Information processing method and device, edge gateway, medium and product
By distributing and scheduling inference tasks of large AI model slices within home devices through edge gateways, the problem of data privacy leakage in edge-cloud collaborative distributed inference is solved, and the security of user data is improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-30
- Publication Date
- 2026-04-10
AI Technical Summary
Distributed inference using edge-cloud collaboration poses a risk of user data privacy breaches and impacts user data security.
The edge gateway receives request information from electronic devices, distributes and schedules electronic devices in the home to perform inference on large AI model slices, avoiding uploading request information to edge servers or cloud servers, and utilizing electronic devices in the home for inference task allocation and data processing.
It improves the security of user data and reduces the risk of data breaches by performing inference on local home devices.
Smart Images
Figure CN121835931A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of artificial intelligence, and particularly relates to an information processing method and device, an edge gateway, a medium and a product. BACKGROUND
[0002] With the rapid development of artificial intelligence (AI) technology, distributed reasoning of edge-cloud collaboration has also developed.
[0003] Distributed reasoning of edge-cloud collaboration refers to deploying part of the reasoning task of an AI large model on an edge server and deploying another part of the reasoning task on a cloud server. The edge server first performs preliminary reasoning based on request information of a user-side electronic device, outputs data information of intermediate layer calculation of the large model, and then transmits the output of the intermediate layer to the cloud server, so that the cloud server continues to complete the remaining reasoning task.
[0004] However, distributed reasoning of edge-cloud collaboration requires a user to upload request information to an edge server or a cloud server, which has a certain risk of privacy leakage and affects user data security. SUMMARY
[0005] Embodiments of the application provide an information processing method and device, an edge gateway and a product, which can solve the problem of low user data security.
[0006] In a first aspect, an information processing method is provided by embodiments of the application, which is applied to an edge gateway, and the method comprises the following steps: receiving request information sent by a first electronic device; sending the request information to the first electronic device, so that the first electronic device calls a first model slice of an AI large model to perform reasoning on the request information and obtains a first reasoning result; receiving the first reasoning result sent by the first electronic device; in a case where the i-th reasoning result sent by the i-th electronic device is received, sending the i-th reasoning result to the (i+1)-th electronic device, so that the (i+1)-th electronic device calls an (i+1)-th model slice of the AI large model to perform reasoning on the i-th reasoning result and obtains an (i+1)-th reasoning result; wherein i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI large model, the N electronic devices are electronic devices in at least two families, the j-th electronic device is deployed with a j-th model slice of the AI large model, and j is a positive integer less than or equal to N and greater than or equal to 1; in a case where the N-th reasoning result sent by the N-th electronic device is received, sending the N-th reasoning result to the first electronic device.
[0007] In a second aspect, an embodiment of the present application provides an information processing device applied to an edge gateway, and the device comprises: The first receiving module is configured to receive request information sent by the first electronic device. The first sending module is configured to send the request information to the first electronic device, so that the first electronic device calls the first model slice of the AI large model to perform inference on the request information, and obtains a first inference result. The second receiving module is configured to receive the first inference result sent by the first electronic device. The second sending module is configured to, in a case where the i-th inference result sent by the i-th electronic device is received, send the i-th inference result to the (i+1)-th electronic device, so that the (i+1)-th electronic device calls the (i+1)-th model slice of the AI large model to perform inference on the i-th inference result, and obtains an (i+1)-th inference result; wherein i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI large model, the N electronic devices are electronic devices in at least two families, the j-th electronic device is deployed with the j-th model slice of the AI large model, and j is a positive integer less than or equal to N and greater than or equal to 1. The third sending module is configured to, in a case where the N-th inference result sent by the N-th electronic device is received, send the N-th inference result to the first electronic device.
[0008] In a third aspect, an embodiment of the present application provides an edge gateway, which comprises a processor and a memory storing computer program instructions; and the processor implements the information processing method provided by the embodiment of the present application when executing the computer program instructions.
[0009] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, and the computer readable storage medium stores computer program instructions; and the computer program instructions are executed by a processor to implement the information processing method provided by the embodiment of the present application.
