AI capability sharing method and device based on resource scheduling, terminal and medium
By sorting and parsing the AI service architecture, the problem of resource contention for AI service requests in smart terminals is solved, resulting in reduced storage space and increased running speed, simplified development process, improved user experience and voice recognition accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-18
- Publication Date
- 2026-03-13
AI Technical Summary
In existing smart terminals, the disorderly competition for resources among multiple AI software development kits leads to memory exhaustion and computing power overload, causing lag and slow response, which affects the user experience.
The AI capability platform layer, which adopts an AI service architecture, receives and sorts AI service requests, parses and processes them through the resource scheduling module, generates control commands, and transmits them to the application through the service interface layer, thus avoiding multiple AI service requests from competing for processor resources at the same time.
It reduces the storage space occupied by AI software development kits on smart terminals, improves the terminal's running speed and response speed, simplifies the application development process, and enhances user experience and voice recognition accuracy.
Smart Images
Figure CN121658211A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of equipment resource scheduling technology. More specifically, this application relates to a method, apparatus, terminal, and medium for sharing AI capabilities based on resource scheduling. Background Technology
[0002] With the rapid development of artificial intelligence technology, various AI capabilities (such as speech recognition, speech synthesis, image recognition, and natural language processing) are widely integrated into applications on smart devices to enhance user interaction and intelligence. More specifically, each application on a smart device requiring AI functionality independently integrates and runs its own proprietary AI software development kit (SDK) and corresponding models. This leads to the following drawbacks: the installation package of each application integrating the AI SSD is relatively large, consuming significant storage space on the smart device. Furthermore, when multiple applications run simultaneously on the smart device, or when multiple applications are active in the background (e.g., a video application is listening for voice wake-up, and another music application is preparing for voice search), the AI SSDs for multiple applications will be loaded into memory simultaneously, competing for the computing power of the smart device's central processing unit (CPU), graphics processing unit (GPU), or neural network processor. The disorderly and repetitive resource consumption of multiple AI software development kits can easily lead to memory exhaustion and computing overload on the smart terminal, resulting in overall lag, slow response, or even application unresponsiveness or crashes, seriously damaging the user experience.
[0003] Therefore, existing technologies still need to be improved and enhanced. Summary of the Invention
[0004] The purpose of this application is to provide a method, apparatus, terminal, and medium for sharing AI capabilities based on resource scheduling, which can reduce the storage space occupied by AI software development kits in applications on smart terminals and improve the overall operating speed and response speed of smart terminals. This application is mainly achieved through the following technical solutions: The first aspect of this application provides a resource scheduling-based AI capability sharing method, comprising: The session management module of the AI capability platform layer, which adopts an AI service architecture, receives multiple concurrent AI service requests and sorts them to obtain a request queue. The resource scheduling module of the AI capability platform layer performs AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request. The service interface layer of the AI service architecture transmits the first control instruction of each AI service request in the request queue to the application that integrates the target SDK corresponding to each AI service request in the queue order, so that each application executes the corresponding first control instruction.
[0005] According to one embodiment of this application, the sorting process is a priority sorting process, a time-slice round-robin sorting process, or a shortest job priority queuing process.
[0006] According to one embodiment of this application, before the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: The AI model is loaded using the model management module of the AI capability platform layer so that the resource scheduling module can use the AI model.
[0007] According to one embodiment of this application, after the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: The AI model can be uninstalled or updated using the model management module.
[0008] According to one embodiment of this application, after the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: The cache management module of the AI capability platform layer is used to save and process all first control instructions.
[0009] According to one embodiment of this application, the service interface layer is also used to provide the target SDK to all applications.
[0010] According to one embodiment of this application, the target SDK does not integrate the AI model.
[0011] A second aspect of this application provides an AI capability sharing device based on resource scheduling, comprising: The sorting module is used by the session management module of the AI capability platform layer of the AI service architecture to receive multiple concurrent AI service requests, sort the multiple AI service requests, and obtain a request queue. The parsing module is used to perform AI parsing processing on each AI service request in the request queue according to the queue order by the resource scheduling module of the AI capability platform layer, and obtain the first control instruction of each AI service request. The transmission module is used to transmit the first control instruction of each AI service request in the request queue to the application that integrates the target SDK corresponding to each AI service request in the order of the queue, so that each application executes the corresponding first control instruction.
