Apparatus, method, and computer readable medium for activating artificial intelligence and / or machine learning capabilities

Abstract

Aspects of the present disclosure relate to a network device (101) comprising at least one processor (510) and at least one memory (520), the at least one memory (520) comprising computer program code (570), the at least one memory and the computer program code being configured to, together with the at least one processor, generate at least one memory code (570). Causing the network device to perform at least: obtaining (201) information describing capabilities of a network entity (103) of another network device (102), the network entity comprising: capabilities for providing an output based on artificial intelligence and / or machine learning reasoning functions related to one or more network objects or object types; sending (203, 404) a request for partial activation of the capability to other network devices; other aspects relate to other network devices, methods respectively associated with each network device, and computer readable media storing appropriate computer program code.

Description

Technical Field

[0001] Various exemplary embodiments relate to devices, methods, and computer-readable media for activating artificial intelligence and / or machine learning capabilities. Background Art

[0002] Artificial intelligence (AI) and machine learning (ML) technologies are increasingly being used in 5G systems and will be widely used in future systems such as 6G. AI / ML technologies have been proposed to implement network data analysis functions, such as "NWDAF" in 3GPP5G Core and Management Data Analysis Service (MDAS) in 3GPP Operation, Administration and Management (OAM).

[0003] The technical specification 3GPP TS28.105, V17.0.0 of June 15, 2022 specifies capabilities and services related to AI / ML management, mainly focusing on AI / ML training. The technical report 3GPP TR 28.908 currently being drafted aims to discuss use cases, potential requirements, and possible solutions for managing further AI / ML capabilities such as AI / ML verification, testing, deployment, configuration, and performance evaluation.

[0004] This disclosure proposes a flexible method for activating AI / ML capabilities. Summary of the invention

[0005] The protection scope is described by the independent claims. Embodiments, examples, and features described in this specification that do not fall within the protection scope (if any) are to be interpreted as various embodiments and useful examples for understanding the protection scope.

[0006] According to a first aspect, there is provided a method performed by a first network device, the first network device comprising at least one processor and at least one memory, the at least one memory comprising computer program code, the computer program code when executed by the at least one processor causing the device to perform the method, the method comprising:

[0007] - obtaining information describing capabilities of a network entity of the second network device, the network entity comprising: capabilities for providing outputs based on artificial intelligence and / or machine learning reasoning functions related to one or more network objects or object types;

[0008] - Sending a request for partial activation of the capability to the second network device.

[0009] According to one or more embodiments, the method includes: obtaining information describing a plurality of predefined ordered activation levels, the plurality of predefined ordered activation levels identifying gradually increasing capability activation ranges; and wherein requesting to describe a partial activation range by one of the reference predefined activation levels.

[0010] According to one or more embodiments, the method includes: receiving, in response to the request, information describing a status of an activation scope of the capability from the second device.

[0011] According to one or more embodiments, the information describing the capability includes one or more of the following:

[0012] a. One or more objects and / or object types, the network entity is configured to optimize or control one or more objects and / or object types;

[0013] b. one or more parameters of one or more objects and / or object types, the network entity being configured to control or optimize one or more parameters of one or more objects and / or object types to achieve a desired result;

[0014] c. one or more network metrics, the network entity is configured to seek to improve one or more network metrics;

[0015] d. One or more network characteristics associated with one or more objects and / or object types.

[0016] According to one or more embodiments, the request includes: one or more subsets of at least one of the items a to d describing the partial activation range.

[0017] According to one or more embodiments, the request includes: at least one of a network context for describing a partial activation scope, and an artificial intelligence and / or machine learning context.

[0018] According to one or more embodiments, the method includes determining whether data describing performance of the capability upon further activation at a requested activation scope satisfies performance criteria, and in the affirmative, sending a request to the second device to increase the previously requested activation scope.

[0019] According to one or more embodiments, the method includes requesting an increase in the activation range of the capability only if the full activation range has not yet been reached.

[0020] According to one or more embodiments, the method includes:

[0021] If the criterion is not met, a deactivation request is sent to the second device, the deactivation request identifying one of the following:

[0022] ○ Instructions to withdraw the activation request;

[0023] ○ Complete deactivation of capabilities;

[0024] o a reduced activation range to be achieved compared to the activation range where the criteria are not met;

[0025] o one or more active parts in the activation scope of the capability to be deactivated;

[0026] o Levels of increasing range of capability activation in a number of predefined ordered activation levels.

[0027] According to one or more embodiments, the method includes determining that network performance is degraded, and requesting a partial reduction in a capability activation range in response to the determination.

[0028] According to one or more embodiments, the method includes: obtaining from a second network device at least one of an accuracy measurement or data describing a benefit assessment performed by the second device, wherein the accuracy measurement or the data describing the benefit assessment is a function of an output of capabilities for an envisioned partial activation range without applying the output to an operating network environment, and sending a request for partial activation based on at least one of the accuracy measurement or the data describing the benefit assessment.

[0029] According to a second aspect, there is provided a method performed by a second network device, the second network device comprising at least one processor and at least one memory, the at least one memory comprising computer program code, which, when executed by the at least one processor, causes the device to perform the method, the second network device comprising a network entity, the network entity comprising a capability for providing an output based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types, the method comprising partial activation of the capability.

[0030] According to one or more embodiments, the method includes making information describing the capability available to other network devices.

[0031] According to one or more embodiments, the method includes making available to other network devices information describing a plurality of predefined ordered activation levels, the plurality of predefined ordered activation levels identifying increasing ranges of capability activation within an available range.

[0032] According to one or more embodiments, the method includes making at least one of an accuracy measurement or data describing a benefit assessment performed by a second device available to other network devices, wherein the accuracy measurement or data describing the benefit assessment is a function of an output of a capability for a contemplated partial activation range without applying the output to an operating environment.

[0033] According to one or more embodiments, the information describing the capability includes one or more of the following:

[0034] a. One or more objects and / or object types, the network entity is configured to optimize or control one or more objects and / or object types;

[0035] b. one or more parameters of one or more objects and / or object types, the network entity being configured to control or optimize one or more parameters of one or more objects and / or object types to achieve a desired result;

[0036] c. one or more network metrics, the network entity is configured to seek to improve one or more network metrics;

[0037] d. One or more network characteristics associated with one or more objects and / or object types.

[0038] According to one or more embodiments, the partial activation of the capability is performed in response to a request for partial activation from the first network device.

[0039] According to one or more embodiments, the method includes, after activation of the capability at the requested scope, sending, to the first device, state information describing an activation state of the capability with respect to the requested scope.

[0040] According to one or more embodiments, the method includes: receiving in the request a network context, and at least one of an artificial intelligence and / or machine learning context; and in response to the context, limiting the scope of activation of the capability.

[0041] According to one or more embodiments, the method includes determining that network performance is degraded, and partially reducing a capability activation range in response to the determination.

[0042] According to one or more embodiments, activation of a capability includes applying an output of the capability to an operating network environment.

[0043] A third aspect relates to a network device, comprising: at least one processor, at least one memory, the at least one memory comprising computer program code, the at least one memory and the computer program code being configured to, together with the at least one processor, enable the network device to at least perform:

[0044] obtaining information describing capabilities of a network entity of another network device, the network entity including capabilities for providing output based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types;

[0045] A request for partial activation of the capability is sent to another network device.

[0046] According to one or more embodiments, at least one memory and a computer program are also configured to, together with at least one processor, cause the network device to execute: obtaining information describing a plurality of predefined ordered activation levels, the plurality of predefined ordered activation levels identifying gradually increasing capability activation ranges; and wherein a request is made to describe a partial activation range by one of the reference predefined activation levels.

[0047] According to one or more embodiments, at least one memory and a computer program are further configured to, together with at least one processor, cause the network device to execute: in response to the request, receiving information describing a status of an activation scope of a capability from another network device.

[0048] According to one or more embodiments, the information describing the capability includes one or more of the following:

[0049] e. One or more objects and / or object types, the network entity is configured to optimize or control one or more objects and / or object types;

[0050] f. One or more parameters of one or more objects and / or object types, the network entity being configured to control or optimize one or more parameters of one or more objects and / or object types to achieve a desired result;

[0051] g. one or more network metrics, the network entity is configured to seek to improve one or more network metrics;

[0052] h. One or more network characteristics associated with one or more objects and / or object types.

[0053] According to one or more embodiments, the request includes: one or more subsets of at least one of the items a to d describing the partial activation range.

[0054] According to one or more embodiments, the request includes at least one of a network context for describing a partial activation scope, and an artificial intelligence and / or machine learning context.

[0055] According to one or more embodiments, at least one memory and a computer program are also configured to, together with at least one processor, cause the network device to perform: determining whether data describing the performance of the capability after further activation at the requested activation range meets the performance criteria, and in the affirmative case, sending (208) a request to another network device to increase the previously requested activation range.

[0056] According to one or more embodiments, the at least one memory and the computer program are further configured to, with the at least one processor, cause the network device to execute: requesting to increase the activation range of the capability only if the full activation range has not been reached.