[0010] In a fifth aspect, an embodiment of the present application provides a computer program product, and instructions in the computer program product are executed by a processor of an edge gateway to enable the edge gateway to execute the information processing method provided by the embodiment of the present application.
[0011] In the embodiment of the present application, the edge gateway receives the request information sent by the first electronic device, sends the request information to the first electronic device, so that the first electronic device calls the first model slice of the AI large model to infer the request information and obtains the first inference result, receives the first inference result sent by the first electronic device, in the case that the i-th inference result sent by the i-th electronic device is received, sends the i-th inference result to the i+1-th electronic device, so that the i+1-th electronic device calls the i+1-th model slice of the AI large model to infer the i-th inference result and obtains the i+1-th inference result, wherein i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI large model, the N electronic devices are electronic devices in at least two families, the j-th electronic device is deployed with the j-th model slice of the AI large model, and j is a positive integer less than or equal to N and greater than or equal to 1; in the case that the N-th inference result sent by the N-th electronic device is received, the N-th inference result is sent to the first electronic device. In this way, the request information does not need to be uploaded to the edge server or the cloud server, only needs to be sent to the edge gateway, the edge gateway can infer through scheduling electronic devices in each family, and the security of user data can be improved. BRIEF DESCRIPTION OF DRAWINGS
[0012] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly introduced. Those skilled in the art can obtain other drawings according to these drawings without creating any creative labor.
[0013] Figure 1 is a structural schematic diagram of an information processing system provided by the embodiments of the present application; Figure 2 is a flowchart of an information processing method provided by the embodiments of the present application; Figure 3 is a process schematic diagram of loading model slices provided by the embodiments of the present application; Figure 4 is a process schematic diagram of model inference provided by the embodiments of the present application; Figure 5 is a structural schematic diagram of an information processing device provided by the embodiments of the present application; Figure 6 is a structural schematic diagram of an edge gateway provided by the embodiments of the present application. DETAILED DESCRIPTION
[0014] The features and exemplary embodiments of various aspects of the present application will be described in detail below with reference to the drawings. In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, but not to limit the present application. The present application can be implemented without some of these specific details by those skilled in the art. The following description of the embodiments is only intended to provide a better understanding of the present application by showing examples of the present application.
[0015] It should be noted that, in this paper, relational terms such as first and second are used merely to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between such entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a list of elements does not only include those elements, but also includes other elements not expressly listed or inherent to such process, method, article or apparatus. Without more limitations, the elements defined by the statement "comprising" do not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the elements.
[0016] It should be noted that the acquisition, storage, use and processing of data in the embodiments of the present application comply with the relevant provisions of national laws and regulations.
[0017] It should be noted that in the embodiments of the present application, some software, components, models and other existing solutions in the industry may be mentioned, which should be considered as exemplary, and the purpose is only to illustrate the feasibility of the implementation of the technical solutions of the present application, but does not mean that the applicant has or will necessarily use the solution.
[0018] The information processing method, device, edge gateway, medium and product provided by the embodiments of the present application will be described in detail below with reference to the drawings through specific embodiments and application scenarios.
[0019] The information processing method and device provided by the embodiments of the present application are applied to an edge gateway. Figure 1 FIG. 1 is a structural schematic diagram of an information processing system provided by the embodiments of the present application.
[0020] In Figure 1In the embodiment, the information processing system 100 includes M families of electronic devices, M optical network units (ONUs) 101, an optical line terminal (OLT) 102, and a broadband remote access server (BRAS) 103, where M is a positive integer greater than or equal to 2.
[0021] Each family of electronic devices is connected to one optical network unit 101, the M families of electronic devices are connected to the optical line terminal 102 through the optical network units 101, the optical line terminal 102 is connected to the broadband remote access server 103, the broadband remote access server 103 is connected to a core router (CR) 104, and the core router 104 is connected to the Internet 105.
[0022] The edge gateway is deployed at the broadband remote access server 103 side.
[0023] The electronic device in the embodiment includes, but is not limited to, a mobile phone, a tablet, a television, a notebook computer, a smart speaker, and the like.