[0012] A third aspect of this application provides an intelligent device, including a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to perform the steps of the resource scheduling-based AI capability sharing method described in the first aspect of this application.
[0013] A fourth aspect of this application provides a computer-readable storage medium for storing a computer program that causes a computer to perform the steps of the resource scheduling-based AI capability sharing method provided in the first and second aspects of this application.
[0014] The beneficial effects of this application include: This application designs an AI service architecture including an AI capability platform layer and a service interface layer. The AI capability platform layer sorts multiple concurrent AI service requests before performing AI parsing and processing to obtain the first control instruction for each AI service request. This avoids the phenomenon of multiple AI service requests simultaneously competing for the computing power of the smart terminal's processor. Furthermore, applications integrating the target SDK only need to receive the first control instruction transmitted by the service interface layer and do not need to use additional AI functions. Therefore, applications integrating the target SDK do not need to integrate an AI software development kit and thus do not need to download dedicated model files. Compared with existing technologies, this application can reduce the storage space occupied by the AI software development kit in the application on the smart terminal and improve the overall operating speed and response speed of the smart terminal. Attached Figure Description
[0015] To more clearly illustrate the technical solutions in this application or conventional technology, the drawings used in the description of the embodiments or conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 The flowcharts for some embodiments of the AI capability sharing method based on resource scheduling of this application are shown below. Figure 2 The flowcharts for the resource scheduling-based AI capability sharing method of this application are shown in some other embodiments; Figure 3 The flowcharts are shown in some further embodiments of the AI capability sharing method based on resource scheduling of this application; Figure 4 This is a block diagram illustrating the principle of the AI capability sharing device based on resource scheduling in some embodiments of this application. Figure 5 This is a block diagram illustrating the principle of the smart terminal in some embodiments of this application. Detailed Implementation
[0017] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the specific embodiments of this application are described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.
[0018] It should be noted that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0019] The terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.
[0020] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units that are expressly listed, but may include other steps or units that are not expressly listed or that are inherent to such process, method, product, or apparatus.
[0021] Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different networks, processor devices, or microcontroller devices.
[0022] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used in this application includes any and all combinations of one or more of the associated listed items.
[0023] The specific embodiments of this application will be further described below with reference to the accompanying drawings.
[0024] like Figure 1 The diagram shown is a flowchart of an AI capability sharing method based on resource scheduling provided in the first aspect of this application. The resource scheduling-based AI capability sharing method is applied to a smart terminal, which may be a television. In other embodiments, the smart terminal may be other types of terminals, which can be specifically configured by those skilled in the art according to actual needs. Figure 1 The AI capability sharing method based on resource scheduling includes: S1. The session management module of the AI capability middleware layer, which adopts the AI service architecture, receives multiple concurrent AI service requests and sorts the multiple AI service requests to obtain a request queue.
[0025] The AI service architecture includes the AI capability platform layer and the service interface layer.
[0026] The AI capability platform layer includes a session management module, a resource scheduling module, a model management module, and a cache management module. The session management module receives multiple concurrent AI service requests, sorts them, and obtains a request queue. It also maintains session states and binds each AI service request to its corresponding session (i.e., context association). The resource scheduling module performs AI parsing processing on each AI service request in the request queue according to the queue order, obtaining the first control instruction for each request, thus achieving unified allocation and load balancing of processor resources for the smart device. The model management module handles the loading, unloading, version management, and updating of AI models. The cache management module saves all first control instructions and other control instructions to optimize the inference results of the AI model and improve response speed. In other embodiments, the AI capability platform layer also provides capabilities, specifically wake-up detection, speech recognition, semantic understanding, and voice broadcasting.
[0027] Each of the multiple AI service requests can come from a different application or be a voice request issued by the user.