[0057] According to one or more embodiments, at least one memory and a computer program are further configured to, together with at least one processor, cause the network device to execute:

[0058] If the criteria are not met, a deactivation request is sent to another network device, the deactivation request identifying one of the following:

[0059] ○ An instruction to cancel the activation range request;

[0060] ○ Complete deactivation of capabilities;

[0061] o a reduced activation range to be achieved compared to the activation range where the criteria are not met;

[0062] o one or more parts of the activation range of the capability to be deactivated;

[0063] o A level among a plurality of predefined ordered activation levels that identifies a gradually increasing range of capability activation.

[0064] According to one or more embodiments, the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to perform: determining that network performance is degraded, and in response to the determination, requesting a partial reduction in the capability activation range.

[0065] According to one or more embodiments, at least one memory and a computer program are also configured to, together with at least one processor, cause the network device to perform: obtaining from another network device at least one of an accuracy measurement or data describing a benefit assessment performed by another device, wherein the accuracy measurement or the data describing the benefit assessment is a function of an output of capabilities for an envisioned partial activation range without applying the output to an operating environment, and sending a request for partial activation based on at least one of the accuracy measurement or the data describing the benefit assessment.

[0066] A fourth aspect relates to a network device, comprising: at least one processor, at least one memory, the at least one memory including computer program code, the network device including a network entity, the network entity including a capability for providing an output based on artificial intelligence and / or machine learning reasoning functions related to one or more network objects or object types, the at least one memory and the computer program code being configured to, together with the at least one processor, cause the network device to perform at least a partial activation of the capability.

[0067] According to one or more embodiments, the at least one memory and the computer program are further configured to, with the at least one processor, cause the network device to perform: making information describing capabilities available to other network devices.

[0068] According to one or more embodiments, at least one memory and a computer program are also configured to, together with at least one processor, cause the network device to perform: making information describing a plurality of predefined ordered activation levels available to other network devices, the plurality of predefined ordered activation levels identifying gradually increasing capability activation ranges within an available range.

[0069] According to one or more embodiments, at least one memory and a computer program are also configured to, together with at least one processor, cause the network device to perform: making at least one of an accuracy measurement or data describing a benefit assessment performed by another device available to other network devices, wherein the accuracy measurement or data describing the benefit assessment is a function of an output for a capability of a partial activation range envisioned without applying the output to an operating environment.

[0070] According to one or more embodiments, the information about the activated available range includes one or more of the following:

[0071] e. One or more objects and / or object types, the network entity is configured to optimize or control one or more objects and / or object types;

[0072] f. One or more parameters of one or more objects and / or object types, the network entity being configured to control or optimize one or more parameters of one or more objects and / or object types to achieve a desired result;

[0073] g. one or more network metrics, the network entity is configured to seek to improve one or more network metrics;

[0074] h. One or more network characteristics associated with one or more objects and / or object types.

[0075] According to one or more embodiments, partial activation of a capability is performed in response to a partial activation request from another network device.

[0076] According to one or more embodiments, at least one memory and a computer program are further configured to, together with at least one processor, cause the network device to execute:

[0077] After activating the capability at the requested scope, status information describing the activation status of the capability with respect to the requested scope is sent to other network devices.

[0078] According to one or more embodiments, at least one memory and a computer program are also configured together with at least one processor to cause the network device to execute:

[0079] receiving in the request at least one of a network context, and an artificial intelligence and / or machine learning context; and

[0080] Responsive to the context, the scope of activation of the capability is limited.

[0081] According to one or more embodiments, the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to execute: determining that network performance is degraded, and partially reducing a capability activation range in response to the determination.

[0082] According to one or more embodiments, activation of a capability includes applying an output of the capability to an operating network environment.

[0083] A fifth aspect relates to a computer readable medium including program instructions stored thereon, the program instructions being configured to perform at least the following:

[0084] obtaining information describing capabilities of a network entity of the second network device, the network entity comprising: capabilities for providing outputs based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types;

[0085] A request for partial activation of the capability is sent to the second network device.

[0086] A sixth aspect relates to a computer-readable medium including program instructions stored thereon for performing at least the following in a network device including a network entity, the network entity including: a capability for providing an output based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types;

[0087] Partial activation of capabilities.

[0088] According to one or more embodiments, the computer readable medium may be non-transitory. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate exemplary embodiments together with the general description given above and the detailed description given below.

[0090] Figure 1 is a block diagram of a system according to one or more exemplary embodiments.

[0091] Figure 2 is a flow chart of a method performed at an AI / ML management service user according to one or more exemplary embodiments.

[0092] Figure 3a and Figure 3b is a flow chart of a method performed at an AI / ML management service producer according to one or more exemplary embodiments.

[0093] Figure 4 is a message sequence diagram between an AI / ML management service user and an AI / ML management service producer according to one or more exemplary embodiments.

[0094] Figure 5 is an exemplary hardware structure of an apparatus according to one or more exemplary embodiments.

[0095] It should be noted that these drawings are intended to illustrate various aspects of the apparatus, methods, and structures used in the exemplary embodiments described herein. The use of similar or identical reference numerals in the various drawings is intended to indicate the presence of similar or identical elements or features. DETAILED DESCRIPTION

[0096] Detailed exemplary embodiments are disclosed herein. However, the specific structural and functional details disclosed herein are representative only for describing exemplary embodiments. Therefore, these embodiments are shown in the accompanying drawings by way of illustrative examples, and will be described in detail herein, so as to provide a thorough understanding of various aspects. However, it will be appreciated by those of ordinary skill in the art that exemplary embodiments can be variously modified and in other forms, and can be practiced without all the specific details. In addition, systems and processes can be shown in block diagrams so as not to obscure the exemplary embodiments in unnecessary details. In other cases, well-known processes, structures, and techniques can be shown without unnecessary details, so as to avoid obscuring the exemplary embodiments.

[0097] Exemplary embodiments provide apparatus, methods, systems, and computer programs for gradually activating artificial intelligence (AI) and / or machine learning (ML) capabilities.

[0098] The terms artificial intelligence and machine learning (AI / ML) refer to software-implemented approaches based on mathematical algorithms or models that provide reasoning capabilities.

[0099] Such models are typically mathematical algorithms that are trained with information and replicate the decisions that an expert would make when provided with the same information. According to some embodiments, the AI / ML capability may also provide data analysis. For example, an AI / ML training function associated with a model acquires data, runs the data through an AI / ML model, and derives the associated losses, and adjusts the parameterization of the AI / ML model based on the calculated losses. Training methods may include supervised learning, unsupervised learning, and reinforcement learning, and training may be performed offline or continuously. The inference function may be one of a plurality of known categories, such as regression-based, clustering or association-based, reward-based behavior, and an appropriate training method is applied.

[0100] Exemplary applications of AI and / or ML include, but are not limited to: speech recognition, image processing / computer vision, natural language processing, information retrieval, personalized recommendations, robotics, data analytics including predictive and prescriptive analytics, use cases for design and / or optimization and / or configuration and / or control and / or management of communication systems and / or networks.

[0101] Exemplary use cases may be without limitation:

[0102] - Use cases related to the physical layer of communication networks, such as modulation, coding, decoding, signal detection, channel estimation, prediction, compression, interference mitigation;

[0103] - Use cases related to the media access control layer of communication networks, such as multipath access and resource allocation (e.g. power control, scheduling, spectrum management);

[0104] -Channel modeling;

[0105] - Network optimization;

[0106] -Cell capacity estimation in cellular networks;

[0107] -routing;

[0108] - Resource management;

[0109] -Data traffic management;

[0110] - Security and anomaly detection;

[0111] -Root cause analysis;

[0112] -Transmission protocol design and optimization;

[0113] - User / network / application behavior analysis / prediction;

[0114] - transport layer congestion control;

[0115] -User experience modeling and optimization;

[0116] -User mobility and location management;

[0117] - Network slicing, network virtualization, and software-defined networking;

[0118] - Nonlinear impairment compensation in optical networks (such as visible light communication, fiber optic communication, and fiber-wireless converged networks); and

[0119] - Transmission quality estimation and optical performance detection in optical networks.

[0120] The term AI / ML entity refers to any network entity that contains one or more AI and / or ML capabilities.

[0121] Exemplary network entities include, but are not limited to:

[0122] - Radio access network entities, such as base stations (e.g., cellular base stations, such as eNodeBs in LTE and LTE-Advanced networks and gNodeBs used in 5G networks, and femtocells used in homes or business centres);

[0123] -Relay station;

[0124] - Control stations (e.g. radio network controllers, base station controllers, network switching subsystems);

[0125] - Access points in local area networks or ad hoc networks;

[0126] - Network management and radio access network entities;

[0127] - Network management entities (such as operations, administration and management (OAM) entities);

[0128] -Network automation system;

[0129] - Distributed analysis entities, such as autonomous systems (D-SONs);

[0130] - Network functions (e.g., Network Data Analysis Function NWDAF defined in current 3GPP standards);

[0131] -User Equipment (UE).

