Associating machine learning models related to federated learning operations

Abstract

A method is provided that includes assigning a vertical federated learning (VFL) session identifier to a VFL model training session. The method includes selecting network functions of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space. And the method includes sending a request to each network function to perform model training on a respective one of the machine learning models in the common sample space with a respective one of the different feature spaces. The request includes the VFL session identifier to cause the machine learning model to be stored in association with the VFL session identifier. The method can also include training a federated machine learning model to compute a prediction based on intermediate output data from the network functions, and storing the federated machine learning model in association with the VFL session identifier.

Description

Technical Field

[0001] This disclosure generally relates to telecommunications, and more particularly to federated machine learning among multiple network functions in a telecommunications system. Background Technology

[0002] A telecommunications system can be viewed as a facility that enables a communication session between two or more entities (such as user terminals, base stations, and / or other nodes) by providing carriers between various entities involved in the communication path. A telecommunications system can be provided, for example, by means of a communication network and one or more compatible communication devices. The communication session can include, for example, communication of data carrying communications such as voice, video, email, text messages, multimedia, and / or content data. Non-limiting examples of the services provided include two-way or multiplexed calls, data communication or multimedia services, and access to data network systems such as the Internet.

[0003] In a wireless telecommunications system, at least a portion of a communication session between at least two stations occurs over a wireless link. Examples of wireless systems include Public Land Mobile Networks (PLMNs), satellite-based communication systems, and various wireless local area networks (WLANs). Some wireless systems can be divided into cells and are therefore often referred to as cellular systems.

[0004] Users can access the telecommunications system through appropriate communication equipment or terminals. A user's communication equipment may be referred to as User Equipment (UE) or simply user equipment. The communication equipment is equipped with appropriate signal receiving and transmission means for enabling communication, such as access to a communication network or direct communication with other users. The communication equipment can access a carrier provided by a station (e.g., a base station in a cell) and transmit and / or receive communication on that carrier.

[0005] Telecommunications systems and associated equipment typically operate according to a given standard or specification that outlines what the various entities associated with the system are allowed to do and how they should be implemented. The communication protocols and / or parameters used for connectivity are also usually defined. An example of a telecommunications system is the Universal Mobile Telecommunications System (UMTS). Other examples include Long Term Evolution (LTE), Advanced Long Term Evolution (ATE), and so-called 5G or New Radio (NR) networks. NR is being standardized by the 3rd Generation Partnership Project (3GPP). Summary of the Invention

[0006] The example implementations of this disclosure relate to telecommunications, and more particularly to federated machine learning among multiple network functions in a telecommunications system. More specifically, the example implementations relate to vertical federated learning among multiple network functions. This disclosure includes, but is not limited to, the following example implementations.

[0007] Some example implementations provide an apparatus comprising: at least one memory configured to store instructions; and at least one processing circuitry configured to access the at least one memory and execute instructions such that the apparatus at least: assigns a vertical federated learning session identifier to a vertical federated learning model training session; selects a network function of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space; and sends a request to each network function for performing model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier to store the machine learning model in association with the vertical federated learning session identifier.

[0008] Some example implementations provide an apparatus including: components for assigning a vertical federated learning session identifier to a vertical federated learning model training session; components for selecting a network function of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space; and components for sending a request to each network function for performing model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier such that the machine learning model is stored in association with the vertical federated learning session identifier.

[0009] Some example implementations provide a method including: assigning a vertical federated learning session identifier to a vertical federated learning model training session; selecting a network function of a telecommunications system to perform model training on machine learning models with different feature spaces in a common sample space; and sending a request to each network function to perform model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier to store the machine learning model vertical federated learning session identifier in association.

[0010] Some example implementations provide a computer-readable storage medium that is non-transitory and has instructions stored therein that, in response to execution by at least one processing circuit, cause the means to at least: assign a vertical federated learning session identifier to a vertical federated learning model training session; select a network function of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space; and send a request to each network function for performing model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier to store the machine learning model in association with the vertical federated learning session identifier.

[0011] Some example implementations provide an apparatus including: at least one memory configured to store instructions; and at least one processing circuitry configured to access the at least one memory and execute the instructions to cause the apparatus to at least: receive a request from a network function of a telecommunications system for performing model training of a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having different feature spaces, the request including a vertical federated learning session identifier; train the machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; send the intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and store the machine learning model in association with the vertical federated learning session identifier.

[0012] Some example implementations provide an apparatus including: components for receiving a request from a network function of a telecommunications system for performing model training of a machine learning model, which will be one of machine learning models trained in a common sample space but with different feature spaces, the request including a vertical federated learning session identifier; components for training the machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; components for sending the intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and components for storing the machine learning model in association with the vertical federated learning session identifier.