[0024] Figure 2 is a flowchart of an information processing method provided by the embodiment. As shown in Figure 2 The information processing method can include the following steps. Step 201: receiving request information sent by a first electronic device; Step 202: sending the request information to the first electronic device, so that the first electronic device calls a first model slice of an AI large model to perform inference on the request information, and obtains a first inference result; Step 203: receiving the first inference result sent by the first electronic device; Step 204: in a case where the i-th inference result sent by the i-th electronic device is received, sending the i-th inference result to the i+1-th electronic device, so that the i+1-th electronic device calls an i+1-th model slice of the AI large model to perform inference on the i-th inference result, and obtains an i+1-th inference result; where i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI large model, the N electronic devices are electronic devices in at least two families, the j-th electronic device is deployed with a j-th model slice of the AI large model, and j is a positive integer less than or equal to N and greater than or equal to 1; Step 205: in a case where the N-th inference result sent by the N-th electronic device is received, sending the N-th inference result to the first electronic device.
[0025] The specific implementation of each step will be described in detail below.
[0026] In the embodiment of the present application, the edge gateway receives the request information sent by the first electronic device, sends the request information to the first electronic device, so that the first electronic device calls the first model slice of the AI large model to perform inference on the request information and obtains the first inference result, receives the first inference result sent by the first electronic device, in the case that the i-th inference result sent by the i-th electronic device is received, sends the i-th inference result to the (i+1)-th electronic device, so that the (i+1)-th electronic device calls the (i+1)-th model slice of the AI large model to perform inference on the i-th inference result and obtains the (i+1)-th inference result, wherein i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI large model, the N electronic devices are electronic devices in at least two families, the j-th electronic device is deployed with the j-th model slice of the AI large model, and j is a positive integer less than or equal to N and greater than or equal to 1, and in the case that the N-th inference result sent by the N-th electronic device is received, the N-th inference result is sent to the first electronic device. In this way, the request information does not need to be uploaded to the edge server or the cloud server, but only needs to be sent to the edge gateway, the edge gateway can perform inference by scheduling the electronic devices in each family, and the security of user data can be improved.
[0027] In some embodiments of the present application, the first electronic device in the embodiment of the present application can be any one of the N electronic devices.
[0028] In some embodiments of the present application, the information processing method provided in the embodiment of the present application can further include: receiving N electronic device idle resource information sent by N electronic devices; according to the N electronic device idle resource information, the AI large model is divided into N model slices; and the corresponding model slice is sent to the N electronic devices.
[0029] In some embodiments of the present application, the resource information in the embodiment of the present application includes but is not limited to: central processing unit (CPU) utilization, graphics processing unit (GPU) utilization, memory utilization, etc.
[0030] In some embodiments of the present application, before receiving the N electronic device idle resource information sent by the N electronic devices, the information processing method provided in the embodiment of the present application can further include: sending a first plug-in to the N electronic devices; and correspondingly, receiving the N electronic device idle resource information sent by the N electronic devices includes: receiving the N electronic device idle resource information sent by the N electronic devices through the first plug-in.
[0031] In some embodiments of the present application, the first plug-in in the embodiments of the present application can be a resource reporting plug-in for collecting, organizing or reporting system resource usage.
[0032] After receiving the first plug-in sent by the edge gateway, the N electronic devices install the first plug-in and send idle resource information to the edge gateway through the first plug-in. After receiving the idle resource information sent by the N electronic devices, the edge gateway splits the AI large model into N model slices according to the idle resource information of the N electronic devices, and then sends the corresponding model slices to the N electronic devices.
[0033] In some embodiments of the present application, according to the idle resource information of the N electronic devices, the AI large model is split into N model slices, which includes: obtaining the idle resource proportion of the i-th electronic device according to the ratio of the idle resource of the i-th electronic device to the total idle resource of the N electronic devices; and splitting the AI large model into N model slices according to the idle resource proportion of each electronic device in the N electronic devices.
[0034] Exemplarily, taking 4 electronic devices and CPU utilization as an example, the 4 electronic devices are electronic device A, electronic device B, electronic device C and electronic device D.