[0132] AI / ML Management Service (MnS) producers, such as management functions within a network or network entity, apply AI / ML to accomplish specific tasks. AI / ML Management Service (MnS) is considered to include one or more AI / ML entities, each with specific AI / ML capabilities.

[0133] According to one or more embodiments, management services are a set of capabilities provided for managing and coordinating networks and services. An entity that produces MnS is called a MnS producer. An entity that consumes MnS is called a MnS consumer. MnS provided by a MnS producer can be consumed by any entity with proper authorization and authentication.

[0134] The AI / ML entity provides capabilities to the AI / ML MnS consumer through the AI / ML MnS producer. Exemplary outputs of AI / ML entity capabilities include decision or data analysis. Exemplary analysis and / or decision functions include, but are not limited to: coverage analysis, coverage problem analysis, switching problem analysis, fault detection, interference detection, coverage optimization, capacity optimization, switching optimization, interference mitigation, energy saving optimization, etc. The AI / ML MnS producer is a function that can be implemented by software and / or hardware by any appropriate network entity (e.g., a network management function, an automation function, an analysis function, or a network function similar to a gNB, or a cell), which is configured to provide an interface for AI / ML capability exposure on behalf of one or more AI / ML entities. The AI / ML MnS consumer is a function that can be implemented by any network entity. According to one or more embodiments, the MnS consumer is required to have appropriate authorization and authentication.

[0135] According to one or more embodiments, the AI / ML MnS Producer is implemented by software executed by a network entity (or possibly multiple entities), and the AI / ML entity is part of this software.

[0136] According to one or more embodiments, the decision-making capabilities of an AI / ML entity are described as a ((multiple) objects, parameters, metrics) triple, where the entries indicate, respectively: objects or object types that the AI / ML entity can optimize or control, configuration parameters of the objects or object types that the AI / ML entity optimizes or controls to achieve desired results, and network metrics that the AI / ML entity optimizes through its actions.

[0137] The decision may take the form of a recommendation or a prediction. For example, the decision output by the AI / ML entity may be a recommendation based on the AI / ML model on whether to perform a handover of the UE to a new cell, or a prediction of the next target cell for the UE. The AI / ML MnS producer takes action on the network based on the output of the AI / ML entity.

[0138] When the output of the AI / ML entity is a decision, the AI / ML MnS producer executes the decision. For example, if the output of the AI / ML entity is antenna tilt, the AI / ML MnS producer will adjust the antenna accordingly, or if the output is a decision to switch to a new cell, the AI / ML MnS producer initiates the switch. When the output of the AI / ML entity consists of data analysis, the AI / ML MnS producer also takes actions that affect the network. For example, if the provided analysis data is related to the network function (NF) load, the NF can be scaled based on the provided analysis data, such as scaling if the NF is underloaded, or amplifying if the NF is overloaded.

[0139] According to one or more embodiments, the data analysis capabilities of the AI / ML entity are described as a tuple (object, feature), where the entries indicate, respectively: an object or object type for which the AI / ML entity can analyze, and a network feature (related to the object or object type) for which the AI / ML entity generates analysis. As a non-limiting example, the load of a single network function and / or object, or the load of a group of network functions and / or objects, or the load of an object type is a feature that can generate analysis. For example, an AI / ML capability for load of object type "cell" will provide analysis of the load experienced by any cell in the network.

[0140] The AI / ML entity may receive input data from a host device or from another network entity (e.g., a network device such as a sensor, data generator, etc.). The AI / ML entity may generate output data used by the host device or another network entity. The host device may be located within a communication network so that the AI / ML entity can access the required input data and can output data about the network entity.

[0141] In one or more embodiments, the proposed system provides an AI / ML capability discovery process implemented between an AI / ML MnS consumer and an AI / ML MnS producer, or between one or more AI / ML entities hosted by an AI / ML MnS producer. To this end, the AI / ML MnS producer exposes capability information, which can be received by the AI / ML MnS consumer in response to a request sent from the AI / ML MnS consumer to one or more AI / ML MnS producers. The capability information can also be received based on a subscription scheme, in which the AI / ML MnS consumer sends at least one request to receive capability information to one or more AI / ML MnS producers, and receives capability information matching the request when a new AI / ML entity is available in these producers.

[0142] The capability information may be exposed on an interface (e.g., an open interface) between an AI / ML MnS consumer and an AI / ML MnS producer. The interface may be used for various operations, including at least one of the following operations:

[0143] - Send a request for capability information from the AI / ML MnS consumer to the AI / ML MnS producer;

[0144] - Receiving capability information from AI / ML MnS producers by AI / ML MnS consumers;

[0145] - Sending configuration data of AI / ML entities from AI / ML MnS consumers to AI / MLMnS producers, targeting entities hosted by the latter.

[0146] The capability information exposed for available AI / ML capabilities may include various types of information, as described below.

[0147] The capability information may include a description of the functions implemented by the relevant AI / ML entity or the AI / ML model. The description may include at least one of text (such as "predicting QoS level within a QoS range"), keywords, identifiers, functions within a list of predefined functions, etc.

[0148] The capability information of the AI / ML entity may include information indicating the type of function performed by the AI / ML entity for the communication network. The type may be at least one of: an optimization function, a control function, an analysis function, and an AI / ML coordination function. Other relevant types may also be used. The type of function may be identified by a name, an identifier in an identifier list, or using another type of identification method.

[0149] The capability information of an AI / ML entity may include information representing an entity to which the functions implemented by the AI / ML entity are applied. These entities may be one or more objects or one or more object types (such as "UE", "cell", etc.). An object may be used as input or output of an AI / ML entity. For example, an AI / ML entity may provide a target cell to be switched as an output, in which case the object "cell" is part of the output generated by the AI / ML entity. An object may correspond to any entity in a communication network: a UE, a network entity, a functional unit, a virtual function, a database, or other similar / related objects. An object or object type may be identified by name, identifier, or using other types of identification methods.

[0150] The capability information may include information representing at least one configuration parameter for an object type for which an AI / ML capability or function provided by the AI / ML entity is performed. The configuration parameter of an object or object type may be any parameter that can be used to configure the relevant object. For example, a cell needs to have antenna tilt to determine where the antenna should be pointed to maximize coverage and minimize interference. The antenna tilt in this example is a configuration parameter for the object cell. The configuration parameter of an object or object type is an example of a configuration parameter of an AI / ML entity.

[0151] Capability information may include information representing configuration parameters for an AI / ML model implemented by an AI / ML entity. Configuration parameters (meta parameters, input parameters, etc.) for an AI / ML model may be any parameters that can be used to configure the associated model. For example, an AI / ML model may have inputs that the model uses to drive decisions, such as a model that optimizes coverage may use antenna tilt as input data. In this case, antenna tilt is a configuration parameter of the model. Configuration parameters for objects or object types are examples of configuration parameters for AI / ML entities.

[0152] The capability information may include information representing goals and / or constraints for performing the capability or function, wherein the goals and / or constraints are defined based on at least one network metric. Such goals and / or constraints may be encoded in words, or may be identified by a name or identifier within a list of goals, or by another type of identification method.

[0153] The capability information of an AI / ML entity may be encoded and reported in a variety of formats. The capability information of an AI / ML entity may be encoded using descriptive text and / or tuples and / or tables.

[0154] The capability information may represent a decision. For example, the capability information may be in the form of a triple (x, y, z) indicating:

[0155] -x: Object or type of object that the AI / ML entity can optimize or control;

[0156] -y: configuration parameters of object or object type x, which the AI / ML entity optimizes or controls to achieve the desired result;

[0157] -z: The network metric that the AI / ML entity optimizes through its actions.

[0158] The capability information may represent data analysis. For example, the capability information may be in the form of a tuple (x, z) indicating:

[0159] -x: Object or type of object that the AI / ML entity can analyze against;

[0160] -z: The network features (on object x) for which the AI / ML entity generates analysis.

[0161] According to one or more embodiments, the AI / ML MnS consumer is configured to request partial activation of a capability at the AI / ML entity. The partial activation includes activating a capability at an intermediate range between the full activation range and the full deactivation range.

[0162] Partial activation may be performed by selecting an activation scope of the AI / ML capabilities that is initially reduced compared to the full scope. The AI / ML entity then activates the capabilities of the initial scope. At a later stage, a new activation scope that adds and / or removes scope restrictions may be requested to reach a different scope that may partially overlap with the previous scope, completely exclude the previous scope, or include the previous scope and expand on it. The fact that a partial activation may be requested does not mean that a full scope activation may not be requested. In fact, one or more partial activation scopes may be requested consecutively, and then a full activation scope may be requested. Conversely, a full activation scope may be requested, and then a partial activation scope may be returned to at a later stage.

[0163] According to one or more embodiments, partial activation may also be performed by an AI / ML MnS producer instead of or in addition to an AI / ML MnS consumer. The AI / ML MnS producer may trigger partial activation based on one or more network performance criteria.