[0013] Some example implementations provide a method including: receiving a request from a network function of a telecommunications system for performing model training of a machine learning model, which will be one of machine learning models trained in a common sample space but with different feature spaces, the request including a vertical federated learning session identifier; training the machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; sending the intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and storing the machine learning model in association with the vertical federated learning session identifier.

[0014] Some example implementations provide a computer-readable storage medium that is non-transitory and has instructions stored therein that, in response to execution by at least one processing circuit, cause the apparatus to at least: receive a request from a network function of a telecommunications system for performing model training of a machine learning model, which will be one of machine learning models trained in a common sample space but with different feature spaces, the request including a vertical federated learning session identifier; train the machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; send the intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and store the machine learning model in association with the vertical federated learning session identifier.

[0015] Some example implementations provide an apparatus including: at least one memory configured to store instructions; at least one processing circuit configured to access the at least one memory and execute the instructions to cause the apparatus to at least: receive a request for providing at least one of: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; select a network function to provide information about a vertical federated learning session associated with the analysis identifier; send a request to the network function, the request including the analysis identifier, the request being sent to the network function for use with a federated machine learning model; receive a response from the network function, the response including the vertical federated learning session identifier and information about the federated machine learning model; access the federated machine learning model associated with the vertical federated learning session identifier; send a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a common sample space with different feature spaces, the request including the vertical federated learning session identifier and the analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; receive intermediate output data from the network function based on the inference; and perform inference using the federated machine learning model based on the intermediate output data to compute predictions.

[0016] Some example implementations provide an apparatus including: components for receiving a request for providing at least one of: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; components for selecting a network function to provide information about a vertical federated learning session associated with the analysis identifier; components for sending a request to the network function, the request including the analysis identifier, the request being sent to the network function for use with a federated machine learning model; components for receiving a response from the network function, the response including a vertical federated learning session identifier and information about the federated machine learning model; components for accessing a federated machine learning model associated with the vertical federated learning session identifier; components for sending a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a common sample space with different feature spaces, the request to the network function of the telecommunications system including the vertical federated learning session identifier and the analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; components for receiving intermediate output data from the network function based on inference; and components for performing inference using the intermediate output data and the federated machine learning model to compute predictions.

[0017] Some example implementations provide a method including: receiving a request for providing at least one of the following: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; selecting a network function to provide information about a vertical federated learning session associated with the analysis identifier; sending a request to the network function, the request including the analysis identifier, the request being sent to the network function for use with a federated machine learning model; receiving a response from the network function, the response including the vertical federated learning session identifier and information about the federated machine learning model; accessing a federated machine learning model associated with the vertical federated learning session identifier; sending a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a public sample space with different feature spaces, the request including the vertical federated learning session identifier and the analysis identifier, for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; receiving intermediate output data from the network function based on the inference; and performing inference using the federated machine learning model based on the intermediate output data to compute predictions.

[0018] Some example implementations provide a computer-readable storage medium that is non-transitory and has instructions stored therein, the instructions being, in response to execution by at least one processing circuit, to cause a device to at least: receive a request for providing at least one of: analysis, or one or more predictions, of a data analysis process associated with an analysis identifier; select a network function to provide information about a vertical federated learning session associated with the analysis identifier; send a request to the network function, the request including the analysis identifier, the request being sent to the network function for use in a federated machine learning model; and receive a response from the network function, the response including... The process involves: a vertical federated learning session identifier and information about a federated machine learning model; accessing the federated machine learning model, which is associated with the vertical federated learning session identifier; sending a request to the network function of the telecommunications system for performing inference using the machine learning model, which is trained in a public sample space with different feature spaces, the request including the vertical federated learning session identifier and an analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; receiving intermediate output data from the network function based on inference; and performing inference using the federated machine learning model based on the intermediate output data to compute predictions.

[0019] Some example implementations provide an apparatus comprising: at least one memory configured to store instructions; and at least one processing circuitry configured to access the at least one memory and execute the instructions such that the apparatus at least: receives a request from a network function of a telecommunications system for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having a different feature space, the request including a vertical federated learning session identifier; accesses the machine learning model associated with the vertical federated learning session identifier; performs inference using the machine learning model to compute intermediate output data; and sends the intermediate data to the network function for inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0020] Some example implementations provide an apparatus including: components for receiving a request from a network function of a telecommunications system, the request being for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having a different feature space, the request including a vertical federated learning session identifier; components for accessing the machine learning model associated with the vertical federated learning session identifier; components for performing inference using the machine learning model to compute intermediate output data; and components for sending the intermediate data to the network function to perform inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0021] Some example implementations provide a method comprising: receiving a request from a network function of a telecommunications system for performing inference using a machine learning model, the machine learning model being one of several machine learning models trained in a common sample space but having different feature spaces, the request including a vertical federated learning session identifier; accessing the machine learning model associated with the vertical federated learning session identifier; performing inference using the machine learning model to compute intermediate output data; and sending the intermediate data to the network function for inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0022] Some example implementations provide a computer-readable storage medium that is non-transitory and has instructions stored therein that, in response to execution by at least one processing circuitry, cause the apparatus to at least: receive a request from a network function of a telecommunications system for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having a different feature space, the request including a vertical federated learning session identifier; access the machine learning model associated with the vertical federated learning session identifier; perform inference using the machine learning model to compute intermediate output data; and send the intermediate data to the network function for inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0023] These and other features, aspects, and advantages of this disclosure will become apparent from the following detailed description and the accompanying drawings, which are briefly described below. This disclosure includes any combination of two, three, four, or more features or elements set forth in this disclosure, whether or not such features or elements are explicitly combined in the particular exemplary embodiments described herein or otherwise narrated. This disclosure is intended to be read holistically such that any separable feature or element of this disclosure should be considered composable in any aspect and exemplary embodiment, unless the context of this disclosure expressly provides otherwise.