[0035] Suppose the CPU utilization of electronic device A is 90%, the CPU utilization of electronic device B is 80%, the CPU utilization of electronic device D is 70%, and the CPU utilization of electronic device D is 60%, then the idle CPU of electronic device A is 10%, the idle CPU of electronic device B is 20%, the idle CPU of electronic device C is 30%, and the idle CPU of electronic device D is 40%.
[0036] The idle CPU proportion of electronic device A is: 10% / (10%+20%+30%+40%) = 10%; the idle CPU proportion of electronic device B is: 20% / (10%+20%+30%+40%) = 20%; the idle CPU proportion of electronic device C is: 30% / (10%+20%+30%+40%) = 30%; and the idle CPU proportion of electronic device D is: 40% / (10%+20%+30%+40%) = 40%.
[0037] Suppose the AI large model includes 10 layers of neural network layers, and the 10 layers of neural network layers are the 1st layer of neural network layer to the 10th layer of neural network layer.
[0038] According to the idle CPU ratios of the electronic device A, the electronic device B, the electronic device C and the electronic device D, the AI large model is divided into four model slices, wherein the first model slice includes the first neural network layer of the 10 neural network layers, the second model slice includes the second neural network layer and the third neural network layer of the 10 neural network layers, the third model slice includes the fourth neural network layer to the sixth neural network layer of the 10 neural network layers, and the fourth model slice includes the seventh neural network layer to the tenth neural network layer of the 10 neural network layers.
[0039] The first model slice is sent to the electronic device A, the second model slice is sent to the electronic device B, the third model slice is sent to the electronic device C, and the fourth model slice is sent to the electronic device D.
[0040] In some embodiments of the present application, sending the corresponding model slices to the N electronic devices includes: generating a correspondence between the N model slices and the N electronic devices; sending the correspondence to the N electronic devices; accordingly, receiving slice request information sent by the N electronic devices; wherein the slice request information is used to request the model slice corresponding to the electronic device itself; and sending the model slices corresponding to the N electronic devices respectively to the N electronic devices.
[0041] Exemplarily, taking four electronic devices and CPU utilization as an example, the four electronic devices are the electronic device A, the electronic device B, the electronic device C and the electronic device D.
[0042] Suppose that the CPU utilization of the electronic device A is 90%, the CPU utilization of the electronic device B is 80%, the CPU utilization of the electronic device D is 70%, and the CPU utilization of the electronic device D is 60%, then the idle CPU of the electronic device A is 10%, the idle CPU of the electronic device B is 20%, the idle CPU of the electronic device C is 30%, and the idle CPU of the electronic device D is 40%.
[0043] The idle CPU ratio of the electronic device A is: 10% / (10%+20%+30%+40%) = 10%; the idle CPU ratio of the electronic device B is: 20% / (10%+20%+30%+40%) = 20%; the idle CPU ratio of the electronic device C is: 30% / (10%+20%+30%+40%) = 30%; and the idle CPU ratio of the electronic device D is: 40% / (10%+20%+30%+40%) = 40%.
[0044] Suppose that the AI large model includes 10 neural network layers, and the 10 neural network layers are the first neural network layer to the tenth neural network layer.
[0045] According to the idle CPU ratios of the electronic device A, the electronic device B, the electronic device C and the electronic device D, the AI large model is cut into four model slices, wherein the first model slice includes the first neural network layer of the 10-layer neural network layer, the second model slice includes the second neural network layer and the third neural network layer of the 10-layer neural network layer, the third model slice includes the fourth neural network layer to the sixth neural network layer of the 10-layer neural network layer, and the fourth model slice includes the seventh neural network layer to the tenth neural network layer of the 10-layer neural network layer.
[0046] The correspondence between the first model slice and the electronic device A is established, the correspondence between the second model slice and the electronic device B is established, the correspondence between the third model slice and the electronic device C is established, and the correspondence between the fourth model slice and the electronic device D is established.
[0047] The correspondence between the above model slice and the electronic device is sent to the electronic device A, the electronic device B, the electronic device C and the electronic device D.