[0164] According to one or more embodiments, after activating a capability with an initially reduced range, the activation range is gradually expanded. For example, an initial partial range is checked to see if it meets one or more performance criteria. If one or more criteria are met, an expanded range compared to the initial range is defined, the AI / ML entity activates the capability over the new range, and the performance criteria check is performed again. The process is repeated until the performance criteria are not met for a given activation range, in which case the range is deactivated and rolled back to the previous range. At the end of the process, the AI / ML will be fully activated, partially activated, or not activated at all, depending on the range that last met the performance criteria.

[0165] It is difficult to predict and quantify the benefits of using AI / ML capabilities in a given operational network environment without implementing their results. Partial activation allows AI / ML MnS consumers to test a limited scope of capability activation before rolling out an increased or full scope of activation.

[0166] Furthermore, a full activation range may not bring the desired performance or even provide reduced performance. On the other hand, although it is generally believed that AI / ML capabilities unconditionally improve network performance, partial activation of a partial range may provide benefits that full activation may not necessarily provide. Furthermore, the partial activation range may be modified in various ways, including gradually increasing, for example, until the desired performance level or the highest possible performance level is reached. In other words, a more refined control over the activation of AI / ML capabilities is achieved.

[0167] Figure 1 is a block diagram of a system 100 according to one or more exemplary embodiments. The system comprises an AI / ML MnS consumer 101 and an AI / ML MnS producer 102. The latter comprises an AI / ML entity 103, which itself comprises AI / ML capabilities implemented, for example, using an AI / ML model. The AI / ML MnS consumer may request partial (in particular gradual) activation or deactivation of one or more AI / ML capabilities of the AI / ML entity by sending an appropriate request 105 to the AI / ML MnS producer. The producer sends a response 106 to the request 105 to the consumer to confirm the status.

[0168] An example request for partial activation includes:

[0169] 1. The identity of the AI / ML entity to which the request applies.

[0170] 2. Identification of one or more AI / ML capabilities of the AI / ML entity.

[0171] 3. Information identifying the given activation scope of the AI / ML capability.

[0172] The information identifying the activation scope may be described in various ways: an exhaustive list of individual items defining a capability, higher level characteristics (eg, type), one or more identifiers of a predefined set of items, group and / or type identifiers.

[0173] Detailed example request for activation (such as Figure 1 The request 105 in is as follows:

[0174] 1. Identification of the AI / ML entity to which the request applies;

[0175] 2. Information identifying at least one AI / ML capability of the AI / ML entity to be activated. According to one or more embodiments, for each AI / ML capability, this may include:

[0176] One or more of the following:

[0177] a. A list of one or more network objects and / or one or more network object types that the AI / ML entity can optimize or control.

[0178] For example, the network object may be a specific cell and the object type may be "cell".

[0179] b. A list of one or more parameters of the above objects and / or object types that the AI / ML entity controls or optimizes to achieve the desired results.

[0180] Example parameters may include antenna tilt, transmission power.

[0181] c. A list of one or more network metrics that the AI / ML entity seeks to improve through its operations.

[0182] Example parameters include throughput parameters.

[0183] d. A list of one or more network characteristics associated with one or more objects and / or object types for which the AI / ML entity generates analysis.

[0184] 3. Information defining the activation scope of the AI / ML capability. According to one or more embodiments,

[0185] This can include zero, one, or more of the following:

[0186] Context

[0187] a. Information about the network context

[0188] Network context is data that describes the state of the network when AI / ML capabilities are activated.

[0189] For example, a network context specifies a network technology, geographic, or temporal context.

[0190] For example, the context of the first scope may be “all gNBs and potentially additional RATs are operating”, while the context of the second scope may be “some gNBs or other RATs are experiencing failures or are shut down for energy saving”.

[0191] b. Information about the AI / ML context in which the AI / ML capabilities are to be activated. An activation context is defined. It is a subset of the expected runtime context (used to activate the initial AI / ML capabilities in an iteration) or a subset of the runtime context used to gradually activate additional AI / ML capabilities later in an iteration. The expected runtime context is the context (i.e., specific conditions) in which the AI / ML model is expected to be applied, while the runtime context is the context in which the model is being applied. The AI / ML context can be defined by the statistical characteristics of the data used by the AI / ML entity during training or inference.

[0192] Subrange

[0193] Defines sub-scope information for the items in Section 2 above, such as one or more of the following:

[0194] - Object sub-scope: identifies the subset of objects for which AI / ML capabilities should be activated (from 2a);

[0195] - Control parameter sub-range: identifies the subset of parameters for which the AI / ML capability should be activated (from 2a);

[0196] - Metric sub-scope: Identifies the subset of parameters for which AI / ML capabilities should be activated (from 2c).

[0197] The above is an exemplary request adapted to a specific context. For example, section 2 may contain only information related to a decision triple or a data analysis tuple. For example, for a decision type output, the request will not contain item 2d.

[0198] Other request formats and contents may be used. For example, the request may include additional activation scope definition information beyond that specified in the exemplary request. Such additional scope definition information may include one or more of the following: timing (start time, duration), geographic location information of the network object or type for which the capability is to be activated, information describing the network slice, UE category, PDU session, etc. Such information may then be used, for example, to further narrow the activation scope.

[0199] According to one or more embodiments, the request may explicitly or implicitly indicate full scope activation. For example, by default, the full scope is activated in the absence of any context or sub-scope.

[0200] According to one or more embodiments, multiple ordered activation levels or steps are defined. The activation level predefines: one or more AI / ML capabilities, and the associated gradually increasing activation ranges ordered from the smallest range to the largest range. The activation level is disclosed by the AI / ML MnS producer in the same way as the AI / ML capability information. The AI / ML MnS consumer activates the level through a request involving the activation level without having to specify any further details, such as sub-ranges. Such a level may also be referred to as an "abstract" activation level. When an activation level includes a set of AI / ML capabilities, these capabilities can be grouped according to common characteristics (e.g., complexity). For example, an AI / ML MnS producer may define three abstract activation levels, "low", "medium", and "high / full", and describe their corresponding characteristics, and disclose such information. This information may include AI / ML capability information and ranges related to the activation level, for example, which AI / ML capability should be activated under which network and / or AI / ML context if the AI / ML MnS consumer requests a given activation level. A low activation level / step may include only a single AI / ML capability of low complexity activated on a single object (e.g., a single cell). A medium activation level may include multiple AI / ML capabilities to be activated on a limited set of cells, while a high / full activation may mean full activation of all AI / ML capabilities of the AI / ML entity over the entire available range.

[0201] According to one or more embodiments, an activation level is defined for activation of a capability at the full range of availability. Using this activation level in a request for activation allows an AI / ML MnS consumer to request full activation over the entire range of availability.

[0202] In response to the activation request, the AI / ML MnS producer provides the AI / ML MnS consumer with an activation status. According to one or more embodiments, the status specifies the AI / ML entities involved and in which scope the activation has been successfully performed for each of these entities.

[0203] The deactivation request may contain parameters that are generally similar to those in the activation request. According to one or more embodiments, the deactivation request identifies which part of the activation range of the capability is to be deactivated. For example, if the current activation range of the capability covers cells with identifiers "cell ID1, cell ID2, cell ID3", and the deactivation request contains "cell ID3" in the information defining the deactivation range, the AI / ML MnS producer will reduce the range to "cell ID1, cell ID2". According to an embodiment, the deactivation request identifies the activation range to be implemented and is reduced compared to the current activation range. According to one or more embodiments, the deactivation request may also include a "revocation" of a previously sent activation request. More ways of encoding the deactivation range in the deactivation request may be defined.

[0204] According to one or more embodiments, an AI / ML MnS producer provides an AI / ML capability evaluation message to an AI / ML MnS consumer, such as a message about the characteristics of the AI / ML capability in an operating environment. Such an evaluation message can be provided to the AI / ML MnS consumer in the same manner as the capability information. The AI / ML entity performs an evaluation of the AI / ML capability by executing the AI / ML capability without taking actions based on the resulting solution (i.e., without implementing a decision output) and measuring the accuracy of the AI / ML solution in the operating environment compared to a non-AI / ML solution or determining the benefits and impact of the AI / ML capability in a specific situation. The AI / ML MnS consumer can obtain the measured accuracy or benefit of AI / ML from the AI / ML entity. Based on this information, the AI / ML MnS consumer can decide whether to request activation of the AI / ML capability.

[0205] More specifically, the disclosure of such evaluation information may be (a) requested by an AI / ML MnS consumer and / or (b) initiated by an AI / ML MnS producer. An AI / ML MnS producer may run one or more AI / ML capabilities by feeding operational data as input data to an AI / ML entity, but does not apply the AI / ML capability output to the operational network, e.g., does not implement decisions based on AI / ML capability predictions. Such information about AI / ML operational characteristics may take the form of an accuracy indicator of the AI / ML solution when applied to the operational data. In some use cases, the evaluation message may be expressed in terms of the benefit / impact of applying the AI / ML capability output in the operational environment. For example, a capability of an AI / ML entity may be to predict virtual network function (VNF) load and provide recommendations to scale it up or down. At 1 p.m., an operator examines the output of the capability, which predicts the VNF load one hour later (e.g., 2 p.m.), but the recommended scaling will not be performed. At 2 p.m., the operator determines the actual VNF load and compares it with that predicted by the AI / ML entity. Based on the comparison, the accuracy of the prediction may be calculated. However, the evaluation is only applicable to certain metrics, such as the cost / number of virtual resources in this case. For other metrics, such as the corresponding quality of experience (QoE) of users affected by the changes required for following recommendations / decisions provided by the AI / ML capability output, it may be necessary to apply the recommended changes in the network environment (e.g., scaling down certain NFs) to evaluate the performance.