[0024] Therefore, it should be understood that the content of this invention is provided merely to outline some exemplary embodiments in order to provide a basic understanding of some aspects of this disclosure. Consequently, it should be understood that the above-described exemplary implementations are merely examples and should not be construed as limiting the scope or spirit of this disclosure in any way. Other exemplary embodiments, aspects, and advantages will become apparent from the following detailed description taken in conjunction with the accompanying drawings, which illustrate the principles of some of the described exemplary embodiments by way of example. Attached Figure Description

[0025] Therefore, exemplary embodiments of this disclosure have been described in general terms. Reference will now be made to the accompanying drawings, which are not necessarily drawn to scale, and in which: Figure 1 The present disclosure illustrates a telecommunications system comprising one or more Public Land Mobile Networks (PLMNs) coupled to one or more external data networks, according to some example embodiments thereof; Figure 2 The deployment of a PLMN according to some example implementations is illustrated; Figure 3 More specifically, embodiments based on some examples are described. Figure 2 All aspects of the deployment; Figure 4A trained machine learning (ML) model supply architecture is illustrated according to some example implementations; Figure 5A and 5B Signaling diagrams for vertical federated learning (VFL) training and inference according to some example implementations are shown; Figure 6 A signaling diagram showing VFL sample alignment of a user equipment sample according to some example implementations is shown; Figure 7A , 7B Figures 7 and 7C show signaling diagrams for VFL training and inference according to other example implementations; Figure 8 A signaling diagram showing VFL sample alignment of a user equipment sample according to other example implementations is shown; Figure 9A , 9B 9C, 9D, 9E, 9F and 9G are flowcharts illustrating various steps in a method according to various exemplary embodiments; Figure 10A , 10B 10C and 10D are flowcharts illustrating various steps in a method according to various exemplary embodiments; Figure 11A , 11B 11C and 11D are flowcharts illustrating various steps in a method according to various exemplary embodiments; Figure 12A , 12B 12C and 12D are flowcharts illustrating various steps in a method according to various exemplary embodiments; and Figure 13 An apparatus according to some example implementations is shown. Detailed Implementation

[0026] Some embodiments of this disclosure will now be described more fully below with reference to the accompanying drawings, which illustrate some, but not all, of these embodiments. In fact, various embodiments of this disclosure may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these exemplary embodiments are provided so that this disclosure will be thorough and complete and will fully convey the scope of this disclosure to those skilled in the art. The same reference numerals throughout refer to the same elements.

[0027] Unless otherwise stated or clear from the context, references to first, second, etc., should not be construed as implying a particular order. A feature described as above another feature (unless otherwise stated or clear from the context) may alternatively be below, and vice versa; and similarly, a feature described as to the left of another feature may alternatively be to the right, and vice versa. Furthermore, while references may be made herein to quantitative measurements, values, geometric relationships, etc., any one or more of these (if not all) may be absolute or approximate to account for possible acceptable variations, such as those due to engineering tolerances, etc.

[0028] As used herein, unless otherwise stated or clearly indicated from the context, an "OR" of a set of operands is an "inclusive OR," meaning it is true if and only if one or more operands are true, not a "XOR" which is false if all operands are true. Thus, for example, "[A] OR [B]" is true if [A] is true, or if [B] is true, or if both [A] and [B] are true. Furthermore, the articles "a" and "one" mean "one or more," unless otherwise stated or clearly indicated from the context to refer to the singular form. Additionally, it should be understood that, unless otherwise stated, the terms "data," "content," "digital content," "information," and similar terms are sometimes used interchangeably. The term "network" can refer to a group of interconnected computers including clients and servers; and within a network, these computers can be interconnected directly or indirectly by various means, including via one or more switches, routers, gateways, access points, etc.

