Information processing method, edge agent and system
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING WODONG TIANJUN INFORMATION TECH CO LTD
- Filing Date
- 2026-03-16
- Publication Date
- 2026-06-19
Smart Images
Figure CN122240241A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to an information processing method, edge agent, and system. Background Technology
[0002] Current artificial intelligence (AI) systems mainly consist of client-server (CS) architecture and browser-server (BS) architecture.
[0003] In related technologies, the interaction process of an AI system based on the above architecture includes: the user inputs request content on the client or browser of a smart terminal; the client or browser sends the user's input request content to the server; the server analyzes the request content based on a large model and responds to the request content; the client or browser receives the response and displays the response to the user.
[0004] The aforementioned AI-based interaction process has the following problems: First, since each user request needs to be sent to the server for analysis, and the network request transmission and the server's large model for analyzing the requests take a long time, the response is not timely, affecting the user experience; Second, the large number of user requests transmitted to the server results in high operating and computing costs for the server; Third, when the user's terminal experiences network failure, the user cannot use the AI service, affecting the user experience; Fourth, users need to log in to a specific application page or a specific browser page to interact with AI, affecting the convenience of using the AI service. Summary of the Invention
[0005] To address at least one of the problems in the interaction process of AI-based systems in related technologies, this disclosure provides an information processing method, an edge agent, and a system.
[0006] According to a first aspect of this disclosure, an information processing method is provided, the method being executed by an edge agent in a terminal device, comprising: monitoring whether a user triggers a specified operation event on the terminal device based on a kernel event monitoring mechanism of an operating system; after monitoring that the user has triggered the specified operation event on the terminal device, acquiring user-inputted demand information; using a large model deployed locally on the terminal device to perform intent recognition on the user-inputted demand information to obtain a first intent recognition result; and, if a task execution script matching the first intent recognition result exists on the terminal device, executing the task execution script matching the first intent recognition result to obtain a task execution result.
[0007] In some embodiments, the method further includes: if the terminal device does not have a task execution script that matches the first intent recognition result, sending the user-inputted demand information to an artificial intelligence (AI) server, so that the AI server performs intent recognition on the user-inputted demand information based on a locally deployed large model to obtain a second intent recognition result; receiving a task execution script that matches the second intent recognition result from the AI server; and executing the task execution script that matches the second intent recognition result to obtain a task execution result.
[0008] In some embodiments, sending the user-inputted request information to the AI server when the terminal device does not have a task execution script that matches the first intent recognition result includes: when the terminal device does not have a task execution script that matches the first intent recognition result, outputting a prompt to the user whether to use the large model service provided by the AI server; and after receiving a response from the user confirming the use of the large model service provided by the AI server, sending the user-inputted request information to the AI server via a network request.
[0009] In some embodiments, the method further includes: after receiving a response from the user confirming that they will not use the large model service provided by the AI server, sending the user-inputted demand information to other edge agents, so that the other edge agents can perform intent recognition on the user-inputted demand information to obtain a third intent recognition result; receiving a task execution script matching the third intent recognition result from the other edge agents; and executing the task execution script matching the third intent recognition result to obtain a task execution result.
[0010] In some embodiments, obtaining the user-inputted requirement information includes: taking a screenshot of the text input area or text selection area on the terminal device when the user inputs the requirement information via text input or text selection to obtain an image to be processed; and using an optical character recognition algorithm to perform text recognition on the image to be processed to obtain the user-inputted requirement information.
[0011] In some embodiments, the method further includes: after receiving a startup instruction, loading the task execution script in the terminal device into a cache; and / or obtaining the task execution script from an edge computing server and storing the task execution script.
[0012] In some embodiments, the specified operation event includes at least one of a user's mouse double-click event, mouse text selection event, or keyboard input event.
[0013] According to a second aspect of this disclosure, an edge agent is provided, comprising: a module for performing the information processing method as described above.
[0014] According to a third aspect of this disclosure, an edge agent is provided, comprising: a memory; and a processor coupled to the memory, the processor being configured to perform the information processing method as described above based on instructions stored in the memory.
