Method, apparatus and system for providing AI service

Abstract

The present disclosure relates to a method, apparatus and system for providing an AI service, and a computer storage medium. A method for providing an artificial intelligence (AI) service performed by a first AI service network element includes predicting, in a process of providing an AI service for a mobile terminal, situation information in which an AI service network element providing the AI service for the mobile terminal is changed in the future; in a case where the predicted situation information indicates that the AI service network element is changed at a future time, acquiring model information of an AI model corresponding to the AI service and expected context information of the AI model at the future time, the expected context information including the future time; and transmitting the model information of the AI model and the expected context information of the future time to a second AI service network element to which the mobile terminal is switched at the future time, the model information of the AI model and the expected context information of the future time being used to guide the second AI service network element to make service preparation in advance for the AI service of the mobile terminal at the future time.

Description

Technical Field

[0001] This disclosure relates to the field of communication technology, and in particular to methods, apparatus and systems for providing AI services, and computer-storable media. Background Technology

[0002] In the process of mobile communication networks providing AI (Artificial Intelligence) services to mobile terminals, the service providers for these AI services may need to change when the mobile terminals move between different areas. However, due to the characteristics of AI services, different scenarios or needs place significant differences in the requirements for models and algorithms. Some AI models have high requirements for training latency, training bandwidth, and training resources, or high requirements for the timeliness of inference analysis.

[0003] The relevant 3GPP (3rd Generation Partnership Project) specifications only consider how to support AI service continuity when the mobile terminal has switched from the first AI service network element to the second AI service network element. Summary of the Invention

[0004] The relevant technologies do not consider how to ensure the timeliness of AI services when the AI ​​service network elements providing AI services to mobile terminals change.

[0005] To address the aforementioned technical issues, this disclosure proposes a solution that can ensure the timeliness of AI services and improve their efficiency.

[0006] According to a first aspect of this disclosure, a method for providing artificial intelligence (AI) services is provided, executed by a first AI service network element, comprising: during the process of providing AI services to a mobile terminal, predicting information regarding future changes to the AI ​​service network element providing AI services to the mobile terminal; if the predicted information indicates that the AI ​​service network element will change in the future, acquiring model information of the AI ​​model corresponding to the AI ​​service and expected context information of the AI ​​model in the future time, wherein the expected context information of the future time includes the future time; and sending the model information of the AI ​​model and the expected context information of the future time to a second AI service network element that the mobile terminal switches to in the future time, wherein the model information of the AI ​​model and the expected context information of the future time are used to guide the second AI service network element to prepare for the AI ​​service of the mobile terminal in advance in the future time.

[0007] In some embodiments, the information for predicting future changes in AI service network elements that provide AI services to the mobile terminal includes: predicting the future movement trajectory of the mobile terminal to predict future changes in AI service network elements that provide AI services to the mobile terminal, wherein, if the future movement trajectory indicates that the mobile terminal will move from the current area to another area in the future, the predicted information indicates that the mobile terminal will switch areas in the future, triggering a change in the AI ​​service network element.

[0008] In some embodiments, the information for predicting future changes in the AI ​​service network element providing AI services to the mobile terminal includes: identifying devices participating in the AI ​​service provided by the first AI service network element to the mobile terminal as participating devices; predicting at least one of the future congestion situation of the network between the mobile terminal and the first AI service network element and the future resource consumption situation of the participating devices, to predict the information for predicting future changes in the AI ​​service network element providing AI services to the mobile terminal, wherein, if at least one of the future congestion situation and the future resource consumption situation meets the switching conditions at the future time, the predicted information indicates that a change in the AI ​​service network element will occur at the future time.

[0009] In some embodiments, the future congestion situation includes one or more indicators for measuring network congestion, the resource consumption situation includes one or more indicators for measuring resource consumption, and the switching condition includes at least one of the following: at least one of the one or more indicators for measuring network congestion falls within a first preset indicator value range; at least one of the one or more indicators for measuring resource consumption falls within a second preset indicator range.

[0010] In some embodiments, the method of providing AI services further includes: after sending model information of the AI ​​model and expected context information for the future time to the second AI service network element and before the mobile terminal switches to the second AI service network element, continuing to predict information about future changes to the AI ​​service network element that provides AI services to the mobile terminal; and if the predicted information changes, updating and sending at least one of the model information of the AI ​​model and the expected context information for the future time to the second AI service network element that the mobile terminal switches to at the future time, based on the changed information.