[0029] This document may refer to terminology specific to a particular system, architecture, etc., but it should be understood that the exemplary implementations of this disclosure are equally applicable to any of a number of systems, architectures, etc. For example, reference may be made to 3GPP technologies such as Global System for Mobile Communications (GSM), UMTS, LTE, Advanced LTE, 5G NR, Advanced 5G, and 6G; however, it should be understood that the exemplary implementations of this disclosure are equally applicable to non-3GPP technologies such as IEEE 802, Bluetooth, and Bluetooth Low Energy.

[0030] Furthermore, as used in this application, the term "circuit" may refer to one or more of the following: (a) a hardware circuit implementation only (such as an implementation only in analog and / or digital circuits); (b) a combination of hardware circuits and software, such as (if applicable): (i) a combination of (multiple) analog and / or digital hardware circuits with software / firmware, and (ii) (multiple) hardware processors with any part of the software (including (multiple) digital signal processors), software, and (multiple) memories, which work together to enable a device such as a mobile phone or a server to perform various functions; or (c) (multiple) hardware circuits and / or (multiple) processors, such as (multiple) microprocessors or a portion thereof, which require software (e.g., firmware) to operate, but may be absent when the software is not required to operate.

[0031] The above definition of "circuit" applies to all uses of the term in this application (including in any claim). As another example, as used herein, the term "circuit" also encompasses only hardware circuitry or a processor (or multiple processors) or a portion thereof, and its accompanying software and / or firmware. The term "circuit" also encompasses, for example and if applicable to a particular claim element, baseband integrated circuits or processor integrated circuits for mobile devices, or similar integrated circuits in servers, cellular network devices, or other computing or network devices.

[0032] Figure 1 A telecommunications system 100 according to various example embodiments of the present disclosure is illustrated. A telecommunications system typically includes one or more telecommunications networks. As shown, for example, the system includes one or more Public Land Mobile Networks (PLMNs) 102 coupled to one or more other external data networks 104, particularly including wide area networks (WANs) such as the Internet. Each PLMN includes a core network (CN) 106 backbone, such as an Evolved Packet Core Network (EPC) for LTE, a 5G core network (5GC), etc.; and each core network and the Internet are coupled to one or more Radio Access Networks (RANs) 108 implementing one or more Radio Access Technologies (RATs), air interfaces, etc. As used herein, "network device" refers to any suitable device on the network side of the telecommunications network. Examples of suitable network devices are described in more detail below.

[0033] Furthermore, the system includes one or more radio units, which may be referred to differently as User Equipment (UE) 110, terminal equipment, terminal gear, mobile station, etc. The UE is typically a device configured to communicate with network equipment in a telecommunications network or with another UE. The UE can be a portable computer (e.g., laptop, notebook, tablet), a mobile phone (e.g., cellular, smartphone), a wearable computer (e.g., smartwatch), etc. In other examples, the UE can be an Internet of Things (IoT) device, an Industrial IoT (IIoT) device, a vehicle equipped with Vehicle-to-Everything (V2X) communication technology, etc. In some examples, as cited by 3GPP, the UE can be a Narrowband IoT (NB-IoT) device, an enhanced machine-type communication (eMTC) device, a RedCap device, an environmental IoT device, etc.

[0034] In operation, these UEs can be configured to connect to one or more RANs 108 based on their specific radio access technologies, thereby accessing a specific CN 106 of the PLMN 102, or accessing one or more external data networks 104 (e.g., the Internet). External data networks can be configured to provide Internet access, operator services, third-party services, etc. For example, the International Telecommunication Union (ITU) has categorized 5G mobile network services into three classes: enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC), and massive machine-type communications (mMTC) or massive Internet of Things (MIoT).

[0035] Examples of radio access technologies include 3GPP radio access technologies such as GSM, UMTS, LTE, Advanced LTE, 5G NR, Advanced 5G, and 6G. Other examples of radio access technologies include IEEE 802 technologies such as IEEE 802.11 (Wi-Fi), IEEE 802.15 (including 802.15.1 (WPAN / Bluetooth), 802.15.4 (Zigbee), and 802.15.6 (WBAN)), Bluetooth, Bluetooth Low Energy (BLE), Ultra Wideband (UWB), etc. Generally, radio access technology can refer to any 2G, 3G, 4G, 5G, 6G, or higher generation mobile communication technology and its different versions, as well as any other radio access technology that can be arranged to interoperate with such mobile communication technology to provide access to the CN106 of a mobile network operator (MNO).