[0015] According to a fourth aspect of this disclosure, a system is provided, comprising: an edge agent as described above; an edge computing server configured to obtain a task execution script from an artificial intelligence (AI) server and send the task execution script to the edge agent; and the AI server.
[0016] According to a fifth aspect of this disclosure, a computer-readable storage medium is provided that stores computer instructions thereon, which, when executed by a processor, implement the information processing method as described above.
[0017] According to a sixth aspect of this disclosure, a computer program product is provided having computer program instructions stored thereon, which, when executed by a processor, implement the information processing method as described above.
[0018] Other features and advantages of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description
[0019] The accompanying drawings, which form part of this specification, illustrate embodiments of this disclosure and, together with the specification, serve to explain the principles of this disclosure.
[0020] Figure 1 This is a schematic flowchart of an information processing method according to some embodiments of the present disclosure; Figure 2 This is a schematic flowchart of an information processing method according to other embodiments of this disclosure; Figure 3 This is a schematic diagram of the structure of an edge agent according to some embodiments of the present disclosure; Figure 4 This is a schematic diagram of the structure of a computer system according to some embodiments of the present disclosure; Figure 5 This is a schematic diagram of the structure of an edge computing-based AI system according to some embodiments of the present disclosure; Figure 6 This is a schematic diagram of the structure of an edge computing-based AI system according to other embodiments of the present disclosure.
[0021] This disclosure can be more clearly understood with reference to the accompanying drawings and the following detailed description. Detailed Implementation
[0022] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.
[0023] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.
[0024] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.
[0025] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.
[0026] In all examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.
[0027] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.
[0028] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.
[0029] To address the problems existing in related technologies, this disclosure proposes an information processing method, an edge intelligent agent, and a system that can efficiently and accurately solve common user problems directly on the terminal without a network, reducing the pressure on the server and the cost of computing power.
[0030] Figure 1 This is a flowchart illustrating an information processing method according to some embodiments of this disclosure. Figure 1 As shown, the information processing method is executed by the edge agent and includes steps S1 to S4.
[0031] In step S1, the kernel event listening mechanism of the operating system is used to listen for whether the user triggers a specified operation event on the terminal device.
[0032] Edge agents can be understood as artificial intelligence (AI) applications installed on edge devices. Edge devices can be terminal devices, such as personal computers and mobile phones.
[0033] In practice, the specific operation events that the edge agent needs to listen to may differ depending on the application scenario. For example, in a scenario where the edge agent is deployed on a personal computer, the specified operation events may include at least one of the following: a user's mouse double-click event, a mouse selection of text event, or a keyboard input event. Similarly, in a scenario where the edge agent is deployed on a mobile phone, the specified operation events may include at least one of the following: a control long-press event, a control touch event, a control click event, a control swipe event, or a control drag event.
[0034] In step S2, after listening to the user triggering a specified operation event on the terminal device, the user's input request information is obtained.
[0035] In practice, the methods for obtaining user input information can differ depending on the specific operation event triggered by the user. The following two examples illustrate how to obtain user input information.
[0036] In some examples, when a user inputs their requirements via text input or text selection, a screenshot is taken of the text input or text selection area on the terminal device to obtain the image to be processed; then, an optical character recognition algorithm is used to perform text recognition on the image to obtain the user's input requirements.
[0037] In other examples, when a user inputs their requirements via voice, the local speech recognition model is used to convert the user's speech into text, and the converted text is used as the user's input requirements.
[0038] In step S3, the large model deployed locally on the terminal device is used to perform intent recognition on the user's input request information to obtain the first intent recognition result.
[0039] In some examples, step S3 includes: constructing prompt words based on the user's input demand information; and invoking a locally deployed large model to perform intent recognition on the user's input demand information based on the prompt words, to obtain a first intent recognition result. The first intent recognition result may include key information from the user's input demand information. This key information may be information directly extracted from the user input, or information derived through analysis and reasoning.
