorithms, and select verification data as application requirement data.

[0048] The meta-application construction module 13 includes a combination recommendation unit 131 and a simulation verification unit 132 .

[0049] The combination recommendation unit 131 includes a large language model based on knowledge enhancement, and generates multiple microservice combinations and corresponding visualization graphs according to application demand data.

[0050] In this embodiment, the large language model of the combination recommendation unit 131 can filter microservices that meet the application demand data from the existing microservices of the meta-application intelligent agent service system 100 based on AI technology microservices or the microservices disclosed on the Internet, thereby generating multiple microservice combinations and corresponding visualization maps.

[0051] The simulation verification unit 132 stores a preset simulation verification data set, and performs simulation verification on the microservice combination selected by the application integrator according to the simulation verification data set and the business data in the application requirement data, and generates a simulation verification evaluation result.

[0052] In this embodiment, the large model agent performs data scheduling on each microservice in the microservice combination, thereby using the verification data to perform simulation verification on the microservice combination. The simulation verification evaluation results include execution results, technical indicators and suggestions.

[0053] The meta-application search and evaluation module 14 includes a meta-application recommendation unit 141 and a business verification unit 142 .

[0054] The meta-application recommendation unit 141 includes a meta-application retrieval agent, which is used to generate at least one recommended meta-application according to the application requirement data. In other embodiments, the meta-application retrieval evaluation module 14 does not include the meta-application recommendation unit 141, but the user end includes the meta-application recommendation unit 141, or both the meta-application retrieval evaluation module 14 and the user end include the meta-application recommendation unit 141.

[0055] In this embodiment, the user uses the list of AI technology meta-application agents in personal workspace, AI technology meta-application management, status monitoring, etc., or passes the application field, industry, scenario, technical field, functional requirements, technical requirements, and related AI algorithms of the required AI technology to the enhanced large model agent to obtain recommended meta-applications, namely recommended meta-applications.

[0056] The business verification unit 142 stores a meta-application test data set, and is used to perform business verification on the recommended meta-application selected by the application integrator according to the business data in the application requirement data and the meta-application test data set, and generate a business verification evaluation result.

[0057] The knowledge-enhanced large language model of the meta-application retrieval and evaluation module 14 in this embodiment is a domain-enhanced large language model agent. On the basis of the available mainstream large language model agent, based on the knowledge of fields such as AI algorithms, microservices, and microservice combinations, combined with the knowledge graph of comprehensive user needs, the RAG framework is used to enhance the domain retrieval accuracy and analysis capabilities of the large model, and a domain-enhanced large language model agent is obtained. The domain-enhanced large language model agent can achieve accurate matching for multiple AI microservice combinations and user application integration needs, and perform business verification. The knowledge enhancement content includes user-defined meta-application requirement descriptions, service knowledge and application knowledge obtained from platform service traceability, and any knowledge enhancement technology is used to enhance knowledge based on any selected large language model, so that the domain-enhanced large language model agent can more accurately understand the search goals of application integration users, hit and recommend meta-applications, and perform verification based on business data on this basis.

[0058] The operation and maintenance module 15 includes a containerized microservice management tool, which allocates containers to microservices and maintains the operation of containers and meta-applications according to the operation and maintenance instruction data uploaded by the operation and maintenance management party through the user terminal. The operation and maintenance module 15 visually displays the status of the microservice container and meta-application to the operation and maintenance management party.

[0059] In this embodiment, the operation and maintenance manager can allocate, activate, and shut down corresponding containers for microservices, and perform activation and shutdown operations on meta-applications.

[0060] The traceability module 16 uses blockchain to store the user's evaluation results of microservices, the user's simulation verification and evaluation results of the published meta-applications, the application integrator's business verification and evaluation results of the recommended meta-applications, and the user's uploaded evidence materials for microservices or meta-applications. Among them, the evidence materials include at least one of qualification certificates, intellectual property rights, and application cases.