[0036] In various examples, RAN108 can be configured as one or more macro cells, micro cells, pico cells, femtocells, etc. RAN typically includes one or more radio access nodes configured to interact with UE110. In various examples, radio access nodes may be referred to as base stations (BS), access points (APs), base transceiver stations (BTS), B-nodes (NBs), evolved NBs (eNBs), macro BSs, NBs (MNBs) or eNBs (MeNBs), home BSs, NBs (HNBs) or eNBs (HeNBs), next-generation NBs (gNBs), enhanced gNBs (en-gNBs), next-generation eNBs (ng-eNBs), etc. RAN may include some type of network control / management entity responsible for controlling the radio access nodes. The network control / management entity and the radio access nodes may be separate or integrated into a single device. The network control / management entity may include processing circuitry configured to perform various management functions, etc. The processing circuitry may be associated with a memory, computer-readable storage medium, or database used to maintain information required for management functions.

[0037] RAN108 can be centralized or distributed. In various examples, RAN components can be interconnected via Ethernet, Gigabit Ethernet, Asynchronous Transfer Mode (ATM), fiber optic, dark fiber, passive wavelength division multiplexing (WDM), WDM passive optical network (WDM-PON), optical transport network (OTN), time-sensitive network (TSN), and / or any other data link layer network (potentially including radio links). RAN can be connected to CN106 via one or more gateways, network functions, etc.

[0038] As will be understood, PLMN102 can be deployed in several different ways. In 4G LTE deployment, EPC is CN106, and Evolved UMTS Terrestrial Radio Access Network (E-UTRAN) is RAN108; E-UTRAN includes one or more eNBs (radio access nodes) for connecting UE110 to E-UTRAN to access EPC. Figure 2 The diagram illustrates deployment 200, such as a 5G or 6G deployment. As shown, 5GC202 is the CN, and Next Generation (NG) Radio Access Network (NG-RAN) 204 is the RAN; and the NG-RAN includes one or more gNBs 206 (radio access nodes) configured to connect the UE 208 to the NG-RAN for access to the 5GC. The term "gNB" in 5G can correspond to the eNB in ​​4G LTE.

[0039] Some 4G LTE and 5G deployments are considered standalone (SA) deployments. Other deployments combine 4G LTE and 5G technologies and are referred to as non-standalone (NSA) deployments. In some deployments, the E-UTRAN includes one or more ng-eNBs configured to communicate with the 5GC and may also be configured to communicate with one or more gNBs. Similarly, in another deployment, the NG-RAN may include one or more en-gNBs configured to communicate with the EPC and may also be configured to communicate with one or more eNBs. In various cases, a single UE110 (dual-mode or multi-mode UE) can support multiple (two or more) RANs, thus being configured to connect to multiple RANs, such as 4G LTE and 5G.

[0040] Figure 3 More specifically, aspects of a deployment 200 for an MNO, implemented according to several examples, are depicted. As shown, the deployment includes a 5GC 202 and an NG-RAN 204, where one or more gNBs 206 are configured to connect a UE 208 to the NG-RAN, thereby accessing the 5GC. The 5GC may include multiple network functions (NFs) partitioned between the control plane and the user plane. Specifically, the 5GC may include, for example, an Access and Mobility Management Function (AMF) 302, a Session Management Function (SMF) 304, a User Plane Function (UPF) 306, a Network Exposure Function (NEF) 308, a Network Data Analysis Function (NWDAF) 310, and / or an Application Function (AF) 312. Other examples of suitable NFs include a Network Repository Function (NRF) 314, a Network Slice Selection Function (NSSF), a Policy Control Function (PCF), a Unified Data Management (UDM) 316, and so on.

[0041] In the control plane, AMF302 is configured to provide UE-based authentication, authorization, mobility management, etc. SMF304 is configured to provide various functions, including Session Management (SM), UE Internet Protocol (IP) address allocation and management, selection and control of (multiple) UPF306s, control portion of policy enforcement and Quality of Service (QoS), lawful interception, termination of the SM portion of NAS messages, Downlink Data Notification (DNN), roaming functions, handling local implementation to apply QoS against Service Level Agreements (SLAs), charge data collection, and charge interface, etc. If UE208 has multiple sessions, different SMFs can be assigned to each session for individual management, and different functions may be provided for each session.

[0042] UPF306 supports various user plane operations and functions, such as packet routing and forwarding, service processing (e.g., QoS enforcement), anchor points for intra / inter-RAT mobility (where applicable), packet inspection and policy rule enforcement, lawful interception (UP collection), service accounting and reporting, etc. The UPF is the interconnection point between the 5GC and at least one external data network (DN) 318 (i.e., the DN's ingress or egress point), and routes packets to and from the DN. The DN can be configured to provide Internet access, carrier services, third-party services, etc. The NWDAF310 collects and analyzes network data to provide insights into the performance, optimization, and overall health of the 5GC202. This information can be used in network management, optimization, and decision-making processes to improve network efficiency.

[0043] The AF312 can interact with the 5GC202 to deploy specific services and applications. The AF communicates with other NFs to request and manage network resources, ensuring the network adapts to the requirements of different applications and services. The NEF308 allows authorized third-party applications and services to access specific network functions and services in a controlled manner. NEFs enable network capabilities to be exposed to external entities, thereby facilitating the innovation and development of new services.