[0040] After obtaining the first intent recognition result in step S3, it is determined whether the terminal device possesses a task execution script that matches the first intent recognition result. The task execution script is a pre-generated file containing instructions for resolving the user's problem. Specifically, the task execution script can be an encrypted Lua script or a script written in another programming language. If it is determined that the terminal device possesses a task execution script that matches the first intent recognition result, step S4 is executed. If it is determined that the terminal device does not possess a task execution script that matches the first intent recognition result, step S4 is not executed.
[0041] For example, the terminal device pre-stores various task execution scripts, such as scripts for diagnosing and repairing computer lag and scripts for diagnosing computer network lag. Suppose a user enters the following request information through the input box: "Please help solve the computer lag problem." The intent recognition based on the local large model yields the following first intent recognition results: "computer," "lag," "diagnosis and repair." Then, by searching the local task execution scripts, it is determined that a script matching the first intent recognition result exists locally, namely, a script for diagnosing and repairing computer lag. Then, step S4 can be executed. Suppose a user enters "Please buy me a plane ticket to Chengdu on the 12th of this month" through the input box. The intent recognition based on the local large model yields the following first intent recognition results: "12th of this month," "local to Chengdu," "plane ticket," "purchase." Then, by searching the local task execution scripts, it is determined that no script matching the first intent recognition result exists locally. Then, step S4 is not executed.
[0042] In step S4, if the terminal device has a task execution script that matches the first intent recognition result, the task execution script that matches the first intent recognition result is executed to obtain the task execution result.
[0043] For example, assuming the task execution script that matches the first intent recognition result is a script for diagnosing and repairing computer lag, then the instructions in that script will be executed to meet the user's needs.
[0044] In some embodiments, the information processing method further includes: after step S4, generating response information based on the task execution result and outputting the response information to the user.
[0045] In this embodiment, the above process can quickly resolve common user problems locally, meeting users' needs for AI services. This not only reduces server load, operating costs, and computing power costs, but also allows for more timely responses to user needs, improving the user experience. Furthermore, the information processing method provided in this embodiment can operate offline and does not require users to log in to specific application pages or browser pages, enhancing the convenience of using AI services and further improving the user experience.
[0046] Figure 2 This is a schematic flowchart illustrating an information processing method according to other embodiments of this disclosure. For example... Figure 2 As shown, the information processing method includes edge agent startup and initialization (step 21) and edge agent operation (step 22). The edge agent startup and initialization includes steps 211 to 215, and the edge agent operation includes steps 221 to 227.
[0047] In step 211, the edge computing server obtains the task execution script from the AI server.
[0048] Edge computing servers, also known as primary servers or management servers, are mainly used to manage edge intelligent agents, including the acquisition, distribution, and management of task execution scripts. AI servers, also known as secondary servers, are mainly used to provide online analysis services based on large models and to generate task execution scripts.
[0049] In some examples, the edge computing server periodically and proactively retrieves task execution scripts from the AI server according to a strategy, or the edge computing server proactively retrieves task execution scripts from the AI server after receiving an instruction.
[0050] In other examples, the AI server periodically and proactively pushes task execution scripts, or the AI server proactively pushes task execution scripts to the edge computing server after receiving instructions; the edge server receives the task execution scripts pushed by the AI server.
[0051] In step 212, the edge agent receives a start command.
[0052] In some examples, the edge agent receives a user's start command and executes steps 213 to 215 after starting according to the command.
[0053] In other examples, the edge agent receives a startup command from the edge computing server and executes steps 213 to 215 after startup according to the command.
[0054] In step 213, the edge agent obtains the task execution script from the edge computing server.
[0055] In some examples, the edge agent proactively retrieves task execution scripts from the edge computing server. For instance, the edge agent periodically retrieves task execution scripts from the edge computing server, or the edge agent retrieves task execution scripts from the edge computing server after receiving specific instructions.
[0056] In some examples, the edge agent passively obtains task execution scripts from the edge computing server. For instance, the edge computing server periodically pushes task execution scripts to the edge agent, or the edge computing server pushes task execution scripts to the edge agent after receiving specific instructions.
[0057] In step 214, the edge agent loads the local task execution script into the cache.