[0048] After the electronic device A receives the correspondence between the model slice and the electronic device sent by the edge gateway, the electronic device A requests the first model slice from the edge gateway; the edge gateway sends the first model slice to the electronic device A.
[0049] After the electronic device B receives the correspondence between the model slice and the electronic device sent by the edge gateway, the electronic device B requests the second model slice from the edge gateway; the edge gateway sends the second model slice to the electronic device B.
[0050] After the electronic device C receives the correspondence between the model slice and the electronic device sent by the edge gateway, the electronic device C requests the third model slice from the edge gateway; the edge gateway sends the third model slice to the electronic device C.
[0051] After the electronic device D receives the correspondence between the model slice and the electronic device sent by the edge gateway, the electronic device D requests the fourth model slice from the edge gateway; the edge gateway sends the fourth model slice to the electronic device D.
[0052] Each electronic device loads the received model slice after receiving the model slice sent by the edge gateway.
[0053] In some embodiments of the present application, the edge gateway can number and identify each electronic device according to a customer virtual local area network (CVLAN) through a cloud computing service related protocol (for example, a qing protocol), establish a corresponding relationship between the electronic device and the idle resource information according to the idle resource information reported by each electronic device, determine the number of neural network layers carried by each electronic device according to the idle resource information of each electronic device, and split the AI large model according to the number of neural network layers carried by each electronic device. Each model slice is numbered and identified by a service virtual local area network (SVLAN), and each model slice represents a part of the inference task of the AI large model and can be processed independently. The corresponding relationship between the SVLAN and the electronic device is established, and the corresponding relationship is sent to each electronic device; each electronic device reads and loads the corresponding model slice through the SVLAN to the corresponding address.
[0054] As shown in Figure 3 , the Figure 3 is a process diagram of loading a model slice provided by an embodiment of the present application. The process of loading a model slice includes the following steps: Step 301: The edge gateway sends a resource reporting plug-in to each electronic device; Step 302: Each electronic device reports the idle resource information of each electronic device to the edge gateway through the resource reporting plug-in; Step 303: The edge gateway splits the AI large model according to the idle resource information of each electronic device, numbers and identifies each model slice by SVLAN, establishes the corresponding relationship between the SVLAN and the electronic device, and sends the corresponding relationship to each electronic device; Step 304: Each electronic device reads and loads the corresponding model slice through the SVLAN to the corresponding address.
[0055] The specific implementation of each step of the process of loading a model slice can refer to the description in the above embodiments, which will not be described here by the embodiments of the present application.
[0056] In some embodiments of the present application, a client-server form of connection can be established between the edge gateway and each electronic device to ensure smooth transmission of data and model slices.
[0057] Figure 4 is a process diagram of model inference provided by an embodiment of the present application.
[0058] In Figure 4 , when the edge gateway receives the request information sent by the first electronic device, the edge gateway sends the request information to the first electronic device according to the corresponding relationship between the SVLAN and the electronic device.
[0059] The first electronic device uses the first model slice to reason about the request information, obtains the first reasoning result, and sends the first reasoning result to the edge gateway.
[0060] The edge gateway sends the first inference result to the second electronic device based on the correspondence between SVLAN and electronic device.
[0061] The second electronic device uses the second model slice to infer the first inference result, obtains the second inference result, and sends the second inference result to the edge gateway.
[0062] The edge gateway sends the second inference result to the third electronic device based on the correspondence between SVLAN and electronic device.
[0063] After the edge gateway receives the (N-1)th inference result sent by the (N-1)th electronic device, the edge gateway sends the (N-1)th inference result to the Nth electronic device according to the correspondence between SVLAN and electronic device.
[0064] The Nth electronic device uses the Nth model slice to infer the (N-1)th inference result, obtains the Nth inference result, and sends the Nth inference result to the edge gateway.
[0065] The edge gateway sends the Nth inference result to the first electronic device.