[0061] In this embodiment, the traceability module 16 supports technology suppliers, application integrators and operation and maintenance managers to trace the entire life cycle of microservices and meta-applications published on the platform 10, including microservice publishing, testing, combination, integration, calling, operating status, container information, and various intellectual property and usage certification materials.

[0062] The traceability module 16 includes active evidence storage and automatic evidence storage functions. Active evidence storage is a list of personal workspaces, AI technology microservice / meta-application intelligent management, traceability, and testing places on the user side by the technology supplier, application integrator, and operation and maintenance manager, or by searching and finding and selecting the microservice or meta-application to be added for evidence storage, and then opening the blockchain evidence storage interface that appears, selecting the category to be added such as intellectual property, qualification certificate, application case, submitting relevant certification materials, and submitting evidence storage materials. Automatic evidence storage is to automatically update the traceability information of the evidence storage every time microservice technology testing, publishing, container management, calling, and meta-application construction, simulation verification, calling, and other operations are performed.

[0063] The master control module 17 stores a control program for controlling the operation of each module.

[0064] The following, in conjunction with the accompanying drawings, describes the usage of the meta-application intelligent service system 100 based on AI technology microservices in the cross-border payment anomaly monitoring application scenario.

[0065] Figure 4 It is a schematic diagram of the principle of cross-border payment anomaly analysis in an embodiment of the present invention.

[0066] like Figure 4As shown in the figure, external application technology suppliers, key technology research and development of the subject, and external cutting-edge technology suppliers act as technology suppliers, and provide technology to the AI ​​technology service full-cycle operation and maintenance and integrated management platform through the user end to generate microservices. The retrieval and integration agent generates multiple meta-applications through the microservices in the AI ​​technology service full-cycle operation and maintenance and integrated management platform. The anti-money laundering reporting subsystem, the cross-border trade abnormal transaction monitoring subsystem, the cross-border trade foreign exchange risk monitoring subsystem, and the cross-border trade anti-money laundering monitoring subsystem respectively call the retrieval and integration agent to run the corresponding meta-applications. Among them, the cross-border trade abnormal transaction monitoring subsystem is used to issue abnormal warnings for cross-border e-commerce, the cross-border trade foreign exchange risk monitoring subsystem is used to determine whether cross-border tuition / remittances are blocked, and the cross-border trade anti-money laundering monitoring subsystem is used to issue abnormal warnings for cross-border trade.

[0067] In the above cross-border payment anomaly analysis, the AI ​​technology service full-cycle operation and maintenance and integrated management platform and retrieval and integration agent are supported by the platform end 10. Therefore, through the meta-application agent service system 100 based on AI technology microservices, full use is made of cutting-edge technologies such as blockchain, microservices, containerization, knowledge graphs, enhanced large models, and large model agents to achieve AI technology microservice release, microservice containerization management, meta-application dynamic construction, and full-cycle technical service trusted traceability, solving the problems of inconsistent release standards and difficult management involved in cross-border payment anomaly identification technology resource sharing, complex and error-prone service deployment, high service orchestration costs, and low efficiency of technical resource sharing due to trust, providing full-cycle integrated support for technical collaboration and application construction, and improving efficiency, convenience, and credibility. Specific advantages include:

[0068] By using technical service interface specifications and visual definitions, the efficiency of microservice release is improved, supporting flexible integration, collaborative development and wide applicability.

[0069] Through systematic classification based on business scenarios, application construction and technical properties, as well as full-cycle status management, three-dimensional information of AI technical services is formed to achieve efficient retrieval and accurate matching.

[0070] Through RAG enhancement, that is, knowledge-enhanced large-model intelligent agents, and visual service orchestration, we can achieve an efficient combination of AI technologies that collaborate with human intelligence, and quickly build an integrated, efficient AI technology meta-application based on enhanced intelligent agents.