[0058] In some examples, step 214 includes: after the edge agent successfully obtains the task execution script from the edge computing server through step 213, it determines the latest task execution script by comparing the obtained task execution script with the existing task execution script locally, and loads the latest task execution script into the cache.
[0059] For example, if the edge agent obtains the computer lag diagnosis and repair script (version V2) and the computer network diagnosis script (version P3) from the edge computing server through step 213, and has the computer lag diagnosis and repair script (version V1) and the computer network diagnosis script (version P2) stored locally, then the computer lag diagnosis and repair script (version V2) and the computer network diagnosis script (version P3) will be loaded into the cache.
[0060] In some examples, step 214 includes: if the edge agent fails to obtain the task execution script through step 213, it loads the locally existing task execution script into the cache. For example, if the edge agent encounters network failures or other problems that prevent it from obtaining the task execution script from the edge computing server, it will only load the locally existing task execution script.
[0061] In step 215, the edge agent initiates kernel event listening.
[0062] In this embodiment, the execution order of steps 214 and 215 is not limited. In practice, steps 214 and 215 can be executed in parallel or sequentially.
[0063] In this embodiment of the disclosure, the startup and initialization of the edge agent are completed through steps 211 to 215, which facilitates the subsequent provision of local AI services to users.
[0064] In step 221, the edge agent listens for a user-triggered specified operation event.
[0065] In step 222, the edge agent obtains the user's input requirements.
[0066] For details on how to perform steps 221 to 222, please refer to the relevant content of the embodiments described above.
[0067] In step 223, the edge agent determines the user's intent.
[0068] After obtaining the user's input request information, the edge agent performs intent recognition on the user's input request information based on the local large model to obtain the first intent recognition result.
[0069] In step 224, the edge agent queries whether there is a matching task execution script locally.
[0070] In some examples, the edge agent queries whether there is a task execution script locally that matches the first intent recognition result. If it is determined that a task execution script matching the first intent recognition result exists locally, step 227 is executed. In the above examples, the information processing method may also include: if it is determined that there is no task execution script matching the first intent recognition result locally, steps 225 to 227 are executed directly. Through the above processing, the AI service capabilities of the terminal device locally can be combined with the AI service capabilities of the server, better meeting user needs and improving the user experience of using AI services.
[0071] In other examples, if a task execution script matching the first intent recognition result exists locally, step 227 is executed. If no task execution script matching the first intent recognition result exists locally, a prompt message is displayed to the user asking whether to use the large model service provided by the AI server. After receiving a response message from the user confirming the use of the large model service provided by the AI server, steps 225 to 227 are executed. In this embodiment of the disclosure, by sending the user-inputted request information to the AI server only after the user confirms the use of the server's AI service, not only is the user interaction during the AI service process more user-friendly, but unnecessary network data transmission is also reduced, lowering network communication overhead and the pressure on the AI server.
[0072] Furthermore, in some of the other examples mentioned above, after receiving a response from the user confirming that they will not use the large model service provided by the AI server, an error message can be displayed to the user. For example, the message could say, "Sorry, we cannot meet your needs at this time." Alternatively, after receiving the response from the user confirming that they will not use the large model service provided by the AI server, the user's input request information can be sent to other edge agents. These edge agents can then perform intent recognition on the user's input request information to obtain a third intent recognition result; and the task execution script matching the third intent recognition result can be received from the other edge agents. By seeking the help of other edge agents, user needs can be better met, and the user experience of using AI services can be improved.
[0073] In step 225, the edge agent sends the user-inputted demand information to the AI server.
[0074] In some examples, the edge agent sends user-inputted request information to the AI server via a network request. Furthermore, in specific implementations, the edge agent may also send other information, such as the initial intent recognition result, to the AI server. The AI server, based on a locally deployed large model, performs intent recognition on the user-input request information to obtain a second intent recognition result. Then, the AI server generates a matching task execution script based on the second intent recognition result and sends this script to the edge agent.
[0075] In step 226, the edge agent receives the matched task execution script.
[0076] In some examples, the edge agent receives task execution scripts directly from the AI server that match the results of the second intent recognition.