[0044] In some deployments (such as deployment 200), the operation of the gNB206 or other radio access nodes can be performed, at least in part, in a centralized unit (CU) (such as a server, host, or node) that is operationally coupled to a distribution unit (DU) (such as a radio head / node). Node operations can also be distributed among multiple servers, hosts, or nodes. It should also be understood that the operational distribution between the 5GC202 (or other CN) operations and the gNB (or other radio access node) operations can vary depending on the specific implementation.

[0045] 5G network architecture can be based on a so-called CU-DU split. A gNB-CU (central node) can control one or more gNB-DUs. A gNB-CU can control multiple spatially separated gNB-DUs, at least acting as transmit / receive (Tx / Rx) nodes. However, in some example implementations, a gNB-DU (also called a DU) can include, for example, the Radio Link Control (RLC), Media Access Control (MAC) layer, and Physical (PHY) layer, while a gNB-CU (also called a CU) can include layers above the RLC layer, such as the Packet Data Convergence Protocol (PDCP) layer, Radio Resource Control (RRC), and Network Protocol (IP) layer. Other functional splits are also possible. It is assumed that those skilled in the art are familiar with the Open Systems Interconnection (OSI) model and the functions within each layer.

[0046] In some example implementations, the server or CU can generate a virtual network through which the server communicates with radio nodes. Typically, virtual networking can involve the process of combining hardware and software network resources and functions into a single software-based management entity (virtual network). Such a virtual network can provide flexible distribution of operations between the server and the radio head end / node. In practice, any digital signal processing task can be performed in the CU or DU, and the boundaries of responsibility transferred between the CU and DU can be selected depending on the implementation.

[0047] Many applications in telecommunications involve using large amounts of data from multiple components (such as UE208) to train a single common machine learning (ML) model (sometimes more simply called a "model"). In this regard, federated learning (FL) is a form of machine learning where, instead of training the model on a single node, different versions of the model are trained on different distributed nodes. This differs from distributed machine learning, where a single ML model is trained on distributed nodes to utilize the computing power of the different nodes.

[0048] Multiple federated learning architectures follow a client-server model involving a server and multiple distributed clients. In federated learning, each distributed client computes the parameters of its local ML model, and the server, instead of computeing a version or part of the ML model, combines the parameters of all distributed ML models to generate a federated model. In this respect, after training the local ML model, each individual client sends its local ML model parameters, rather than the original training dataset, to the server. The server uses the local ML model parameters to update the federated model, which is ultimately fed back to the distributed clients for their use.

[0049] Federated learning can be categorized into several scenarios, including horizontal federated learning (HFL) and vertical federated learning (VFL). In HFL, distributed clients utilize a shared feature space but employ different sampling spaces (e.g., UEID) to train their local ML models. As described in this paper, a feature space refers to a single feature or a set of features. In HFL, each client keeps its training dataset local to where it was generated. Clients benefit from the datasets of other learners, but only through the federated model shared by the server, without accessing large amounts of privacy-sensitive data.

[0050] In VFL, distributed clients utilize a common sample space but different feature spaces (different features or feature sets) to train their local ML models. In this respect, each client can train its local ML model in the common sample space using a corresponding feature space from its respective feature space. Compared to HFL, federated models in VFL can be trained using fewer samples shared among clients. However, federated models in VFL can provide deeper federated models by utilizing a larger number of features organized into different sample spaces, which are then distributed among the clients to train their respective local ML models.

[0051] In 3GPP, the key issue is to provide a solution for enabling 5GC support for VFLs that involve one or more NFs (such as (multiple) NWDAF310 and / or (multiple) AF312). Figure 4 A trained ML model provisioning architecture 400 according to some example implementations is illustrated. As shown, the trained ML model provisioning architecture includes an Analysis Logic Function (AnLF) 402 and a Model Training Logic Function (MTLF) 404, which are logical functions of an NWDAF and can be connected to each other via an Nnwdaf interface. In this respect, an NWDAF containing an AnLF can use trained ML model provisioning services from another NWDAF containing an MTLF.

[0052] AnLF402 is a logical function of NWDAF310, configured to perform inference, derive analytics information (e.g., derive statistics and / or predictions based on analytics consumer requests), and expose analytics services (e.g., Nnwdaf_AnalyticsSubscription, Nnwdaf_AnalyticsInfo). MTLF404 is a logical function of NWDAF, configured to train ML models and expose new training services (e.g., provide trained ML models). NWDAF may include AnLF, MTLF, or both AnLF and MTLF.