[0206] like Figure 1 As shown in the relevant part of the above description, deactivation of an AI / ML capability may be requested by an AI / ML MnS consumer. In response to the deactivation request, the AI / ML MnS producer deactivates the specified AI / ML capability (or capabilities) and provides status information to the AI / ML MnS consumer in return.

[0207] Deactivation may be complete, or deactivation may be partial, for example by identifying specific items to be excluded from the activation scope of the current application, or as a revocation of an AI / ML capability activated by a previous activation request. The above-mentioned abstract activation levels may also be used in a request to deactivate an AI / ML capability.

[0208] In response to the deactivation request, the AI / ML MnS producer provides the AI / ML MnS consumer with a deactivation status. According to one or more embodiments, the status specifies the AI / ML entities involved and, for each of these entities, within which scope the deactivation was successfully performed.

[0209] Figure 2is a flow chart of a method performed at an AI / ML MnS consumer according to one or more embodiments. In step 201, the AI / ML MnS consumer first obtains AI / ML capability information describing the capabilities of the AI / ML entity. Note that this step is optional if the AI / ML MnS consumer has already disposed of the information, e.g., it previously obtained the information. The AI / ML MnS consumer then selects (202) a partial scope of activation of the AI / ML capability and sends (203) an activation request characterizing the partial scope to the AI / ML MnS producer that hosts the AI / ML entity.

[0210] According to one or more embodiments, the scope of activation to be implemented may be defined in the request, for example but not limited to, based on a subset of information describing previously acquired capabilities, and / or at least one of a network context, and / or an AI / ML context. The network and AI / ML context will typically be provided by the AI / ML MnS consumer in the activation request (203).

[0211] Optionally, the request identifies one of a plurality of abstract activation levels defining a gradual capability activation step as previously described. Once the AI / ML entity has activated the capability at the requested level, the AI / ML MnS consumer receives (204) a status report from the AI / ML MnS producer regarding the scope over which the capability has been effectively activated. The AI / ML MnS consumer then obtains information describing the performance over the selected scope after activation (e.g., by measuring network performance) and determines, based on one or more criteria, whether the performance is satisfactory (205). If this is the case (and unless the full scope of activation has been reached (208)), the AI / ML MnS consumer increases the scope of activation (209) and sends another request (iteration at 203), followed by another performance check (205). Thus, the method gradually and iteratively increases the scope of activation. If the performance is not satisfactory, the AI / ML MnS consumer instead determines a scope to deactivate (206), issuing a deactivation request (207). The deactivation request identifies a reduced activation scope compared to the scope that resulted in the unsatisfactory performance. For example, the reduced activation scope may be identified in the same manner as in the activation request.The AI / ML MnS consumer then receives a status report on the deactivation request from the AI / ML MnS producer (210).

[0212] Figure 3a301 is a flow chart of a first method performed at an AI / ML MnS producer according to one or more embodiments. At 301, the AI / ML MnS producer provides an interface for exposing information describing capabilities of an AI / ML entity to one or more AI / ML MnS consumers to obtain information about the scope in which the capabilities can be activated. According to one or more embodiments, other information can be used to replace or supplement the information exposed by the AI / ML MnS producer to define a partial activation scope. For example, a context such as a network context or an AI / ML context can be used.

[0213] Optionally, the information includes multiple abstract activation levels defining progressive capability activation steps as described above. At 302, the AI / ML MnS producer receives a request from the AI / ML MnS consumer identifying a partial activation scope for a given capability. At 303, the AI / ML MnS producer activates the AI / ML capability of the AI / ML entity at the requested partial scope. At 304, the AI / ML MnS producer then sends a status report to the AI / ML MnS consumer that issued the request. The report provides information about the scope of the effective activated AI / ML capability.

[0214] Figure 3b Flowchart of a second method performed at an AI / ML MnS producer according to one or more embodiments. At 311, the AI / ML MnS producer receives a request from an AI / ML MnS consumer identifying a partial deactivation scope of a given capability. At 311, the AI / ML MnS producer deactivates the AI / ML capability of the AI / ML entity according to the request. At 312, the AI / ML MnS producer sends a status report to the requesting AI / ML MnS consumer. The report provides information about the scope to which the AI / ML capability has been effectively deactivated, such as "predefined abstraction level 3 has been successfully deactivated."

[0215] According to the first exemplary use case, the AI / ML MnS producer is a user equipment (UE) running an AI / ML model for predicting UE location / trajectory. The AI / ML MnS consumer is a base station gNB, which uses the predicted UE information to trigger handover. The AI / ML MnS consumer (i.e., gNB) gradually activates / deactivates AI / ML capabilities at a single UE or at multiple selected UEs.

[0216] According to a second exemplary use case, the AI / ML MnS producer is a Capacity and Coverage Optimization (COO) Self-Organizing Network (SON) function that runs an AI / ML model to adjust control parameters in order to achieve efficient network resource usage and optimal end-user experience. The AI / ML MnS consumer is an operator's Operator / Network Management System (OAM) that aims to improve network performance and UE Quality of Service (QoS). The AI / ML MnS consumer gradually activates / deactivates CCO capabilities in a cell or at selected cells among all available cells and may initially be used for selected contexts. The scope is subsequently expanded.

[0217] In the above example, the CCO (Capacity and Coverage Optimization) SON (Self-Organizing Capacity) AI / ML capabilities are exposed by specifying the following tuple:

[0218] (a) Control parameters

[0219] The control parameters may include any one or more control parameters of the CCO (eg, configured maximum transmit power, configured maximum EIRP, coverage shape, digital tilt, digital azimuth).

[0220] (b) Measurement

[0221] The metrics may include any or multiple CCO related performance measurements (e.g., distribution of SS-RSRP per SSB, distribution of SS-RSRQ, distribution of number of active UEs per SSB, number of requested handover executions, number of failed handover executions, distribution of DL total PRB usage, DL PRBs used for data traffic, total DL available PRBs, DL PRBs used for data traffic, DL total available PRBs, UL PRBs used for data traffic, UL total available PRBs, average DL UE throughput in gNB, distribution of DL UE throughput in gNB, average UL UE throughput in gNB, distribution of UL UE throughput in gNB, average number of RRC connections, maximum number of RRC connections, number of PDU sessions requested to be setup, number of PDU sessions successfully setup, number of PDU sessions failed to setup).

[0222] (c) Object type: cell.

[0223] Using the same tuple in the activation request, the AI / ML MnS consumer (i.e. the operator's operator / network management system (OAM)) initially specifies in the request context information (e.g., geographic / temporal / technical context) and additional sub-range information, such as a sub-range limited to a single control parameter (e.g., antenna tilt (digital tilt)), and a subset of cells for which the AI / ML capability is to be activated. Once satisfactory performance is achieved, the sub-range is extended to include further control parameters, in this case such as transmit power, and also to include a wider geographic / temporal / technical context.

[0224] An exemplary request within the frame of the second use case is as follows:

[0225] -AIMLEntityID="CCO_ID1"

[0226] -AI / ML capabilities:

[0227] ○ Object: Cell ID1, Cell ID2, Cell ID3, Cell ID4, Cell ID5, Cell ID6, Object type: Cell.

[0228] ○Control parameters: antenna tilt, transmission power

[0229] - The scope to be activated:

[0230] ○ Network context: 5G RAT, FR1

[0231] ○ AI / ML context: Geographic area of ​​northern and southern suburbs of Munich, nighttime, low traffic.

[0232] ○ Target sub-range: Cell ID1, Cell ID2, Cell ID3

[0233] ○ Parameter sub-range: Antenna tilt

[0234] An exemplary status report sent by the AI / ML MnS producer after activating the scope in this request is as follows:

[0235] -AIMLEntity="CCO_ID1"

[0236] -Activation time: e.g. 22:00CET

[0237] - Status: Activatable

[0238] - Activation capabilities: Cell ID1, Cell ID2, Cell ID3, Antenna Tilt

[0239] - Range: 5G RAT FR1, Suburban North and Suburban South, 5% of maximum traffic in the selected areas.

[0240] After this step, the AI / ML MnS producer and / or consumer monitors the network performance and decides if further expansion of capacity and / or range is needed. For example, assuming that the performance of the initial range of cells with IDs 1, 2, 3 is satisfactory, this range can be gradually increased. The first expansion of the range can be performed by adding cells with IDs 4, 5, 6 and requested in the second request. The transmission power as a control parameter can be added at the same time as the activation range or in the third request. Other areas can be added to the activation range, for example, other areas such as Suburban East and Suburban West, and / or other times of the day, such as afternoon time, can be added to the activation range, which can be a second frequency range FR2.