[0028] The sorting process can be priority sorting, time-slice round-robin sorting, or shortest job first queuing. In other embodiments, the sorting process can also be other sorting methods, which can be set by those skilled in the art according to actual needs.
[0029] S2. The resource scheduling module of the AI capability middleware layer performs AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request.
[0030] refer to Figure 2 As shown, when one of the AI service requests in the request queue is a user's original voice request, the resource scheduling module performs AI parsing processing on the original voice request to obtain the first control instruction of the original voice request, which includes steps S21, S22 and S23.
[0031] S21. Preprocess the original voice request to obtain the target voice information.
[0032] Step S21 can be understood as... Figure 3 The "audio preprocessing" step in the audio.
[0033] The original voice request can be obtained by continuously capturing ambient audio through a microphone array.
[0034] Further, step S21 includes: performing noise suppression processing on the original voice request to obtain voice information to be processed; and performing echo cancellation processing on the voice information to be processed to obtain the target voice information.
[0035] S22. Call the AI model to perform speech-to-text processing on the target speech information to obtain the first text information. Step S22 can be understood as follows: Figure 3 The “speech recognition” step in the process.
[0036] The AI model is an ASR model (i.e., a speech recognition model). More specifically, the ASR model can be the Doubao model, the Whisper model (Whisper is an open-source automatic speech recognition model launched by OpenAI, which is based on the Transformer architecture and has powerful speech recognition capabilities), the DeepSpeech model (DeepSpeech is an open-source speech recognition engine developed by Mozilla, based on deep learning technology, using an end-to-end architecture, and can efficiently convert speech into text), or the Baixiao model. In other embodiments, the ASR model can also be other models, which can be set by those skilled in the art according to actual needs.
[0037] S23. Call the cloud-based big model to perform semantic understanding processing on the first text information to obtain the second text information, and convert the second text information into the first control command of the original voice request.
[0038] Step S23 can be understood as... Figure 3 The "semantic understanding" step in the process.
[0039] The cloud-based large model refers to an artificial intelligence model with a large number of parameters that runs in the cloud.
[0040] Furthermore, prior to step S22, the step of the resource scheduling module performing AI parsing processing on the original voice request to obtain the first control command of the original voice request further includes: waking up the AI model to put the AI model into a listening state. This step can be understood as... Figure 3 The "wake-up detection" step in the process.
[0041] S3. The service interface layer of the AI service architecture transmits the first control instruction of each AI service request in the request queue to the application that integrates the target SDK corresponding to each AI service request, in the order of the queue, so that each application executes the corresponding first control instruction. Step S3 can be understood as... Figure 3 The "Results Distribution" step in the process.
[0042] The service interface layer is also used to provide the target SDK for all applications and a standardized AI service interface for each application.
[0043] The target SDK does not integrate the AI model. The target SDK is a lightweight SDK. Because the target SDK does not require a built-in AI model, the size of each application is optimized. Furthermore, the target SDK's configuration standardizes the interfaces of the corresponding applications, providing a unified interface for each stage; it also enables asynchronous processing in the corresponding applications, supporting non-blocking calls and preventing application interface lag; and it allows for intelligent scheduling based on the application's state, ensuring that the foreground application responds first.
[0044] The application can be a smart assistant application, a voice control application, a meeting recording application, or a voice translation application. In other embodiments, the application can also be other applications, which can be specifically configured by those skilled in the art according to actual needs.
[0045] Through the above implementation methods, this application designs an AI service architecture including an AI capability platform layer and a service interface layer. The AI capability platform layer sorts multiple concurrent AI service requests before performing AI parsing and processing to obtain the first control instruction for each AI service request. This avoids the phenomenon of multiple AI service requests simultaneously competing for the computing power of the smart terminal's processor. Furthermore, applications integrating the target SDK only need to receive the first control instruction transmitted by the service interface layer and do not need to use additional AI functions. Therefore, applications integrating the target SDK do not need to integrate an AI software development kit and thus do not need to download dedicated model files. Compared with existing technologies, this application can reduce the storage space occupied by the AI software development kit in the application on the smart terminal and improve the overall operating speed and response speed of the smart terminal.