[0053] In the NWDAF310 architecture, MTLF404 can be configured to train ML models, and AnLF402 can be configured to perform inference analytics based on the trained models. ML models can also be stored in and retrieved from the Analytics Data Repository Function (ADRF), which is an NF that can be used to store, retrieve, and remove data or for analysis by NF consumers (e.g., NWDAF) that access data using the Nadrf service.

[0054] The example implementations of this disclosure involve VFLs between multiple NFs (such as NWDAF310 and / or AF312). Some example implementations assume that the VFL is controlled by an NWDAF acting as a server (or active participant), and the client (or passive participant) can be another NWDAF or AF. Some example implementations use an internal NWDAF architecture where MTLF404 can be configured to train the model, and AnLF402 can be configured to perform inference based on the trained model. In some examples, the model can be stored at ADRF.

[0055] According to some examples, when a VFL model training session begins, an MTLF404 operating as a VFL training server can discover and select one or more MTLFs and / or AF312s that act as or otherwise operate as VFL training clients for one or more specific feature spaces. This discovery can use NRF314, where MTLFs register their ability to act as or otherwise operate as VFL training servers or VFL training clients, along with associated analytics identifiers (IDs). Similarly, VFL training clients can also register their supported feature spaces, which can be associated with feature IDs. In some examples, the feature spaces and associated intermediate results to be exchanged between training clients and training servers are not normalized, and feature IDs may include the vendor's vendor ID and an identifier assigned by the vendor.

[0056] In an example where training involves UE-related data as samples, the VFL training server (MTLF404) can selectively register and reach UE208 as a sample, and monitor whether the UE remains reachable during training. For AF312, which can operate as a VFL training client, the VFL training server can send a request to check whether the AF can reach the UE.

[0057] The MTLF404, acting as the VFL training server, can assign a VFL session identifier to the VFL model training session and send a request to start training to the selected VFL clients (MTLF or AF312). Each request may indicate an analytics ID, feature ID, desired samples (e.g., a list of UEs), and / or a VFL session ID. The MTLF itself, acting as the VFL training server, can train a model that combines information related to several features across the feature space (a federated model). Each VFL training client can train a model related to the feature space. These models are trained together over several epochs. After each epoch, each VFL training client provides the server with so-called “gradients” as intermediate results used by the server to update its federated model, and the server also provides each client with intermediate results used by the client to update its feature-related models.

[0058] After training is complete, both the MTLF404, which acts as the VFL training server, and the MTLF, which acts as the VFL training client, store their trained models at ADRF along with the associated analytics ID, feature ID (for the client) or server model indicator and VFL session ID.

[0059] Inference for analysis or predictions associated with a VFL-trained model can be performed by AnLF402, which acts as the inference server, and an inference client (AnLF or AF312) for a specific feature space. Models already trained together can be used together for inference.

[0060] When AnLF402, acting as an inference server, receives a request or subscription for analysis or prediction for a specific analysis ID, AnLF can discover MTLF404 for that analysis ID and request the relevant model from that MTLF. For VFLs, MTLFs acting as VFL training servers can be contacted. MTLFs acting as VFL training servers can indicate that a federated model is associated with a VFL and provide a VFL session ID and feature IDs for training the federated model. AnLF acting as a VFL inference server can use NRF314 to discover VFL inference clients for the indicated feature space, sending requests or subscriptions to each VFL inference client for inference outputs associated with each feature space. Requests to each VFL inference client can include the analysis ID, feature ID, and training session ID. AnLF acting as an inference client can use NRF to discover MTLFs acting as training clients with the same analysis ID and feature ID, and request models with the analysis ID, feature ID, and VFL session ID from those MTLFs.

[0061] Each AnLF402, acting as a VFL inference client, can retrieve input data and use the retrieved model in the feature space to compute output data associated with the feature space and analytics ID, providing the output data to the AnLF acting as the inference server. The AnLF402 acting as the VFL inference server can then use the retrieved federated model and the received feature-related output data to compute analytics or one or more predictions.

[0062] In some examples, it is assumed that there is no separate controller. In these examples, the MTLF404, acting as the VFL training server, can also train the federated model and can therefore be considered an "active participant" for training. In other examples, a separate NWDAF310 or other NF can be provided as the "VFL coordinator," as the NWDAF can select active and passive participants for VFL training and request the active participants to train the federated model and the passive participants to train models associated with different feature spaces. The MTLF, acting as a VFL training client, can also be referred to as a "passive participant" for training.

[0063] In some examples, the AnLF402, acting as the VFL inference server, can also use the federated model and can therefore be considered an "active participant" for inference. In other examples, a separate NWDAF310 or other NF can be provided as the "VFL coordinator," since the NWDAF can select active and passive participants for inference. The NWDAF can request active participants to perform inference on the federated model, and passive participants to perform inference related to the feature model. The AnLF, acting as a VFL inference client, can also be considered a "passive participant" for inference. In some examples, the VFL coordinator can be the same for both training and inference.