[0241] In case the AI / ML MnS consumer requests partial or gradual activation using an abstract activation level / step (e.g. "low"). The AI / ML MnS producer may provide the status and description of the activated capabilities in subsequent steps. Based on the monitored network performance, the AI / ML MnS consumer may decide to increase the scope of activation by specifying the next level / step (e.g. "medium") and subsequently "high / full" in the activation request.

[0242] When monitoring network performance, an AI / ML MnS consumer or producer may identify a degradation in that performance. Based on such an event, an AI / ML MnS consumer may request to deactivate an AI / ML capability, either by requesting to deactivate all capabilities of the entire AI / ML entity, or by requesting to deactivate selected AI / ML capabilities, assuming that the entity provides multiple capabilities. Selective deactivation of capabilities may be useful if, during gradual activation, network performance degrades after certain capabilities have been activated to a certain extent. An AI / ML MnS consumer (or producer) may request to gradually deactivate (rollback) key AI / ML capabilities.

[0243] Figure 4The present invention is a message sequence diagram showing an AI / ML MnS consumer and an AI / ML MnS producer, and illustrating different requests and responses exchanged between the two functions, according to one or more exemplary embodiments. At 401, capability discovery is performed between an AI / ML MnS consumer and an AI / ML MnS producer. At 402, the AI / ML MnS producer optionally exposes an abstract activation level for acquisition by the AI / ML MnS consumer. At 403, the operational characteristics of the AI / ML capabilities of one or more AI / ML MnS entities of the AI / ML MnS producer are optionally exposed for acquisition by the AI / ML MnS consumer. These operational characteristics may include accuracy measurements and / or the benefits and impact of the AI / ML solution. At 404, the AI / ML MnS consumer sends a request for partial activation to the AI / ML MnS producer. At 405, the AI / ML MnS producer provides a status report for the partial activation of one or more capabilities identified in the request. At 406, the AI / ML MnS consumer sends a request for partial deactivation to the AI / ML MnS producer. At 407, the AI / ML MnS producer sends a status report for the deactivation of one or more capabilities identified in the request.

[0244] Regarding deactivation, Figure 4 The case where deactivation is initiated by the AI / ML MnS consumer (at 406) is illustrated. The partial deactivation request may contain parameters similar to those of the activation request. For example, a complete deactivation of an AI / ML entity may be requested by providing only an identifier of the AI / ML entity in the request.

[0245] Similar to activation, deactivation can also be performed based on the abstract activation levels / steps exposed by the AI / ML MnS producer. The deactivation request issued by the AI / ML MnS consumer then includes the level / step that needs to be deactivated by the AI / ML MnS producer. In response, the activation scope is set back to the immediately previous level of the smaller scope.

[0246] It will be understood by those skilled in the art that any functions, engines, block diagrams, flow charts, state transition diagrams, flowcharts, and / or data structures described herein identify conceptual views of illustrative circuit systems embodying the present disclosure. Similarly, it will be understood that any flow charts, flow diagrams, state transition diagrams, pseudocodes, etc., identify various processes that may be substantially represented in a computer-readable medium and thus executed by a computer or processing device, regardless of whether such a computer or processor is explicitly shown.

[0247] Although a flowchart may describe operations as a sequential process, many operations may be performed in parallel, concurrently, or simultaneously. Some operations may also be omitted, combined, or performed in a different order. A process may terminate when its operations are completed, but may also have additional steps not described in the drawings or description. A process may correspond to a method, function, procedure, subroutine, subprogram, etc. When a process corresponds to a function, its termination may correspond to a function call to the function or a main function.

[0248] Each function, engine, module, step described herein may be implemented in hardware, software, firmware, middleware, microcode, or any suitable combination thereof.

[0249] When implemented in software, firmware, middleware, or microcode, instructions for performing the necessary tasks may be stored in a computer-readable medium, which may be included or not included in a host device or system. Instructions may be sent on a computer-readable medium and loaded onto a host device or system. Instructions are configured so that the host device / system performs one or more functions disclosed herein. For example, as described above, according to one or more examples, at least one memory may include or store instructions, and at least one memory and instructions are configured to, together with at least one processor, cause the host device / system to perform one or more functions. In addition, processors, memories, and instructions serve as components for being provided by a host device / system or causing the host device / system to perform one or more functions disclosed herein.

[0250] The host device or system may be a general purpose computer and / or computing system, a special purpose computer and / or computing system, a programmable processing device and / or system, a machine, etc. The host device or system may be or include: a user device, a client device, a mobile phone, a laptop, a computer, a network element, a data server, a network resource controller, a network device, a router, a network management, a network node, a computer, a cloud server, a web server, an application server, a proxy server, etc.

[0251] Figure 5 An exemplary embodiment of an apparatus 500 is illustrated. The apparatus 500 may be an AI / ML consumer or an AI / ML producer as disclosed herein. The apparatus may be configured to host at least one AI / ML entity disclosed herein.

[0252] like Figure 5As shown in the exemplary embodiment, the device 500 may include at least one processor 510 and at least one memory 520. The device 500 may include one or more communication interfaces 540 (for example, a network interface for accessing a wired / wireless network, including an Ethernet interface, a WIFI interface, a USB interface, etc.), one or more communication interfaces 540 are connected to the processor and are configured to communicate via (multiple) wired / non-wired communication links. The device 500 may include other associated hardware that communicates with the processor, such as a user interface 530 (such as a keyboard, a mouse, a display screen, etc.). The device 500 may also include one or more media drives 550 for reading computer-readable storage media (such as a digital storage disk 560 (CD-ROM, DVD, Blu-ray, etc.), a USB key 580, etc.). The processor 510 is operably connected to each of the other components 530, 540, 550 to control their operation.

[0253] The memory 520 may include random access memory (RAM), cache memory, non-volatile memory, backup memory (e.g., programmable memory or flash memory), read-only memory (ROM), hard disk drive (HDD), solid state drive (SSD), or any combination thereof. The ROM of the memory 520 may be configured to store one or more computer program codes of an operating system and / or one or more software applications of the storage device 500. The RAM of the memory 520 may be used by the memory 510 for temporary storage of data.

[0254] The processor 510 can be configured to store, read, load, execute and / or otherwise process instructions 570 stored in computer-readable storage media 560, 580 and / or memory 520, which, when executed by the processor, causes the device 500 to perform one or more or all steps of the method described herein for the device 500 involved.

[0255] Instructions may correspond to computer program instructions, computer program codes, and may also include one or more code segments. A code segment may represent a program, function, subprogram, program, routine, subroutine, module, software package, category, or any combination of instructions, data structures, or program statements. A code segment may be coupled to another code segment or hardware circuit by transmitting and / or receiving information, data, independent variables, parameters, or memory contents. Information, independent variables, parameters, data, etc. Information, independent variables, parameters, data, etc. may be transmitted, forwarded, or sent via any suitable technology including memory sharing, message passing, token passing, network transmission, etc.

[0256] When provided by a processor, the function may be provided by a single dedicated processor, a single shared processor, or by a plurality of individual processors, some of which may be shared. The term "processor" should not be interpreted as referring specifically to hardware capable of executing software, and may implicitly include one or more processing circuits, whether programmable or not. A processor or similar processing circuit may correspond to a digital signal processor (DSP), a network processor, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single chip system (SOC), a central processing unit (CPU), an arithmetic logic unit (ALU), a programmable logic unit (PLU), a processing core, programmable logic, a microprocessor, a controller, a microcontroller, a microcomputer, any device capable of responding and / or executing instructions in a defined manner and / or according to a defined logic. Other conventional or customized hardware may also be included. A processor or processing circuit may be configured to execute instructions, which are suitable for causing a host device or system to perform one or more functions disclosed herein for the host device or system involved.

[0257] Computer readable media or computer readable storage media can be any tangible storage medium suitable for storing data that can be read by a computer or processor. Generally speaking, computer readable media can be any storage medium that can store and / or contain and / or carry instructions and / or data. Computer readable media can be portable or fixed storage media. Computer readable media can include one or more storage devices, such as permanent mass storage devices, magnetic storage media, optical storage media, digital storage disks (CD-ROM, DVD, Blu-ray, etc.), USB keys or dongles or peripheral devices, memory for storing computer or processor readable instructions.

[0258] A memory suitable for storing computer or processor readable instructions may be, for example, a read-only memory (ROM), a permanent mass storage device (e.g., a disk drive, a hard disk drive (HDD), a solid-state drive (SSD), a memory card, a core memory, a flash memory), or any combination thereof.

[0259] In this specification, the term "component configured to perform one or more functions" or "component for performing one or more functions" may correspond to one or more functional blocks including circuit systems, which are suitable for performing or configured to perform (multiple) the functions involved. The block can perform the function by itself, or cooperate and / or communicate with one or more other blocks to perform the function. The "component" may correspond to or be implemented as "one or more modules", "one or more devices", "one or more units", etc. The component may include at least one processor and at least one memory including computer program code, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, enable the device or system to perform the functions involved.