[0011] In some embodiments, the method of providing AI services further includes: monitoring the expected context information of the AI ​​model at a future time before the mobile terminal switches to the second AI service network element; and updating and sending the expected context information of the AI ​​model at the future time to the second AI service network element if the expected context information of the AI ​​model at the future time changes.

[0012] In some embodiments, the method of providing AI services further includes: when the mobile terminal switches to the second AI service network element at the future time, sending the actual context information of the AI ​​model at the future time to the second AI service network element, wherein the actual context information is used to guide the second AI service network element to update the service preparation.

[0013] In some embodiments, the method of providing AI services further includes: sending the subscription information of the mobile terminal to the second AI service network element when the mobile terminal switches to the second AI service network element at the future time.

[0014] In some embodiments, the method of providing AI services, including predicting future changes to AI service network elements providing AI services to the mobile terminal, includes requesting other AI service network elements besides the first AI service network element to predict future changes to AI service network elements providing AI services to the mobile terminal.

[0015] In some embodiments, the model information includes model structure information and model parameter information, and the expected context information further includes the resource computing power requirements of the AI ​​model.

[0016] In some embodiments, when the AI ​​service is a service related to a learning task, the expected context information also includes the expected progress of the learning task; when the AI ​​service is a service related to an analysis task, the expected context information also includes the model splitting points of the analysis task.

[0017] In some embodiments, the method of providing AI services further includes: when the AI ​​service is a service related to an analysis task: if the mobile terminal switches to the second AI service network element at the future time and the first AI service network element has not yet completed the analysis task, based on the subscription operation of the second AI service network element, sending the analysis results obtained after completing the analysis task to the second AI service network element, so that the second AI service network element sends the analysis results to the mobile terminal.

[0018] In some embodiments, the method of providing AI services further includes: receiving an AI service request from a consumer for the mobile terminal, wherein the AI ​​service request includes a service type; and providing the mobile terminal with an AI service corresponding to the service type.

[0019] According to a second aspect of this disclosure, an apparatus for providing artificial intelligence (AI) services is provided, deployed in a first AI service network element corresponding to a first region, comprising: a prediction module configured to predict, during the process of providing AI services to a mobile terminal, information regarding future changes to the AI ​​service network element providing AI services to the mobile terminal; an acquisition module configured to acquire model information of an AI model corresponding to the AI ​​service and expected context information of the AI ​​model at the future time, provided that the predicted information indicates a future change in the AI ​​service network element, wherein the expected context information at the future time includes the future time; and a transmission module configured to transmit the model information of the AI ​​model and the expected context information at the future time to a second AI service network element that the mobile terminal switches to at the future time, wherein the model information of the AI ​​model and the expected context information at the future time are used to guide the second AI service network element to prepare for the AI ​​service of the mobile terminal at the future time.

[0020] According to a third aspect of this disclosure, an apparatus for providing artificial intelligence (AI) services is provided, comprising: a memory; and a processor coupled to the memory, the processor being configured to execute the method for providing AI services as described in any of the above embodiments based on instructions stored in the memory.

[0021] According to a fourth aspect of this disclosure, a system for providing artificial intelligence (AI) services is provided, comprising: the apparatus for providing AI services as described in any of the foregoing embodiments.

[0022] In some embodiments, the system providing AI services further includes: a second AI service network element configured to prepare for the AI ​​service of the mobile terminal in advance based on model information of the AI ​​model and expected context information of the future time.

[0023] In some embodiments, the system providing AI services further includes: other AI service network elements besides the first AI service network element, configured to, in response to a request from the device providing AI services, predict future changes to the AI ​​service network elements providing AI services to the mobile terminal, and send the predicted information to the device providing AI services.

[0024] In some embodiments, the system for providing AI services further includes: a consumer configured to send an AI service request for a mobile terminal to the device for providing AI services, wherein the AI ​​service request includes a service type; wherein the device for providing AI services is further configured to provide the mobile terminal with an AI service corresponding to the service type.

[0025] According to a fifth aspect of this disclosure, a computer-storeable medium is provided having computer program instructions stored thereon, which, when executed by a processor, implement the method for providing AI services as described in any of the above embodiments.

[0026] The above embodiments can ensure the timeliness of AI services and improve the efficiency of AI services. Attached Figure Description

[0027] 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.