[0077] In other examples, the AI server first sends the task execution script that matches the second intent recognition result to the edge computing server; after the edge computing server approves the task execution script, it then sends the script to the edge agent. In other words, the edge agent receives the approved task execution script forwarded by the edge computing server.
[0078] In step 227, the edge agent executes the matched task execution script.
[0079] In some examples, if step 224 determines that a task execution script exists locally that matches the first intent recognition result, the task execution script is executed to obtain the task execution result.
[0080] In some examples, if step 224 determines that no task execution script matching the first intent recognition result exists locally, a task execution script matching the second intent recognition result obtained from the AI server is executed to obtain the task execution result. Alternatively, if step 224 determines that no task execution script matching the first intent recognition result exists locally, a task execution script matching the third intent recognition result obtained from other edge agents is executed to obtain the task execution result.
[0081] In this embodiment, the above process can quickly resolve common user problems locally, meeting users' needs for AI services. This not only reduces server load, operating costs, and computing power costs, but also allows for more timely responses to user needs, improving the user experience. Furthermore, the information processing method provided in this embodiment can operate offline and does not require users to log in to specific application pages or browser pages, enhancing the convenience of using AI services and further improving the user experience.
[0082] Figure 3 This is a schematic diagram of the structure of an edge agent according to some embodiments of this disclosure. For example... Figure 3 As shown, the edge agent 30 includes a listening module 31, an acquisition module 32, an identification module 33, and an execution module 34.
[0083] The listening module 31 is used to listen for whether the user has triggered a specified operation event on the terminal device based on the kernel event listening mechanism of the operating system.
[0084] The acquisition module 32 is used to acquire the user's input request information after listening to the user triggering a specified operation event on the terminal device.
[0085] The recognition module 33 is used to perform intent recognition on the user's input request information using a large model deployed locally on the terminal device, so as to obtain the first intent recognition result.
[0086] The execution module 34 is used to execute the task execution script that matches the first intent recognition result when there is a task execution script on the terminal device, so as to obtain the task execution result.
[0087] In some embodiments, the edge agent 30 also includes modules required for performing other steps of the information interaction method as described above.
[0088] In the embodiments of this disclosure, the above-described device can quickly and directly resolve common user problems locally, meeting users' needs for AI services. This not only reduces server-side pressure, operating costs, and computing power costs, but also allows for more timely responses to user needs, improving the user experience. Furthermore, the information processing method provided in these embodiments can operate without a network connection and does not require users to log in to specific application pages or browser pages, enhancing the convenience of using AI services and further improving the user experience.
[0089] Figure 4 This is a schematic diagram of the structure of a computer system according to some embodiments of the present disclosure. For example... Figure 4 As shown, edge intelligent agents can be represented in the form of general computing devices. The computer system 40 includes a memory 41, a processor 42, and a bus 43 connecting different system components.
[0090] The memory 41 may include, for example, system memory, non-volatile storage media, etc. The system memory may store, for example, an operating system, application programs, a boot loader, and other programs. The system memory may include volatile storage media, such as random access memory (RAM) and / or cache memory. The non-volatile storage media may store, for example, instructions for executing at least one embodiment of the information processing method. Non-volatile storage media include, but are not limited to, disk storage, optical storage, flash memory, etc.
[0091] Processor 42 can be implemented using a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA) or other programmable logic devices, discrete hardware components such as discrete gates or transistors. Correspondingly, each module, such as the monitoring module, acquisition module, identification module, and execution module, can be implemented by the central processing unit (CPU) running instructions in memory to execute the corresponding steps, or by dedicated circuitry to execute the corresponding steps.
[0092] Bus 43 can use any of the various bus architectures. For example, bus architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MCA) bus, and the Peripheral Component Interconnect (PCI) bus.
[0093] These interfaces 44, 45, and 46 in computer system 40, as well as memory 41 and processor 42, can be connected via bus 43. Input / output interface 44 provides a connection interface for input / output devices such as monitors, mice, and keyboards. Network interface 45 provides a connection interface for various networked devices. Storage interface 46 provides a connection interface for external storage devices such as floppy disks, USB flash drives, and SD cards.