[0260] As used in this application, the term "circuitry" may refer to one or more or all of the following:

[0261] (a) implemented only in hardware circuits (e.g., implemented only in analog and / or digital circuit systems); and

[0262] (b) a combination of hardware circuitry and software, such as (where applicable): (i) a combination of analog and / or digital hardware circuitry and software / firmware, and (ii) any portion of hardware processor(s) and software (including digital signal processor(s), software, and memory(s) that work together to enable a device such as a mobile phone or server to perform a variety of functions; and

[0263] (c) Hardware circuit(s) and / or processor(s), such as microprocessor(s) or portions of microprocessor(s), that require software (e.g., firmware) in order to operate, but in which case the software may not be present when the software is not required for operation.

[0264] This definition of circuitry applies to all uses of the term in this application, including in any claims. As a further example, as used in this application, the term circuitry also covers only hardware circuits or processors (or multiple processors) or parts of hardware circuits or processors and their (or their) accompanying software and / or firmware implementations. For example and if applicable to a particular claim element, the term circuitry also covers an integrated circuit or any other computing device or network device of a network element or network node.

[0265] The term circuit system may cover digital signal processor (DSP) hardware, network processor, application specific integrated circuit (ASIC), field programmable gate array (FPGA), etc. The circuit system may be or include, for example, hardware, programmable logic, a programmable processor (e.g., processor, control unit / entity, controller) that executes software or firmware and / or any combination thereof to execute instructions or software and control the transmission and reception of signals, and a memory that stores data and / or instructions.

[0266] The circuit system may also make decisions or determinations, generate frames, packets, or messages for transmission, decode received frames or messages for further processing, and other tasks or functions described herein. The circuit system may control the transmission of signals or messages over a radio network, and may control the reception of signals or messages via one or more communication networks, etc.

[0267] Although the terms first, second, etc. can be used here to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, without departing from the scope of the present disclosure, the first element can be referred to as the second element, and similarly, the second element can be referred to as the first element. As used herein, the term "and / or" includes any and all combinations of one or more associated listed items.

[0268] The terms used herein are only used for the purpose of describing specific embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" used in this Agreement are also intended to include plural forms. It should be further understood that the terms "include", "comprise", "have", and / or "contain" when used in this article specify the presence of the features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations of the foregoing.

[0269] Although various aspects of the present disclosure have been specifically shown and described with reference to the above embodiments, it will be understood by those skilled in the art that various additional embodiments may be provided by modifying the disclosed machines, systems, and methods without departing from the disclosed scope. Such embodiments should be understood to fall within the scope of the present disclosure as determined based on the claims and any equivalents thereof.

[0270] Abbreviations

[0271] 5G Fifth Generation

[0272] CCO Capacity and Coverage Optimization

[0273] DL Downlink

[0274] MDAS Management Data Analysis Service

[0275] MnS Management Services

[0276] NWDAF network data analysis function

[0277] OAM Operations, Administration and Management

[0278] PDU Protocol Data Unit

[0279] PRB Physical Resource Block

[0280] QoE Quality of Experience

[0281] QoS Quality of Service

[0282] SON Self-Organizing Network

[0283] RRC Radio Resource Control

[0284] SSB Synchronous Signal Block

[0285] SS-RSRP Secondary synchronization signal reference signal power

[0286] SS-RSRQ Secondary synchronization signal reference signal reception quality

[0287] UL Uplink

[0288] Example

[0289] 5 Use cases, potential requirements, and possible solutions

[0290] 5.A AI / ML Activation

[0291] 5.A.1 Description

[0292] TR 28.908 defines AI / ML deployment as the process of enabling AI / ML functions to be available in an operating environment. After training, AI / ML-enabled functions can be deployed in a 3GPP system and subsequently activated. The current description of AI / ML entity activation assumes that users request and producers perform activation of already deployed AI / ML entities and that activation of AI / ML capabilities will unconditionally improve network performance. But this is not always the case.

[0293] 5.A.2 Use cases

[0294] 5.A.2.1 Enabling gradual activation of AI / ML capabilities

[0295] For a given AI / ML-enabled capability, first, it is difficult to “predict” and quantify the benefit of using the AI / ML capability in a given operational network context before the AI / ML capability is used. Second, testing the AI / ML capability using test data does not give a “full picture” of how the AI / ML model will impact the network once activated in an operational environment. For example, testing can provide insights into the accuracy of the AI / ML entity that can be expected once it is deployed and activated.

[0296] It is therefore necessary to ensure that AI / ML capabilities activated in an operational network will bring benefits and will not further exacerbate existing network performance issues. In addition, it is important to provide means for checking which specific AI / ML capabilities are beneficial to activate in a given operational network context. Taken together, these mean that it is important to ensure that AI / ML MnS consumers have more granular control over AI / ML capability activation and deactivation.

[0297] 5.A.3 Potential requirements

[0298] REQ-AI / MLUPDATE-1 The 3GPP management system shall have the capability to allow authorized consumers to partially or gradually activate the AI / ML capabilities of producers of AI / ML reasoning through an abstract set of activation steps.

[0299] REQ-AI / MLUPDATE-2 The 3GPP management system shall have the capability to allow an authorized consumer to activate the AI / ML capabilities of a producer of AI / ML reasoning for a specified sub-scope of the applicable expected temporal context of the producer of AI / ML reasoning.

[0300] REQ-AI / MLUPDATE-3 The 3GPP management system shall have the capability to allow authorized consumers to partially or gradually deactivate the AI / ML capabilities of producers of AI / ML reasoning through an abstract set of activation steps.

[0301] REQ-AI / MLUPDATE-4 The 3GPP management system shall have the capability to enable producers of AI / ML reasoning to notify authorized consumers that a new sub-range of AI / ML capabilities has been activated.

[0302] 5.A.4 Possible solutions Note: The above requirements may not be made in the 3GPP management system, but in the specific provider, as follows:

[0303] REQ-AI / MLUPDATE-1 A producer of AI / ML inferences SHOULD have the capability to allow authorized consumers to partially or incrementally activate the AI / ML capabilities of the producer of AI / ML inferences through a set of abstract activation steps.

[0304] REQ-AI / MLUPDATE-2 A producer of AI / ML reasoning shall have the capability to allow authorized consumers to activate the AI / ML capability of the producer of AI / ML reasoning within a specified sub-scope of the applicable intended temporal context of the producer of AI / ML reasoning.

[0305] REQ-AI / MLUPDATE-3 A producer of AI / ML reasoning SHOULD have the capability to allow authorized consumers to partially or gradually deactivate the producer's AI / ML reasoning through an abstract set of activation steps.

[0306] REQ-AI / MLUPDATE-4 Providers of AI / ML reasoning SHOULD have the capability to enable producers of AI / ML reasoning to notify authorized consumers that a new sub-scope of AI / ML capability has been activated.

Claims

1. A method performed by a first network device (101), the first network device comprising at least one processor (510) and at least one memory (520), the at least one memory (520) comprising a computer program code (570), the computer program code (570) when executed by the at least one processor causing the device to perform the method, the method comprising: obtaining (204, 401) information describing capabilities of a network entity (103) of a second network device (102), the network entity comprising: capabilities for providing outputs based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types; A request for partial activation of the capability is sent (203, 404) to the second network device.

2. The method according to claim 1, comprising: Information describing a plurality of predefined ordered activation levels is obtained, the plurality of predefined ordered activation levels identifying gradually increasing capability activation ranges; and wherein the request describes a partial activation range by reference to one of the predefined activation levels.

3. The method according to any one of claims 1 or 2, comprising: In response to the request, information describing a status of the activation scope of the capability is received (204, 405) from the second device.

4. The method according to any one of claims 1 to 3, wherein: The information describing the capability includes one or more of the following: a. one or more objects and / or object types, the network entity being configured to optimize or control the one or more objects and / or object types; b. one or more parameters of the one or more objects and / or object types, the network entity being configured to control or optimize the one or more parameters of the one or more objects and / or object types to achieve a desired result; c. one or more network metrics, the network entity being configured to seek to improve the one or more network metrics; d. One or more network characteristics associated with the one or more objects and / or object types.

5. The method of claim 4, wherein the request comprises: One or more subsets of at least one of the items a to d used to describe the partial activation range.

6. The method according to any one of claims 1 to 5, wherein: The request includes: a network context for describing the partial activation range, and at least one of an artificial intelligence and / or machine learning context.

7. The method according to any one of claims 1 to 6, comprising: It is determined (205) whether data describing the performance of the capability upon further activation at the requested activation scope meets performance criteria, and in the affirmative, a request is sent (208) to the second device to increase the activation scope previously requested.

8. The method according to claim 7, comprising: Only if (207) the full activation range has not yet been reached, a request is made to increase the activation range of the capability.

9. The method according to any one of claims 7 or 8, comprising: If the criterion is not met, a deactivation request is sent (206, 406) to the second device, the deactivation request identifying one of: ○ an instruction to cancel the activation request; o Complete deactivation of said capability; o a reduced activation range to be achieved compared to said activation range where said criterion is not met; o one or more active parts in the activation scope of the capability to be deactivated; ○ A level among the plurality of predefined ordered activation levels that identifies the gradually increasing capability activation range as described in claim 2.