[0066] This application also provides an information processing device, such as... Figure 5 As shown. Figure 5 This is a schematic diagram of the structure of the information processing device provided in the embodiments of this application. The information processing device 500 may include: The first receiving module 501 is used to receive request information sent by the first electronic device; The first sending module 502 is used to send request information to the first electronic device so that the first electronic device can call the first model slice of the AI big model to reason about the request information and obtain the first reasoning result; The second receiving module 503 is used to receive the first inference result sent by the first electronic device; The second sending module 504 is configured to, in a case where the i-th inference result sent by the i-th electronic device is received, send the i-th inference result to an (i+1)th electronic device, so that the (i+1)th electronic device calls an (i+1)th model slice of the AI large model to perform inference on the i-th inference result, to obtain an (i+1)th inference result; wherein i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI large model, the N electronic devices are electronic devices in at least two families, the jth electronic device is deployed with a jth model slice of the AI large model, and j is a positive integer less than or equal to N and greater than or equal to 1; The third sending module 505 is configured to, in a case where the Nth inference result sent by the Nth electronic device is received, send the Nth inference result to the first electronic device.
[0067] In the embodiments of the present application, the edge gateway receives the request information sent by the first electronic device, sends the request information to the first electronic device, so that the first electronic device calls the first model slice of the AI large model to perform inference on the request information, to obtain the first inference result, receives the first inference result sent by the first electronic device, in a case where the i-th inference result sent by the i-th electronic device is received, sends the i-th inference result to the (i+1)th electronic device, so that the (i+1)th electronic device calls the (i+1)th model slice of the AI large model to perform inference on the i-th inference result, to obtain the (i+1)th inference result, wherein i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI large model, the N electronic devices are electronic devices in at least two families, the jth electronic device is deployed with a jth model slice of the AI large model, and j is a positive integer less than or equal to N and greater than or equal to 1; in a case where the Nth inference result sent by the Nth electronic device is received, send the Nth inference result to the first electronic device. In this way, the request information does not need to be uploaded to the edge server or the cloud server, but only needs to be sent to the edge gateway, and the edge gateway can perform inference by scheduling electronic devices in each family, thereby improving the security of user data.
[0068] In some embodiments of the present application, the information processing apparatus provided in the embodiments of the present application further includes: The third receiving module is configured to receive N pieces of idle resource information of N electronic devices sent by the N electronic devices; The splitting module is configured to split the AI large model into N model slices according to the N pieces of idle resource information of the N electronic devices; The fourth sending module is configured to send corresponding model slices to the N electronic devices.
[0069] In some embodiments of the present application, the splitting module is specifically configured to: According to a ratio of the idle resource of the i-th electronic device to total idle resources of the N electronic devices, an idle resource proportion of the i-th electronic device is obtained. According to the idle resource proportion of each of the N electronic devices, the AI large model is divided into N model slices.
[0070] In some embodiments of the present application, the fourth sending module is specifically configured to: generate a corresponding relationship between the N model slices and the N electronic devices; send the corresponding relationship to the N electronic devices; receive slice request information sent by the N electronic devices; wherein the slice request information is used to request a model slice corresponding to the own electronic device; send the model slices corresponding to the N electronic devices to the N electronic devices respectively.
[0071] In some embodiments of the present application, the information processing device provided by the embodiments of the present application further includes: a fifth sending module configured to send a first plug-in to the N electronic devices; Correspondingly, the third receiving module is specifically configured to: receive idle resource information of the N electronic devices sent by the N electronic devices through the first plug-in.
[0072] Figure 6 is a structural schematic diagram of an edge gateway provided by the embodiments of the present application.
[0073] The edge gateway 600 can include a processor 601 and a memory 602 having stored computer program instructions.
[0074] Specifically, the above processor 601 can include a central processing unit (CPU), or a specific integrated circuit (ASIC), or can be configured as one or more integrated circuits implementing the embodiments of the present application.
[0075] The memory 602 can include mass storage for data or instructions. As an example and not by way of limitation, the memory 602 can include a Hard Disk Drive (HDD), a floppy disk drive, a flash memory, a Compact Disc (CD) or other optical disk, a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 602 can include removable or non-removable (or fixed) media, where appropriate. The memory 602 can be internal or external to the edge gateway, as appropriate. In some particular embodiments, the memory 602 is non-volatile solid-state memory.