[0028] This disclosure will become clearer with reference to the accompanying drawings and the following detailed description, wherein:

[0029] Figure 1 This is a flowchart illustrating a method for providing AI services according to some embodiments of the present disclosure;

[0030] Figure 2 This is a block diagram illustrating an apparatus for providing AI services according to some embodiments of the present disclosure;

[0031] Figure 3 This is a block diagram illustrating an apparatus for providing AI services according to other embodiments of this disclosure;

[0032] Figure 4 This is a block diagram illustrating a system for providing AI services according to some embodiments of the present disclosure;

[0033] Figure 5 This is a signaling diagram illustrating a method for providing AI services according to some embodiments of this disclosure;

[0034] Figure 6 This is a block diagram illustrating a computer system for implementing some embodiments of the present disclosure. Detailed Implementation

[0035] 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.

[0036] 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.

[0037] 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.

[0038] 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.

[0039] 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.

[0040] 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.

[0041] Figure 1 This is a flowchart illustrating a method for providing AI services according to some embodiments of the present disclosure.

[0042] like Figure 1 As shown, the method for providing AI services includes steps S110-S130. The method for providing AI services is executed by a first AI service network element. The AI ​​service network element mentioned in this disclosure is a network element that provides AI services to a mobile terminal. For example, the AI ​​service network element includes an NWDAF (Network Data Analytics Function) network element. The mobile terminal is also referred to as UE (User Equipment).

[0043] In step S110, during the process of providing AI services to the mobile terminal, information is provided on the future changes of the AI ​​service network element that provides AI services to the mobile terminal.

[0044] In some embodiments, the future movement trajectory of the mobile terminal is predicted to anticipate potential changes in the AI ​​service network elements providing AI services to the mobile terminal. If the future movement trajectory indicates that the mobile terminal will move from its current area to another area in the future, the predicted information indicates that a region switch will occur in the future, triggering a change in the AI ​​service network elements. In some embodiments, the future movement trajectory of the mobile terminal is predicted based on its historical movement trajectory. For example, a machine learning model can be used to predict the future movement trajectory of the mobile terminal based on its historical movement trajectory.

[0045] In other embodiments, a device is identified as a participating device that provides AI services to the mobile terminal through the first AI service network element. At least one of the future congestion conditions of the network between the mobile terminal and the first AI service network element and the future resource consumption conditions of the participating device are predicted to predict information regarding future changes to the AI ​​service network element providing AI services to the mobile terminal. If at least one of the future congestion conditions and future resource consumption conditions meets the handover conditions at a future time, the predicted information indicates that a change in the AI ​​service network element will occur at a future time. In this case, the mobile terminal is expected to switch from the first AI service network element to another AI service network element at a future time. In some embodiments, the participating device may include the first AI service network element itself, or it may include other network elements or devices besides the first AI service network element.

[0046] In the above embodiments, by predicting future movement trajectories and / or at least one of network congestion or resource consumption, the AI ​​service network element to which the mobile terminal is about to switch can be triggered to prepare for service in advance. In this way, the timeliness of AI services can be further guaranteed, the efficiency of AI services can be further improved, and the occurrence of AI service interruptions or failures, such as model training and inference analysis, can be further reduced.

[0047] The above embodiments are only some specific examples of changes to AI service network elements and do not represent all implementation methods.

[0048] In some embodiments, future congestion conditions include one or more metrics measuring network congestion. Resource consumption conditions include one or more metrics measuring resource consumption. Switching conditions include at least one of the following: at least one of the one or more metrics measuring network congestion falls within a first preset metric range; and at least one of the one or more metrics measuring resource consumption falls within a second preset metric range. For example, metrics measuring network congestion include at least one of network latency, network packet loss rate, and network throughput. For example, metrics measuring resource consumption include the amount of remaining available resources.

[0049] In some embodiments, other AI service network elements besides the first AI service network element can be requested to predict future changes in the AI ​​service network elements providing AI services to the mobile terminal. For example, the other AI service network elements may have a wider service coverage than the first AI service network element. As another example, the other AI service network elements may have the ability to predict device switching or region switching of the mobile terminal.

[0050] In some embodiments, an aggregated AI service network element can be requested to predict the aforementioned situation information. The aggregated AI service network element, as a global AI service network element, has a service area larger than that of the first AI service network element and is used to coordinate AI service network elements in various regions. The aggregated AI service network element can analyze the global handover situation of mobile terminals, thereby further improving the accuracy of predicting future handover information, further ensuring the timeliness of AI services, further improving the efficiency of AI services, and further reducing the occurrence of AI service interruptions or failures such as model training and inference analysis.

[0051] In step S120, if the predicted situation information indicates that a change in the AI ​​service network element will occur in the future, the model information of the AI ​​model corresponding to the AI ​​service and the expected context information of the AI ​​model in the future are obtained. The expected context information in the future includes the future time.