[0094] Figure 5 This is a schematic diagram of the structure of an edge computing-based AI system according to some embodiments of this disclosure. Figure 5 As shown, the AI system based on edge computing includes an edge agent 30, an edge computing server 51, and an AI server 52.
[0095] Edge agent 30 is used to perform the information processing method as described above.
[0096] Edge computing server 51 is configured to obtain task execution scripts from AI server 52 and send the task execution scripts to edge agent 30.
[0097] AI Server 52 is used to provide online AI services and generate task execution scripts.
[0098] In this embodiment, the system described above can quickly resolve common user problems locally on the terminal device, meeting users' needs for AI services. This not only reduces server load, operating costs, and computing power costs, but also allows for more timely responses to user needs, improving the user experience. Furthermore, the information processing method provided in this embodiment can operate offline and does not require users to log in to specific application pages or browser pages, enhancing the convenience of using AI services and further improving the user experience.
[0099] Figure 6 This is a schematic diagram of the structure of an edge computing-based AI system according to other embodiments of this disclosure. For example... Figure 6 As shown, the AI system based on edge computing includes an edge agent 30, an edge computing server 51, and an AI server 52.
[0100] Edge agent 30 is used to execute the information processing method described above. The edge agent 30 can be modularized in various ways. For example, in this embodiment, the edge agent 30 includes a system kernel event monitoring module, an image processing module, an OCR module, a language model module, a forward agent module, and a task management and execution module.
[0101] The kernel event listening module is used to listen for whether a user triggers a specified operation event on the terminal device, based on the kernel event listening mechanism of the operating system.
[0102] The image processing module is used to take a screenshot of the user's input text or selected text area after listening to the user triggering a specified operation event on the terminal device, so as to obtain the image to be processed.
[0103] The OCR module is used to perform text recognition on the image to be processed using optical character recognition algorithms in order to obtain the user's input requirements.
[0104] The language model module is used to utilize a large model deployed locally on the terminal device to perform intent recognition on the user's input information in order to obtain the first intent recognition result.
[0105] The task management and execution module is used to execute the task execution script that matches the first intent recognition result when there is a task execution script on the terminal device, so as to obtain the task execution result.
[0106] The forward proxy module is used to send the user's input request information to the AI server when there is no task execution script on the terminal device that matches the first intent recognition result. The AI server then performs intent recognition on the user's input request information based on the locally deployed large model to obtain the second intent recognition result; and receives the task execution script that matches the second intent recognition result from the AI server.
[0107] The task management and execution module is also used to execute a task execution script obtained from the AI server that matches the second intent recognition result when there is no task execution script on the terminal device that matches the first intent recognition result, so as to obtain the task execution result.
[0108] In addition, the edge agent 30 also includes modules such as a skill package management module and a NAT type detection module. The skill package management module manages skill packages (i.e., the "task execution scripts" mentioned earlier). The NAT type detection module detects the Network Address Translation (NAT) type of the edge computing network.
[0109] The edge computing server 51 mainly includes a terminal permission management module, a skill package management module, a version management module, and a policy management module. In addition to managing edge AI bodies, the edge computing server 51 can periodically retrieve the latest skill packages from the AI server according to policies, or retrieve the latest skill packages from the AI server after receiving a notification message. Furthermore, the edge computing server 51 can maintain the addition, deletion, query, and modification of skill packages, and supports administrator approval management of skill packages.
[0110] AI Server 52 can be an AI gateway running on the server side, a server based on Model Context Protocol (MCP), or a server that supports large model computation, etc., and supports online real-time analysis.
[0111] In this embodiment, the system described above can quickly resolve common user problems locally on the terminal, meeting users' needs for AI services. This not only reduces server load, operating costs, and computing power costs, but also allows for more timely responses to user needs, improving the user experience. Furthermore, the information processing method provided in this embodiment can operate offline and does not require users to log in to specific application pages or browser pages, enhancing the convenience of using AI services and further improving the user experience.