[0076] In some particular embodiments, the memory can include Read-Only Memory (ROM), Random Access Memory (RAM), a disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical, or other physical / tangible memory storage device. Thus, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions that, when executed (e.g., by one or more processors), are operable to carry out operations described with reference to the information processing method according to the present application.
[0077] The processor 601 implements the information processing method provided by the embodiments of the present application by reading and executing computer program instructions stored in the memory 602.
[0078] In one example, the edge gateway can further include a communication interface 603 and a bus 610. Wherein, as shown, the processor 601, the memory 602, the communication interface 603 are connected through the bus 610 and complete the communication between each other. Figure 6 The processor 601, the memory 602, the communication interface 603 are connected through the bus 610 and complete the communication between each other.
[0079] The communication interface 603 is mainly used to realize the communication between each module, device, unit and / or equipment in the embodiments of the present application.
[0080] Bus 610 includes hardware, software, or both, to couple components of the edge gateway to each other and to couple components of the edge gateway to other components, such as those of a network. While Fig. 1b provides an example of a bus, it is to be understood that other bus architectures can be used in conjunction with the various embodiments. For example, an industry standard architecture (ISA), extended industry standard architecture (EISA), point-to-point, or some other bus architectures can be used. Bus 610 can include any interconnection fabric that is used to interconnect integrated circuit (IC) cores and other design components, such as those discussed herein. In some embodiments, bus 610 can be a multi-drop bus, a point-to-point connection, or some combination thereof. In some embodiments, bus 610 can be a peripheral component interconnect (PCI) bus, PCI-Express (PCI-X) bus, or some other bus. In some embodiments, bus 610 can be a serial bus, such as a serial advanced technology attachment (SATA) bus, a universal serial bus (USB) bus, or some other serial bus. In some embodiments, bus 610 can be a parallel bus, such as a video electronics standards association local (VLB) bus, or some other parallel bus. In some embodiments, bus 610 can be a proprietary bus, such as one used in a system-on-a-chip (SoC).
[0081] The edge gateway can execute the information processing method provided by the embodiments of the present application, so as to achieve the corresponding technical effects of the information processing method provided by the embodiments of the present application.
[0082] In addition, in combination with the information processing method in the above embodiments, the embodiments of the present application also provide a computer readable storage medium to realize. The computer readable storage medium has computer program instructions stored thereon; the computer program instructions are executed by the processor to realize the information processing method provided by the embodiments of the present application. Examples of the computer readable storage medium include non-transitory computer readable media, such as ROM, RAM, magnetic disk or optical disk, etc.
[0083] The embodiments of the present application provide a computer program product, the instructions in the computer program product are executed by the processor of the edge gateway, so that the edge gateway executes the information processing method provided by the embodiments of the present application, and can achieve the same technical effects. To avoid repetition, it will not be repeated here.
[0084] It is to be understood that the application is not limited to the particular configurations and processes described hereinabove and shown in the figures. For the sake of brevity, detailed descriptions of known methods and processes are omitted. In the above embodiments, several specific steps are described and illustrated as examples. However, the method processes of the present application are not limited to the specific steps described and illustrated, and various changes, modifications and additions can be made thereto by one skilled in the art without departing from the spirit of the present application, and the order of the steps can be changed.
[0085] The functional blocks shown in the above described block diagrams of the structure can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, Application Specific Integrated Circuits (ASICs), appropriate firmware, plug-ins, functional cards, etc. When implemented in software, the elements of the present application are program or code segments that are used to perform the required tasks. The program or code segments can be stored in a machine-readable medium or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. A "machine-readable medium" includes any medium that can store or transport information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, Erasable Read Only Memory (EROM), floppy disks, Compact Disc Read-Only Memory (CD-ROM), optical disks, hard disks, fiber optic media, Radio Frequency (RF) links, etc. The code segments can be downloaded via computer networks such as the Internet, intranets, etc.
[0086] It is also to be understood that the exemplary embodiments mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiments, or in an order different from the embodiments, or several steps can be performed simultaneously.
[0087] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.