[0052] In some embodiments, model information includes model structure information and model parameter information. For example, model parameter information includes the model's inherent parameters. In some embodiments, model parameter information also includes model aggregation-related information, such as aggregation weights in federated learning.

[0053] In some embodiments, the contextual information may also include the resource computing power requirements of the AI ​​model. For example, the resource computing power requirements include the amount of resources the AI ​​model requires, such as CPU.

[0054] In some embodiments, when the AI ​​service is related to a learning task, the expected contextual information may also include the anticipated progress of the learning task. When the AI ​​service is related to an analytics task, the expected contextual information may also include the model splitting points of the analytics task. For example, the expected contextual information may also include user face information, etc.

[0055] In some embodiments, the expected progress of a learning task is referred to as the expected split point of the learning task. For example, for future time, the expected split point of the learning task represents the expected level of training of the AI ​​model at that future time (e.g., the number of training epochs). In some embodiments, the model split point of an analysis task characterizes the functional breakdown of the analysis task of the AI ​​model. For example, part of the analysis task of the AI ​​model is completed in the network element, and part of the analysis task is completed on the mobile terminal side.

[0056] In step S130, the model information of the AI ​​model and the expected context information of the AI ​​model at a future time are sent to the second AI service network element that the mobile terminal will switch to at the future time. The model information of the AI ​​model and the expected context information of the AI ​​model at the future time are used to guide the second AI service network element to prepare for the AI ​​service of the mobile terminal in advance. That is, after the model information of the AI ​​model and the expected context information at the future time are sent to the second AI service network element, the second AI service network element is triggered to prepare for the AI ​​service of the mobile terminal at the future time based on the model information of the AI ​​model and the expected context information at the future time. The service preparation of this disclosure includes, but is not limited to, models, resource deployment, etc. In some embodiments, the first AI service network element also sends the data source of the AI ​​model to the second AI service network element.

[0057] In the above embodiments, by predicting future changes to AI service network elements on the mobile terminal, and in the event of such changes, the system proactively sends the model information of the AI ​​model corresponding to the AI ​​service and the expected context information of the AI ​​model for the future to the second AI service network element that the mobile terminal will switch to in the future. This triggers the second AI service network element to prepare for service in advance. In this way, the second AI service network element can provide AI services to the mobile terminal as quickly as possible in the future, thereby ensuring the timeliness of AI services, improving the efficiency of AI services, and reducing the occurrence of interruptions or failures in AI services such as model training and inference analysis.

[0058] In some embodiments, after sending the model information of the AI ​​model and the expected context information for the future time to the second AI service network element, and before the mobile terminal switches to the second AI service network element, it continues to predict whether the AI ​​service network element providing AI services to the mobile terminal will change in the future. If the predicted context information changes, at least one of the model information of the AI ​​model and the expected context information for the future time is updated and sent to the second AI service network element that the mobile terminal switches to at the future time, based on the changed context information.

[0059] In some embodiments, the change in situation information may be a change at a future time, a change in the second AI service network element, or both. If the second AI service network element changes, the first AI service network element may also notify the second AI service network element that was not updated to delete the service update information.

[0060] In the above embodiments, by continuously predicting future changes, the system monitors whether the future change information of the mobile terminal changes. When the future change information changes, the model information and expected context information sent to the second AI service network element are updated in a timely manner, thereby triggering the second AI service network element to update its service preparation. This approach further ensures the timeliness of AI services, improves the efficiency of AI services, and reduces the occurrence of interruptions or failures in AI services such as model training and inference analysis.

[0061] In some embodiments, if the predicted situation information remains unchanged, but the situation differs from the predicted situation information when the mobile terminal switches to the second AI service network element in the future, the second AI service network element indicated in the predicted situation information can be notified to cancel the previously performed service preparation and delete the relevant information. The mobile terminal is then notified to perform service preparation at the second AI service network element it actually switches to in the future.

[0062] In some embodiments, before the mobile terminal switches to the second AI service network element, the expected context information of the AI ​​model in the future is monitored. If the expected context information of the AI ​​model in the future changes, the expected context information in the future is updated and sent to the second AI service network element. For example, in this case, the future movement trajectory of the mobile terminal may change or may not change. For example, if the future movement trajectory of the mobile terminal does not change, the expected context information in the future may change due to the service progress of the first AI service network element being advanced or delayed.

[0063] In the above embodiments, monitoring the expected context information separately can take into account information updates caused by other situations besides region switching, thereby further ensuring the timeliness of AI services, further improving the efficiency of AI services, and further reducing the occurrence of AI service interruptions or failures such as model training and inference analysis.