[0071] Through the RAG-based large-model agent, intelligent retrieval, evaluation and integration of AI technology meta-application agents are realized, providing convenience for monitoring or other application system selection.

[0072] Through single technology evaluation and meta-application intelligent agent simulation for real application scenarios, a comprehensive evaluation combining single algorithm technology evaluation and application verification can be achieved, thus achieving deep matching of AI technology selection.

[0073] Based on blockchain, AI technology and application evidence storage and traceability services can achieve the credibility of microservice publishing, use, management and transaction process monitoring, eliminate data and transaction security concerns, and promote technology sharing.

[0074] In summary, the meta-application intelligent service system 100 based on AI technology microservices implemented in this embodiment, first, provides a collaborative working mechanism and service system for technology suppliers, application integrators and operation and maintenance managers, and provides microservice encapsulation, publishing, containerized operation and maintenance, meta-application combination, business verification, trusted traceability, application integration, and full-process services until the application system covering AI technology algorithm data. It can be used to build a flexible supply enterprise technology middle platform and a technology transaction public service platform to enable technology supply, application integration and operation and maintenance management to play their respective strengths and accurately connect. In particular, through technical model description specifications, multi-layer classification of application technology, and visual interactive publishing services, it supports and constrains technology suppliers to publish AI technology that meets business needs. Through technology combination simulation and monomer application construction based on business data and visualization assistance, the application integrator builds a lightweight monomer application that is easy to integrate. Through containerized microservice management tools, it supports the operation and maintenance manager to build and manage the status of AI technology on demand. Second, adopt AI technology microservice components based on microservices to enhance the combinatoriality, coordination, applicability and publishing efficiency of AI technology, adopt lightweight meta-applications based on intelligent agents to achieve flexible integration based on interaction, and improve integration efficiency and compatibility. Third, propose a microservice combination based on large model intelligent agents, that is, meta-applications based on intelligent agents, which are lightweight AI monomer applications that can run independently and support intelligent interaction and flexible integration. Fourth, for AI technology publishers, a microservice publishing method based on knowledge-enhanced large model intelligent agents is adopted to support intelligent division and visual interactive selection of AI algorithm model microservices. Fifth, for AI technology users, services such as auxiliary retrieval, business verification, and full-process trusted traceability based on knowledge-enhanced large language model intelligent agents are provided to achieve precise technology matching for business needs and effects.

[0075] Functions and Effects of the Embodiments

[0076] According to the meta-application intelligent agent service system based on AI technology microservices involved in this embodiment, the AI ​​technology algorithm data is split and customized to generate microservices through the microservice module, and multiple microservices are collaboratively called according to user needs through the meta-application construction module to generate a meta-application based on a large model intelligent agent, and the microservices and meta-applications are operated and managed through the operation and maintenance module. In short, this method can generate corresponding meta-applications flexibly integrated by multiple microservices according to any task, solve the problems of weak composability of technical components, low flexibility of technical combination integration, and little consideration of the needs of technology publishers and integrators in the current technical service system, and realize the effective coordination of technical resource release, application integration and operation and maintenance management.

[0077] Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A meta-application agent service system based on AI technology microservices, used for microservice encapsulation, publishing, containerized operation and maintenance, meta-application combination, business verification, trusted traceability, and application integration of AI technology algorithm data, characterized in that: include: The platform side and multiple user sides owned by different users, Each of the user terminals is connected to the platform terminal for communication. The users include technology suppliers, application integrators and operation and maintenance managers. The platform end includes: The microservice module publishes microservices with specific functions based on the AI ​​technology algorithm data uploaded by the technology supplier through the user end through the technical model description specification, multi-layer classification of application technology, and visual interactive publishing services; A meta-application construction module, including a large language model based on knowledge enhancement, generates multiple microservice combinations and corresponding visualization maps according to the application demand data uploaded by the application integrator through the user terminal, performs meta-application simulation evaluation according to the microservices to be combined selected by the application integrator, and publishes a meta-application integrated with the microservices selected by the application integrator and using the large model agent as the software carrier; The operation and maintenance module includes a containerized microservice management tool, which allocates containers to the microservices and maintains the operation of the containers according to the operation and maintenance instruction data uploaded by the operation and maintenance manager through the user terminal.