10. The method according to any one of claims 1 to 9, comprising: A determination is made that network performance is degraded, and in response to the determination, a partial reduction in a capability activation range is requested.

11. The method according to any one of claims 1 to 10, comprising: Obtaining from the second network device at least one of an accuracy measurement or data describing a benefit assessment performed by the second device, wherein the accuracy measurement or data describing the benefit assessment is a function of an output of the capability for an envisioned partial activation range without applying the output to an operating network environment, and sending a request for the partial activation based on at least one of the accuracy measurement or data describing the benefit assessment.

12. A method performed by a second network device (102), the second network device comprising at least one processor (510) and at least one memory (520), the at least one memory (520) comprising a computer program code (570), the computer program code (570) when executed by the at least one processor, causing the device to perform the method, the second network device comprising a network entity, the network entity comprising a capability for providing an output based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types, the method comprising (303) partial activation of the capability.

13. The method according to claim 12, comprising (301) making information describing the capabilities available (401) to other network devices.

14. The method according to claim 13, wherein: The information describing the capability includes one or more of the following: a. one or more objects and / or object types, the network entity being configured to optimize or control the one or more objects and / or object types; b. one or more parameters of the one or more objects and / or object types, the network entity being configured to control or optimize the one or more parameters of the one or more objects and / or object types to achieve a desired result; c. one or more network metrics, the network entity being configured to seek to improve the one or more network metrics; d. One or more network characteristics associated with the one or more objects and / or object types.

15. The method according to any one of claims 13 or 14, comprising: Information describing a plurality of predefined ordered activation levels identifying increasing ranges of capability activation within an available range is made available to other network devices (402).

16. A method according to any one of claims 13 to 15, comprising: Making at least one of an accuracy measurement or data describing a benefit assessment performed by the second device available to other network devices (403), wherein the accuracy measurement or data describing the benefit assessment is a function of the output of the capability for variations of an envisioned activation range without applying the output to an operating environment.

17. The method according to any one of claims 12 to 16, wherein: Partial activation of the capability is performed in response to a request (302) for partial activation from a first network device.

18. The method according to claim 17, comprising: After activation of the capability at the requested scope, status information describing an activation status of the capability with respect to the requested scope is sent (304, 405) to the first device.

19. The method according to claim 18, comprising: receiving in the request at least one of a network context, and an artificial intelligence and / or machine learning context; as well as Responsive to the context, limiting the activation scope of the capability.

20. A method according to any one of claims 12 to 19, comprising: A determination is made that network performance is degraded, and the capability activation range is partially reduced in response to the determination.

21. The method of any one of claims 12 to 20, wherein activation of the capability comprises: The output of the capabilities is applied to an operational network environment.

22. A network device (101), comprising: At least one processor (510) and at least one memory (520), the at least one memory (520) comprising computer program code (570), the at least one memory and the computer program code being configured to, together with the at least one processor, cause the network device to at least perform: Obtaining (201) information describing capabilities of a network entity (103) of another network device (102), the network entity comprising: capabilities for providing outputs based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types; A request for partial activation of the capability is sent (203, 404) to the other network device.

23. The device of claim 22, wherein the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to perform: obtaining information describing a plurality of predefined ordered activation levels, the plurality of predefined ordered activation levels identifying gradually increasing capability activation ranges; and wherein the request describes the partial activation range by reference to one of the predefined activation levels.

24. A device according to any one of claims 22 or 23, wherein the at least one memory and the computer program are also configured to, together with the at least one processor, cause the network device to execute: in response to the request, receiving (204, 405) information describing the status of the activation scope of the capability from the other network device.

25. Apparatus according to any one of claims 22 to 24, wherein The information describing the capability includes one or more of the following: a. one or more objects and / or object types, the network entity being configured to optimize or control the one or more objects and / or object types; b. one or more parameters of the one or more objects and / or object types, the network entity being configured to control or optimize the one or more parameters of the one or more objects and / or object types to achieve a desired result; c. one or more network metrics, the network entity being configured to seek to improve the one or more network metrics; d. One or more network characteristics associated with the one or more objects and / or object types.

26. The apparatus of claim 25, wherein the request comprises: One or more subsets of at least one of the items a to d used to describe the partial activation range.

27. The apparatus of any one of claims 22 to 26, wherein the request comprises: At least one of a network context and an artificial intelligence and / or machine learning context is used to describe the partial activation range.

28. A device according to any one of claims 22 to 27, wherein the at least one memory and the computer program are also configured to, together with the at least one processor, cause the network device to perform: determining (205) whether data describing the performance of the capability after further activation at the requested activation range meets performance criteria, and in the affirmative case, sending (208) a request to increase the activation range previously requested to the other network device.

29. The device of claim 28, wherein the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to execute: requesting to increase the activation range of the capability only if (207) a full activation range has not yet been reached.

30. The device according to claim 29 or 30, wherein the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to perform: If the criterion is not met, a deactivation request is sent (206) to the other network device, the deactivation request identifying one of: ○ an instruction to cancel the activation range request; o Complete deactivation of said capability; o a reduced activation range to be achieved compared to said activation range where said criterion is not met; o one or more portions of the activation range of the capability to be deactivated; ○ A level among the plurality of predefined ordered activation levels, identifying the gradually increasing capability activation range as described in claim 2.

31. The device according to any one of claims 22 to 30, wherein the at least one memory and the computer program are also configured to, together with the at least one processor, cause the network device to perform: determining that network performance is degraded, and in response to the determination, requesting a partial reduction in the scope of capability activation.

32. The device according to any one of claims 22 to 31, wherein the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to perform: obtaining from the other network device at least one of an accuracy measurement or data describing a benefit assessment performed by the other device, wherein: The accuracy measurement or data describing the benefit assessment is a function of the output of the capability for an envisioned range of partial activation without applying the output to an operating network environment, and the request for the partial activation is sent based on at least one of the accuracy measurement or the data describing the benefit assessment.

33. A network device comprising: At least one processor (510), at least one memory (520), the at least one memory (520) including computer program code (570), the network device including a network entity, the network entity including a capability for providing an output based on artificial intelligence and / or machine learning reasoning functions related to one or more network objects or object types, the at least one memory and computer program code being configured to, together with the at least one processor, cause the network device to perform (303) at least a partial activation of the capability.

34. The device of claim 33, wherein the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to perform: making information describing the capability available to other network devices (301, 401).

35. The apparatus of claim 34, wherein: The information describing the capability includes one or more of the following: a. one or more objects and / or object types, the network entity being configured to optimize or control the one or more objects and / or object types; b. one or more parameters of the one or more objects and / or object types, the network entity being configured to control or optimize the one or more parameters of the one or more objects and / or object types to achieve a desired result; c. one or more network metrics, the network entity being configured to seek to improve the one or more network metrics; d. One or more network characteristics associated with the one or more objects and / or object types.

36. A device according to any one of claims 34 or 35, wherein the at least one memory and the computer program are also configured to, together with the at least one processor, cause the network device to perform: making information describing a plurality of predefined ordered activation levels available to other network devices (402), the plurality of predefined ordered activation levels identifying gradually increasing capability activation ranges within an available range.

37. The device of any one of claims 34 to 36, wherein the at least one memory and the computer program are further configured to, together with the at least one processor, cause the network device to perform: making at least one of the accuracy measurement or the data describing the benefit evaluation performed by another device available to other network devices (403), wherein The accuracy measure or data describing the benefit assessment is a function of the output of the capability for variations in the range of activations envisioned, without applying the output to an operational environment.

38. The device according to any one of claims 33 to 37, wherein the partial activation of the capability is performed in response to a request (302, 404) for partial activation from another network device.

39. The device of claim 38, wherein the at least one memory and the computer program are further configured to, with the at least one processor, cause the network device to perform: After activation of the capability at the requested scope, status information describing the activation status of the capability with respect to the requested scope is sent (304, 405) to the other network devices.

40. The device of claim 39, wherein the at least one memory and the computer program are further configured to, with the at least one processor, cause the network device to perform: receiving in the request at least one of a network context, and an artificial intelligence and / or machine learning context; and Responsive to the context, limiting the activation scope of the capability.

41. The device according to any one of claims 33 to 40, wherein the at least one memory and the computer program are also configured to, together with the at least one processor, cause the network device to perform: determining that network performance is degraded, and partially reducing the capability activation range in response to the determination.

42. The apparatus of any one of claims 33 to 41, wherein activation of the capability comprises: The output of the capabilities is applied to an operational network environment.

43. A computer readable medium comprising program instructions stored thereon, the program instructions being configured to perform at least the following: Acquiring (201, 401) information describing capabilities of a network entity (103) of a second network device (102), the network entity comprising: Capabilities for providing output based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types; A request for partial activation of the capability is sent (203, 404) to the second network device.

44. A computer readable medium comprising program instructions stored thereon, the program instructions being configured to perform at least the following in a network device comprising a network entity, the network entity comprising: Capabilities for providing output based on artificial intelligence and / or machine learning reasoning functions associated with one or more network objects or object types; Partial activation of said capability.

Images