[0046] In some implementations, before the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction for each AI service request, the resource scheduling-based AI capability sharing method further includes: using the model management module of the AI capability platform layer to load an AI model so that the resource scheduling module can use the AI model.
[0047] In some implementations, after the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: using the model management module to unload the AI model or update the AI model.
[0048] Updating the AI model refers to updating the version information of the AI model.
[0049] In some implementations, after the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: using the cache management module of the AI capability platform layer to save all the first control instructions.
[0050] In some implementations, the AI service architecture further includes an application layer, which comprises all applications, each of which invokes the AI services of the AI capability platform layer via a lightweight SDK.
[0051] In practical applications, the AI capability sharing method based on resource scheduling can be implemented as follows: The session management module of the AI capability middleware layer of the AI service architecture receives the original voice request as "switch TV desktop to Easy Desktop". Then, the resource scheduling module of the AI capability middleware layer preprocesses the original voice request and calls the AI model and the cloud big model to analyze and process the preprocessed original voice request (that is, to perform recognition and parsing processing). It is determined that the preprocessed original voice request is a desktop switching request, and then a desktop switching control command (that is, the first control command) is generated. Subsequently, the service interface layer of the AI service architecture transmits the desktop switching control command to the TV settings module of the smart terminal.
[0052] Upon receiving a desktop switching control command, the TV settings module does not immediately execute the desktop switching operation. Instead, it first initiates an interactive confirmation process: a prompt box appears on the TV screen, explicitly asking the user, "Please confirm whether to switch to Easy Desktop?" This prompt box provides two options: confirm and cancel, which the user must actively choose. If the user selects "confirm," the TV settings module will execute the desktop switching control command, switching the current TV desktop environment to Easy Desktop mode. If the user selects "cancel" or there is no response within a certain time, the switching process is terminated, and the TV interface remains unchanged.
[0053] This application scenario combines voice initiation, AI capability platform scheduling, and front-end confirmation to achieve voice-controlled desktop switching while effectively preventing accidental operations through a secondary confirmation mechanism, thereby improving the reliability of the TV system and the user experience.
[0054] It should also be noted that in existing technologies, due to the different AI service providers, SDK versions, and model algorithms used by developers of different applications, the AI capabilities they provide (such as the accuracy of speech recognition, response speed, supported languages, and instruction sets) vary greatly. When using different applications, users may need to adapt to different voice wake-up words, different interaction logics, and inconsistent recognition results, failing to obtain a unified, smooth, and high-quality intelligent experience. This sense of fragmentation in the experience reduces users' trust in and willingness to use the AI functions of smart terminals. Unlike existing technologies, the application in this application does not need to call an AI model, that is, the application does not need to integrate AI functions. Furthermore, this application only uses one AI model. Therefore, when using different applications, users do not need to adapt to different voice wake-up words, different interaction logics, and inconsistent recognition results, thereby improving the user experience.
[0055] In existing technologies, application developers must perform tedious SDK selection, model conversion, and performance tuning to ensure that the AI functionality of their applications runs correctly on different brands and models of smart terminals. This involves different main chips, different NPU (Neural Processing Unit) hardware, and different system versions. This significantly increases the complexity and cost of development, testing, and maintenance. Unlike existing technologies, the application in this application does not need to call AI models, meaning it does not need to integrate AI functionality. Therefore, this application eliminates the need for tedious SDK selection, model conversion, and performance tuning, thus simplifying the development process and reducing development costs.
[0056] In existing technologies, the speech recognition rate of lightweight SDKs for individual applications drops significantly in low signal-to-noise ratio (SNR) environments, and the wake-up success rate decreases sharply with increasing far-field distance, failing to meet practical application needs. However, in this application, none of the applications need to individually call AI models to recognize speech. Therefore, this application can improve the overall speech recognition accuracy of the method in low SNR environments without considering the AI capabilities of individual applications.