[0064] In some embodiments, if the mobile terminal switches to the second AI service network element at a future time, the actual context information of the AI ​​model at that future time is sent to the second AI service network element. This context information is actually used to guide the second AI service network element in updating its service preparation.

[0065] In the above embodiments, under the premise of using estimated context information to guide the second AI service network element to prepare for service in advance, the first AI service network element will also send the actual context information to the second AI service network element in the future to trigger the target associated device or the second AI service network element to update the service preparation. In this way, the situation where AI service fails due to the deviation between actual context information and estimated context information can be reduced, thereby further ensuring the timeliness of AI services, further improving the efficiency of AI services, and further reducing the occurrence of AI service interruptions or failures such as model training and inference analysis.

[0066] In some embodiments, if the mobile terminal switches to the second AI service network element at a future time, the mobile terminal's subscription information is sent to the second AI service network element. For example, the subscription information includes the consumer's callback URI (Uniform Resource Identifier) ​​and the data source ID for data collection. In some embodiments, the mobile terminal's subscription information is used to guide the second AI service network element in authenticating the mobile terminal's subscription.

[0067] In some embodiments, if the AI ​​service is a service related to the analysis task, and the mobile terminal switches to the second AI service network element in the future while the first AI service network element has not yet completed the analysis task, the analysis results obtained after completing the analysis task are sent to the second AI service network element based on the subscription operation of the second AI service network element, so that the second AI service network element can send the analysis results to the mobile terminal.

[0068] In some embodiments, an AI service request, including the service type, is received from a consumer for a mobile terminal. The AI ​​service corresponding to the service type is then provided to the mobile terminal. For example, the consumer could be a core network element, an application provider, or an AI network element from another domain.

[0069] Figure 2 This is a block diagram illustrating an apparatus for providing AI services according to some embodiments of the present disclosure. The apparatus for providing AI services is deployed in a first AI service network element.

[0070] like Figure 2 As shown, the device 21 that provides AI services includes a prediction module 211, an acquisition module 212, and a transmission module 213.

[0071] Prediction module 211 is configured to predict, during the process of providing AI services to the mobile terminal, information regarding future changes to the AI ​​service network element providing the AI ​​services to the mobile terminal, such as the execution of... Figure 1 The step S110 is shown. The prediction module can also be called the AI ​​analysis and prediction module.

[0072] The acquisition module 212 is configured to acquire, when the predicted situation information indicates that a change in the AI ​​service network element will occur at a future time, the model information of the AI ​​model corresponding to the AI ​​service and the expected context information of the AI ​​model at that future time. The expected context information for the future time includes the future time itself, for example, the execution of... Figure 1 The step S120 shown.

[0073] The sending module 213 is configured to send model information of the AI ​​model and expected context information for the future time to the second AI service network element that the mobile terminal will switch to in the future time. The model information of the AI ​​model and the expected context information for the future time are used to guide the second AI service network element to prepare for the AI ​​service provided by the mobile terminal in advance, such as performing actions like... Figure 1 The step S130 shown.

[0074] Figure 3 This is a block diagram illustrating an apparatus for providing AI services according to other embodiments of this disclosure.

[0075] like Figure 3 As shown, the apparatus 31 for providing AI services includes a memory 311 and a processor 312 coupled to the memory 311. The memory 311 is used to store instructions for performing methods corresponding to embodiments of the method for providing AI services. The processor 312 is configured to perform methods for providing AI services in any of the embodiments of this disclosure based on the instructions stored in the memory 311. The apparatus 31 for providing AI services is deployed in a first AI service network element.

[0076] Figure 4 This is a block diagram illustrating a system for providing AI services according to some embodiments of the present disclosure.

[0077] like Figure 4 As shown, the system 4 for providing AI services includes an apparatus 41 for providing AI services. The apparatus 41 for providing AI services is deployed in a first AI service network element and executes the methods for providing AI services in any of the embodiments of this disclosure.

[0078] In some embodiments, the system 4 providing AI services further includes a second AI service network element 42. The second AI service network element 42 is configured to prepare for the AI ​​service provided by the mobile terminal in the future time based on model information of the AI ​​model and expected context information of the future time.

[0079] In some embodiments, the system 4 providing AI services further includes other AI service network elements 43 besides the first AI service network element. These other AI service network elements 43 are configured to predict the future movement trajectory of a mobile terminal in response to a request from the device providing the AI ​​services, and send the predicted future movement trajectory to the device 41 providing the AI ​​services. For example, the other AI service network element 43 may include a second AI service network element 42.