[0088] The above only is a specific implementation of the present application, and those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, module and unit can refer to the corresponding process in the foregoing method embodiments, which will not be described herein. It should be understood that the protection scope of the present application is not limited to this, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed in the present application, and these modifications or replacements shall be covered within the protection scope of the present application.
Claims
1. An information processing method, characterized in that, The method is applied to an edge gateway; the method includes: Receive request information sent by the first electronic device; The request information is sent to the first electronic device so that the first electronic device can call the first model slice of the artificial intelligence big model to reason about the request information and obtain the first reasoning result; Receive the first inference result sent by the first electronic device; Upon receiving the i-th inference result sent by the i-th electronic device, the i-th inference result is sent to the (i+1)-th electronic device, so that the (i+1)-th electronic device can call the (i+1)-th model slice of the AI big model to infer the i-th inference result and obtain the (i+1)-th inference result; where i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the AI big model, N electronic devices are electronic devices in at least two households, the j-th electronic device deploys the j-th model slice of the AI big model, and j is a positive integer less than or equal to N and greater than or equal to 1; Upon receiving the Nth inference result from the Nth electronic device, the Nth inference result is sent to the first electronic device.
2. The method according to claim 1, characterized in that, Before receiving the request information sent by the first electronic device, the method further includes: Receive the idle resource information of the N electronic devices sent by the N electronic devices; Based on the idle resource information of the N electronic devices, the large artificial intelligence model is divided into N model slices; Send the corresponding model slices to the N electronic devices.
3. The method according to claim 2, characterized in that, The step of dividing the large artificial intelligence model into N model slices based on the idle resource information of the N electronic devices includes: The idle resource ratio of the i-th electronic device is obtained by comparing the idle resources of the i-th electronic device with the total idle resources of the N electronic devices. Based on the proportion of idle resources of each of the N electronic devices, the large artificial intelligence model is divided into N model slices.
4. The method according to claim 2, characterized in that, Sending the corresponding model slices to the N electronic devices includes: Generate the correspondence between the N model slices and the N electronic devices; Send the correspondence to the N electronic devices; Receive slice request information sent by the N electronic devices; wherein, the slice request information is used to request a model slice corresponding to its own electronic device; Send model slices corresponding to each of the N electronic devices to the N electronic devices respectively.
5. The method according to claim 2, characterized in that, Before receiving the idle resource information of the N electronic devices sent by the N electronic devices, the method further includes: Send the first plug-in to the N electronic devices; The receipt of idle resource information from the N electronic devices includes: Receive the idle resource information of the N electronic devices sent by the first plug-in.
6. The method according to claim 1, characterized in that, The edge gateway is deployed on the broadband access server side.
7. An information processing device, characterized in that, The device is used in an edge gateway; the device includes: The first receiving module is used to receive request information sent by the first electronic device; The first sending module is used to send the request information to the first electronic device, so that the first electronic device can call the first model slice of the artificial intelligence big model to reason about the request information and obtain the first reasoning result; The second receiving module is used to receive the first inference result sent by the first electronic device; The second sending module is used to send the i-th inference result to the (i+1)-th electronic device after receiving the i-th inference result sent by the i-th electronic device, so that the (i+1)-th electronic device can call the (i+1)-th model slice of the artificial intelligence big model to infer the i-th inference result and obtain the (i+1)-th inference result; where i is a positive integer less than or equal to N-1 and greater than or equal to 1, N is the number of model slices of the artificial intelligence big model, N electronic devices are electronic devices in at least two households, the j-th electronic device deploys the j-th model slice of the artificial intelligence big model, and j is a positive integer less than or equal to N and greater than or equal to 1; The third sending module is used to send the Nth inference result to the first electronic device when it receives the Nth inference result sent by the Nth electronic device.
8. An edge gateway, characterized in that, The edge gateway includes: a processor and a memory storing computer program instructions; The processor reads and executes the computer program instructions to implement the information processing method as described in any one of claims 1-6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the information processing method as described in any one of claims 1-6.
10. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the edge gateway, the edge gateway performs the information processing method as described in any one of claims 1-6.