[0057] refer to Figure 4 The diagram shown is a schematic block diagram of an AI capability sharing device based on resource scheduling, provided in the second aspect of this application. Figure 4 The AI capability sharing device 100 based on resource scheduling includes: The sorting module 101 is used by the session management module of the AI capability platform layer of the AI service architecture to receive multiple concurrent AI service requests and sort the multiple AI service requests to obtain a request queue. The parsing module 102 is used to perform AI parsing processing on each AI service request in the request queue according to the queue order by the resource scheduling module of the AI capability platform layer, and obtain the first control instruction of each AI service request. The transmission module 103 is used to transmit the first control instruction of each AI service request in the request queue to the application with the target SDK corresponding to each AI service request in the order of the queue, so that each application executes the corresponding first control instruction.
[0058] refer to Figure 5 The diagram shown is a schematic block diagram of a smart terminal provided in the third aspect of this application. Figure 5 In the process, the smart terminal 200 includes a processor 201 and a memory 202. The memory 202 is used to store computer programs, and the processor 201 is used to call and run the computer programs stored in the memory 202 to execute the steps of the resource scheduling-based AI capability sharing method provided in the first aspect of this application.
[0059] A fourth aspect of this application provides a computer-readable storage medium for storing a computer program that causes a computer to perform the steps of the resource scheduling-based AI capability sharing method provided in the first aspect of this application.
[0060] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0061] The technical features of the above embodiments can be combined without changing the basic principles of this application. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0062] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the patent protection scope of this application should be determined by the appended claims.
Claims
1. A method for sharing AI capabilities based on resource scheduling, characterized in that, include: The session management module of the AI capability platform layer, which adopts an AI service architecture, receives multiple concurrent AI service requests and sorts them to obtain a request queue. The resource scheduling module of the AI capability platform layer performs AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request. The service interface layer of the AI service architecture transmits the first control instruction of each AI service request in the request queue to the application that integrates the target SDK corresponding to each AI service request in the queue order, so that each application executes the corresponding first control instruction.
2. The AI capability sharing method based on resource scheduling according to claim 1, characterized in that, The sorting process can be priority sorting, time-slice round-robin sorting, or shortest job priority queuing.
3. The AI capability sharing method based on resource scheduling according to claim 1, characterized in that, Before the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: The AI model is loaded using the model management module of the AI capability platform layer so that the resource scheduling module can use the AI model.
4. The AI capability sharing method based on resource scheduling according to claim 3, characterized in that, After the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: The AI model can be uninstalled or updated using the model management module.
5. The AI capability sharing method based on resource scheduling according to claim 4, characterized in that, The target SDK does not integrate the AI model.
6. The AI capability sharing method based on resource scheduling according to claim 1, characterized in that, After the step of using the resource scheduling module of the AI capability platform layer to perform AI parsing processing on each AI service request in the request queue according to the queue order to obtain the first control instruction of each AI service request, the resource scheduling-based AI capability sharing method further includes: The cache management module of the AI capability platform layer is used to save and process all first control instructions.
7. The AI capability sharing method based on resource scheduling according to claim 1, characterized in that, The service interface layer is also used to provide the target SDK to all applications.
8. An AI capability sharing device based on resource scheduling, characterized in that, include: The sorting module is used by the session management module of the AI capability platform layer of the AI service architecture to receive multiple concurrent AI service requests, sort the multiple AI service requests, and obtain a request queue. The parsing module is used to perform AI parsing processing on each AI service request in the request queue according to the queue order by the resource scheduling module of the AI capability platform layer, and obtain the first control instruction of each AI service request. The transmission module is used to transmit the first control instruction of each AI service request in the request queue to the application that integrates the target SDK corresponding to each AI service request in the order of the queue, so that each application executes the corresponding first control instruction.
9. A smart terminal, characterized in that, include: A processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to perform the steps of the resource scheduling-based AI capability sharing method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, Used to store computer programs that cause a computer to perform the steps of the resource scheduling-based AI capability sharing method according to any one of claims 1 to 7.