[0080] In some embodiments, system 4, which provides AI services, further includes a consumer 44. Consumer 44 is configured to send an AI service request for a mobile terminal to the device providing the AI ​​service, wherein the AI ​​service request includes a service type. Device 41, which provides the AI ​​service, is also configured to provide the mobile terminal with an AI service corresponding to the service type.

[0081] The following will combine Figure 5 This document describes a method for providing AI services in cases of area switching due to changes in movement trajectories. Methods for other cases are similar and will not be elaborated upon here.

[0082] Figure 5 This is a signaling diagram illustrating a method for providing AI services according to some embodiments of the present disclosure.

[0083] like Figure 5 As shown, the method for providing AI services includes steps S500-S505.

[0084] In step S500, the consumer sends an AI service request to the AI ​​service network element in region 1. In some embodiments, the consumer initiates an AI service request, such as a model training or analysis request, to the AI ​​service network element in region 1 based on the location of the mobile terminal UE as needed for analysis.

[0085] In step S501, the AI ​​service network element in region 1 responds to the AI ​​service request sent by the consumer and provides AI services to the UE. For example, it provides AI services such as model training or analysis to the UE.

[0086] In some embodiments, in step S502a, the AI ​​service network element of region 1 autonomously initiates the UE movement trajectory analysis service to predict the UE's future movement trajectory. In this case, step S503a is executed. In step S503a, if the AI ​​service network element of region 1 predicts that the UE is about to leave region 1 and enter region 2, it notifies the AI ​​service network element of region 2 to prepare services for the UE. For example, the notification message carries data source, model information, etc.

[0087] In some embodiments, in step S502b, the AI ​​service network element of region 1 automatically triggers a request to the aggregated AI service network element to start the UE movement trajectory analysis service. In this case, steps S503b and S504 are executed. In step S503b, the aggregated AI service network element notifies the AI ​​service network element of region 1 if it predicts that the UE is about to leave region 1 and enter region 2. In step S504, the AI ​​service network element of region 1 notifies region 2 to prepare for service for the UE.

[0088] Step S501 can be performed before or after step S502a or step S502b.

[0089] After the AI ​​service network element in region 1 notifies the AI ​​service network element in region 2 to prepare for service, step S505 is executed. In step S505, the AI ​​service network element in region 1 notifies the AI ​​service network element in region 2 to interact so that the AI ​​service network element in region 2 can obtain relevant context information. The relevant context information here is the expected context information in the aforementioned embodiment.

[0090] In some embodiments, the method for providing AI services further includes steps S506-S507.

[0091] If the UE has not yet entered Area 2, the UE's future movement trajectory continues to be predicted. If the future movement trajectory changes, step S506 is executed. In step S506, the AI ​​serving network element in Area 1 notifies the AI ​​serving network element in Area 2 to update service preparation. The continued prediction of the UE's future movement trajectory can refer to the implementation process of steps S502a, S502b, S503a, S503b, and S504.

[0092] In step S507, the AI ​​service network element of region 1 notifies the AI ​​service network element of region 2 to interact, so that the AI ​​service network element of region 2 can obtain relevant context information. The relevant context information here is the expected context information in the aforementioned embodiment.

[0093] In some embodiments, the method for providing AI services further includes steps S508-S509.

[0094] When the UE enters region 2, step S508 is executed. In step S508, the AI ​​service network element of region 1 transmits subscription information to the AI ​​service network element of region 2. In step S509, the AI ​​service network element of region 1 notifies the AI ​​service network element of region 2 to interact so that the AI ​​service network element of region 2 can obtain relevant context information. The relevant context information here is the actual context information in the aforementioned embodiments. For example, if the AI ​​service is an analysis service, the actual context information also includes information indicating whether the analysis (also known as inference) is complete.

[0095] In some embodiments, the method of providing AI services further includes step S510. In step S510, after the UE enters area 2, the AI ​​service network element of area 2 continues to provide AI services to the UE based on the previously prepared service.

[0096] In some embodiments, the method for providing AI services further includes steps S511-S512. In step S511, when the AI ​​service network element in region 1 provides analysis services to the UE and the AI ​​service network element in region 1 has not yet completed the analysis service when the UE enters region 2, the AI ​​service network element in region 2 interacts with the AI ​​service network element in region 1 to complete the process of the AI ​​service network element in region 2 subscribing to the analysis results of the AI ​​service network element in region 1.

[0097] In step S512, after the AI ​​service network element in region 1 completes the analysis service, it sends the analysis results to the AI ​​service network element in region 2, so that the AI ​​service network element in region 2 can send the analysis results to the UE.