2. The meta-application agent service system based on AI technology microservices according to claim 1 is characterized by: in, When the meta-application is running, the meta-application agent generates scheduling instructions according to the user's demand instructions, calls each of the microservices corresponding to the meta-application according to the scheduling instructions, transfers data between the microservices, and interacts with microservices, meta-applications, application systems, cloud platforms, and operating systems.

3. The meta-application agent service system based on AI technology microservices according to claim 1, Features: in, The microservice module includes: The application technology multi-layer classification unit is used by the technology supplier to set the type and technical performance indicators corresponding to the AI ​​technology algorithm data. A visualization unit, including a microservice publishing agent, is used to parse each functional function in the AI ​​technology algorithm data and generate a corresponding visualization microservice map; A microservice draft generation unit, comprising a knowledge-enhanced large language model and preset microservice templates of different types, for generating a corresponding program draft for the function to be released selected by the technology supplier from the visualized microservice map; A customization unit, including a preset microservice description specification, is used by the technology provider to customize the program draft according to the microservice description specification, debug and release the microservice, The types include field subtypes, industry subtypes, scenario subtypes, technology subtypes, algorithm subtypes and function subtypes. The microservice description specification is used to set the description interface of the microservice.

4. The meta-application agent service system based on AI technology microservices according to claim 1 is characterized by: in, The platform also includes a meta-application search and evaluation module and a traceability module. The meta-application search and evaluation module and / or the client includes a meta-application recommendation unit, The meta-application recommendation unit includes a meta-application retrieval agent, which is used to generate at least one recommended meta-application according to the application requirement data. The meta-application retrieval and evaluation module comprises: a business verification unit storing a meta-application test data set, and configured to perform business verification on the recommended meta-application selected by the application integrator according to the business data in the application requirement data and the meta-application test data set, and generate a business verification evaluation result; The traceability module uses blockchain to store the business verification and evaluation results of the recommended meta-application by the application integrator.

5. The meta-application agent service system based on AI technology microservices according to claim 1 is characterized by: in, The platform also includes a microservice evaluation module and a traceability module. The microservice evaluation module stores a microservice test data set, which is used by the user to evaluate the microservice selected by the user based on the uploaded own data set and the microservice test data set, and generate a test result. The traceability module uses blockchain to store the user's evaluation results of the microservice. The microservice evaluation module uses the professional AI technology evaluation program released by the platform or the stored large language model to evaluate the microservice.

6. The meta-application agent service system based on AI technology microservices according to claim 1 is characterized by: in, The operation and maintenance module visually displays the status of the container and the meta-application of the microservice to the operation and maintenance manager, The operation and maintenance module also maintains the operation of the meta-application according to the operation and maintenance instruction data.

7. The meta-application agent service system based on AI technology microservices according to claim 1, Features: in, The platform also includes a traceability module. The meta-application building blocks include: A combination recommendation unit, comprising a large language model based on knowledge enhancement, generating a plurality of microservice combinations and corresponding visualization graphs according to the application demand data; A simulation verification unit stores a preset simulation verification data set, performs simulation verification on the microservice combination selected by the application integrator according to the simulation verification data set and the business data in the application requirement data, and generates a simulation verification evaluation result; The traceability module is used to use blockchain to store the simulation verification and evaluation results of the meta-application published by the user.

8. The meta-application agent service system based on AI technology microservices according to claim 1 is characterized by: in, The platform also includes a traceability module. The traceability module is used to store the evidence materials uploaded by the user to the microservice or the meta-application. The evidence materials include at least one of qualification certificates, intellectual property rights, and application cases.

Images