[0112] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations thereof, can be implemented by computer-readable program instructions.
[0113] These computer-readable program instructions are provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable device to produce a machine, such that execution of the instructions by the processor produces means for implementing the functions specified in one or more boxes of the flowchart and / or block diagram.
[0114] These computer-readable program instructions may also be stored in a computer-readable storage medium. These instructions cause a computer to work in a particular manner to produce an article of manufacture, including instructions that implement the functions specified in one or more boxes in a flowchart and / or block diagram.
[0115] This disclosure may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.
[0116] The information processing method, edge agent, and system according to this disclosure have been described in detail above. To avoid obscuring the concept of this disclosure, some details known in the art have not been described. Those skilled in the art can fully understand how to implement the technical solutions disclosed herein based on the above description.
Claims
1. An information processing method, characterized in that, The method is executed by an edge agent in a terminal device and includes: Based on the kernel event listening mechanism of the operating system, it listens to whether the user has triggered a specified operation event on the terminal device; After detecting that the user has triggered a specified operation event on the terminal device, the user's input request information is obtained; Using the large model deployed locally on the terminal device, the user's input request information is used to perform intent recognition to obtain a first intent recognition result; If the terminal device has a task execution script that matches the first intent recognition result, the task execution script that matches the first intent recognition result is executed to obtain the task execution result.
2. The method according to claim 1, characterized in that, The method further includes: If the terminal device does not have a task execution script that matches the first intent recognition result, the user-inputted demand information is sent to the artificial intelligence (AI) server, so that the AI server can perform intent recognition on the user-inputted demand information based on a locally deployed large model to obtain a second intent recognition result. Receive a task execution script from the AI server that matches the second intent recognition result; Execute the task execution script that matches the second intent recognition result to obtain the task execution result.
3. The method according to claim 2, characterized in that, The step of sending the user-inputted request information to the AI server when the terminal device does not have a task execution script that matches the first intent recognition result includes: If the terminal device does not have a task execution script that matches the first intent recognition result, a prompt message is output to the user asking whether to use the large model service provided by the AI server. After receiving the user's confirmation to use the large model service provided by the AI server, the system sends the user's input request information to the AI server via a network request.
4. The method according to claim 3, characterized in that, The method further includes: After receiving the response information from the user confirming that they will not use the large model service provided by the AI server, the user's input request information is sent to other edge agents so that the other edge agents can perform intent recognition on the user's input request information to obtain a third intent recognition result. Receive a task execution script that matches the third intent recognition result from the other edge agents; Execute the task execution script that matches the third intent recognition result to obtain the task execution result.
5. The method according to claim 1, characterized in that, The requirement information for obtaining user input includes: When a user inputs the required information via text input or text selection, a screenshot is taken of the text input area or text selection area on the terminal device to obtain the image to be processed. An optical character recognition algorithm is used to perform text recognition on the image to be processed in order to obtain the user's input requirements.
6. The method according to claim 1, characterized in that, The method further includes: Upon receiving the start command, the task execution script in the terminal device is loaded into the cache; and / or, The task execution script is obtained from the edge computing server and stored.
7. The method according to any one of claims 1 to 6, wherein, The specified operation event includes at least one of the following: a user's mouse double-click event, a mouse text selection event, or a keyboard input event.
8. An edge intelligent agent, characterized in that, include: A module for performing the information processing method as described in any one of claims 1 to 7.
9. An edge agent, comprising: Memory; as well as A processor coupled to the memory, the processor being configured to execute the information processing method as described in any one of claims 1 to 7 based on instructions stored in the memory.
10. A system comprising: Edge agents as described in claim 8 or 9; An edge computing server is configured to retrieve a task execution script from an artificial intelligence (AI) server and send the task execution script to the edge agent; and The AI server.
11. A computer-readable storage medium having stored thereon computer instructions that, when executed by a processor, implement the information processing method as described in any one of claims 1 to 7.
12. A computer program product having stored computer program instructions thereon, which, when executed by a processor, implement the information processing method as described in any one of claims 1 to 7.