[0098] The implementation examples for network congestion and resource consumption monitoring are the same as those described above. Figure 5 The implementation examples are similar and will not be described again here.

[0099] Figure 6 This is a block diagram illustrating a computer system for implementing some embodiments of the present disclosure.

[0100] like Figure 6 As shown, the computer system 60 can be represented in the form of a general computing device. The computer system 60 includes a memory 610, a processor 620, and a bus 600 connecting different system components.

[0101] The memory 610 may include, for example, system memory, non-volatile storage media, etc. The system memory may store, for example, an operating system, applications, 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 performing corresponding embodiments of the method for providing AI services. Non-volatile storage media include, but are not limited to, disk storage, optical storage, flash memory, etc.

[0102] The processor 620 can be implemented using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete hardware components such as discrete gates or transistors. Accordingly, each module, such as the decision module and the determination module, can be implemented by executing instructions in the central processing unit (CPU) memory to perform the corresponding steps, or by implementing dedicated circuitry to perform the corresponding steps.

[0103] Bus 600 can use any of the various bus architectures. For example, bus architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, and Peripheral Component Interconnect (PCI) bus.

[0104] The computer system 60 may also include an input / output interface 630, a network interface 640, and a storage interface 650. These interfaces 630, 640, and 650, as well as the memory 610 and processor 620, can be connected via a bus 600. The input / output interface 630 provides a connection interface for input / output devices such as a monitor, mouse, and keyboard. The network interface 640 provides a connection interface for various networked devices. The storage interface 650 provides a connection interface for external storage devices such as floppy disks, USB flash drives, and SD cards.

[0105] 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.

[0106] 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.

[0107] 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.

[0108] This disclosure may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0109] The methods, apparatus, systems, and computer-storable media for providing AI services described in the above embodiments can ensure the timeliness of AI services and improve their efficiency.

[0110] This concludes the detailed description of the methods, apparatus, systems, and computer-storable media for providing AI services according to this disclosure. To avoid obscuring the concept of this disclosure, some details known in the art have not been described. Those skilled in the art will fully understand how to implement the technical solutions disclosed herein based on the above description.

Claims

1. A method for providing artificial intelligence (AI) services, executed by a first AI service network element, comprising: In the process of providing AI services to mobile terminals, information is provided on the future changes of the AI ​​service network elements that provide AI services to the mobile terminals. If the predicted situation information indicates that an AI service network element will change in the future, obtain the model information of the AI ​​model corresponding to the AI ​​service and the expected context information of the AI ​​model in the future time, wherein the expected context information in the future time includes the future time. The model information of the AI ​​model and the expected context information of the future time are sent to the second AI service network element that the mobile terminal switches to at the future time. The model information of the AI ​​model and the expected context information of the future time are used to guide the second AI service network element to prepare for the AI ​​service of the mobile terminal in advance.

2. The method for providing artificial intelligence (AI) services according to claim 1, wherein, Information regarding future changes to the AI ​​service network elements providing AI services to the mobile terminal includes: The future movement trajectory of the mobile terminal is predicted to predict the future changes of the AI ​​service network element that provides AI services to the mobile terminal. In cases where the future movement trajectory indicates that the mobile terminal will move from the current area to another area in the future, the predicted situation information indicates that the mobile terminal will switch areas in the future, triggering a change of the AI ​​service network element.

3. The method for providing artificial intelligence (AI) services according to claim 1, wherein, Information regarding future changes to the AI ​​service network elements providing AI services to the mobile terminal includes: The device that participates in the AI ​​service provided by the first AI service network element to the mobile terminal is identified as the participating device. The system predicts at least one of the future congestion of the network between the mobile terminal and the first AI service network element and the future resource consumption of the participating devices, in order to predict information on the future changes of the AI ​​service network element that provides AI services to the mobile terminal. The predicted information indicates that the AI ​​service network element will change in the future if at least one of the future congestion and future resource consumption conditions is met at the future time.

4. The method for providing artificial intelligence (AI) services according to claim 3, wherein, The future congestion situation includes one or more metrics for measuring network congestion, the resource consumption situation includes one or more metrics for measuring resource consumption, and the switching conditions include at least one of the following: At least one of the one or more metrics for measuring network congestion falls within the range of the first preset metric value. At least one of the one or more indicators for measuring resource consumption falls within the range of the second preset indicator.

5. The method for providing artificial intelligence (AI) services according to claim 1, further comprising: After sending the model information of the AI ​​model and the expected context information of the future time to the second AI service network element and before the mobile terminal switches to the second AI service network element, the system continues to predict information about future changes in the AI ​​service network element that provides AI services to the mobile terminal. If the predicted situation changes, at least one of the model information of the AI ​​model and the expected context information of the future time is updated and sent to the second AI service network element that the mobile terminal switches to at the future time, based on the changed situation information.

6. The method for providing artificial intelligence (AI) services according to claim 1, further comprising: Before the mobile terminal switches to the second AI service network element, monitor the expected context information of the AI ​​model at the future time. If the expected context information of the AI ​​model at the future time changes, the expected context information at the future time is updated and sent to the second AI service network element.

7. The method for providing artificial intelligence (AI) services according to claim 1, further comprising: When the mobile terminal switches to the second AI service network element at the future time, the actual context information of the AI ​​model at the future time is sent to the second AI service network element, wherein the actual context information is used to guide the second AI service network element to update the service preparation.

8. The method for providing artificial intelligence (AI) services according to claim 1, further comprising: If the mobile terminal switches to the second AI service network element at the future time, the subscription information of the mobile terminal is sent to the second AI service network element.

9. The method for providing artificial intelligence (AI) services according to claim 1, wherein the information predicting future changes in the AI ​​service network element providing AI services to the mobile terminal includes: The system requests information from other AI service network elements besides the first AI service network element to predict future changes in the AI ​​service network elements providing AI services to the mobile terminal.

10. The method for providing artificial intelligence (AI) services according to claim 1, wherein, The model information includes model structure information and model parameter information, and the expected context information also includes the resource computing power requirements of the AI ​​model.

11. The method for providing artificial intelligence (AI) services according to claim 1, wherein, When the AI ​​service is a service related to a learning task, the expected context information also includes the expected progress of the learning task; When the AI ​​service is a service related to the analysis task, the expected context information also includes the model splitting points of the analysis task.

12. The method for providing artificial intelligence (AI) services according to claim 1, further comprising: For cases where the AI ​​service is related to an analysis task: If the mobile terminal switches to the second AI service network element at the future time and the first AI service network element has not yet completed the analysis task, the analysis results obtained after completing the analysis task are sent to the second AI service network element based on the subscription operation of the second AI service network element, so that the second AI service network element can send the analysis results to the mobile terminal.

13. The method for providing artificial intelligence (AI) services according to claim 1, further comprising: Receive AI service requests from consumers for the mobile terminal, wherein the AI ​​service requests include service types; Provide the mobile terminal with AI services corresponding to the service type.

14. An apparatus for providing artificial intelligence (AI) services, deployed in a first AI service network element, comprising: The prediction module is configured to predict, during the process of providing AI services to a mobile terminal, information about future changes in the AI ​​service network element providing AI services to the mobile terminal. The acquisition module is configured to acquire, when the predicted situation information indicates that an AI service network element change will occur in the future, the model information of the AI ​​model corresponding to the AI ​​service and the expected context information of the AI ​​model in the future time, wherein the expected context information in the future time includes the future time. The sending module is configured to send the model information of the AI ​​model and the expected context information of the future time to the second AI service network element that the mobile terminal switches to at the future time, wherein the model information of the AI ​​model and the expected context information of the future time are used to guide the second AI service network element to prepare for the AI ​​service of the mobile terminal in advance.

15. An apparatus for providing artificial intelligence (AI) services, deployed in a first AI service network element, comprising: Memory; as well as A processor coupled to the memory, the processor being configured to perform the method of providing artificial intelligence (AI) services as described in any one of claims 1 to 13, based on instructions stored in the memory.

16. A system for providing artificial intelligence (AI) services, comprising: The apparatus for providing artificial intelligence (AI) services as described in claim 14 or 15.

17. The system for providing artificial intelligence (AI) services according to claim 16, further comprising: The second AI service network element is configured to prepare for the AI ​​service of the mobile terminal in advance based on the model information of the AI ​​model and the expected context information of the future time.

18. The system for providing artificial intelligence (AI) services according to claim 16 or 17, further comprising: Other AI service network elements besides the first AI service network element are configured to, in response to a request from the device providing the AI ​​service, predict future changes to the AI ​​service network elements providing the AI ​​service to the mobile terminal, and send the predicted information to the device providing the AI ​​service.

19. The system for providing artificial intelligence (AI) services according to claim 16, further comprising: A consumer is configured to send an AI service request for a mobile terminal to the device providing the AI ​​service, wherein the AI ​​service request includes a service type; The device for providing AI services is further configured to provide AI services corresponding to the service type to mobile terminals.

20. A computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the method of providing artificial intelligence (AI) services as described in any one of claims 1 to 13.

Images