[0064] To further illustrate some example implementations of this disclosure, Figure 5A and 5B Signaling diagram 500 for VFL training and inference is shown. (See diagram 500.) Figure 5A As shown, in step 501, AnLF402A, acting as a VFL inference server, registers its ability to act as a VFL inference server, along with its supported analytics IDs and vendor IDs, in NRF314. As shown in step 502, AnLF402B, acting as a VFL inference client, registers its ability to act as a VFL inference client, along with its supported analytics IDs and associated feature IDs, in the NRF. In step 503, MTLF404B, acting as a VFL training client, registers its ability to act as a VFL training client, along with its supported analytics IDs and associated feature IDs, in the NRF. And based on the configured information, NEF308, in step 504, registers on behalf of AF312 the ability of AF to act as a VFL client in the NRF, along with the analytics IDs and associated feature IDs supported by AF.

[0065] At step 505, the MTLF404A, acting as the VFL training server, receives or observes a trigger to begin VFL model training for the analytics ID. At step 506, the MTLF, acting as the VFL training server, selects the features required for model training based on the implementation.

[0066] For each feature, in step 507, MTLF404A, acting as the VFL training server, sends a discovery request to NRF314 for either MTLF404B, acting as the VFL training client, or AF312, acting as the VFL client, providing the analysis ID and feature ID. For each feature, when the VFL training server selects such an NF, NRF provides a profile of a candidate NF that matches the discovery request and MTLF in step 508.

[0067] The MTLF404A, acting as the VFL training server, can perform sample alignment with the VFL training client at step 509. If UE208 is used as a sample, the process described below in Figure 7 can be applied.

[0068] In step 510, the MTLF404A, acting as the VFL training server, assigns a unique VFL session ID. In step 511, the MTLF, acting as the VFL training server, sends a request to the VFL training client to start VFL training for each feature, providing the analysis ID, feature ID, and VFL session ID.

[0069] In step 512, the MTLF404A, acting as the VFL training server, and the VFL training clients perform VFL model training. The VFL training server trains a model capable of combining information related to several features. Each VFL training client trains a model related to a specific feature. These models are trained together over several epochs. After each epoch, each VFL training client provides the VFL training server with so-called "gradients" as intermediate results, which the server uses to update its model. The server also provides intermediate results to each VFL training client, which the client uses to update its feature-related model.

[0070] In step 513, the VFL training server stores its model trained in ADRF along with the analytics ID, VFL session ID, and feature IDs of all features used to train the model, as well as an indication of the model's use by the VFL server. In step 514, each VFL training client stores its model trained in ADRF along with the analytics ID, VFL session ID, feature IDs of the features involved in the model, and an indication of the model's use by the VFL client.

[0071] In step 515, AnLF402A, acting as the VFL inference server, receives a request or subscription for analysis or statistics on the analysis ID. In step 516, AnLF, acting as the VFL inference server, queries NRF314 for the VFL training server using the analysis ID and its own vendor ID. In step 517, NRF provides profiles of candidate NFs matching the discovery request, and AnLF, acting as the VFL inference server, selects one such NF. In step 518, AnLF, acting as the VFL inference server, requests the model from the selected MTLF404A, acting as the VFL training server, providing the analysis ID.

[0072] like Figure 5BAs shown, the MTLF404A, acting as the VFL training server, can retrieve the requested model from the ADRF in step 519. The ADRF provides a VFL session identifier and feature IDs for all features used to train the model, along with the model. In step 520, the MTLF acting as the VFL training server provides information related to the requested model it is training (the model or information on how to retrieve it). The VFL training server also provides a VFL session identifier and feature IDs for all features used to train the model. If the MTLF acting as the VFL training server provides information on how to retrieve the model, the AnLF402A, acting as the VFL inference server, retrieves the model, as shown in step 521.

[0073] For each feature, at step 522, AnLF402A, acting as the VFL inference server, queries NRF314 for either AnLF402B, acting as the VFL inference client, or AF312, acting as the VFL client, using the analysis ID and the associated feature ID. For each feature, NRF provides profiles of candidate NFs matching the discovery request at step 523, and AnLF, acting as the VFL inference server, selects one such NF. For the selected AnLF402B acting as the VFL inference client, AnLF, acting as the VFL inference server, sends a VFL inference request or subscription at step 524, providing the analysis ID, VFL session ID, and feature ID within the request.

[0074] At step 525, each AnLF402B acting as a VFL inference client queries the NRF314 for the MTLF404B acting as a VFL training client, using its analysis ID and feature ID. For each AnLF acting as a VFL inference client, the NRF provides a profile of a candidate NF matching the discovery request in step 526, and the AnLF acting as a VFL inference client selects one such NF. At step 527, each AnLF acting as a VFL inference client requests a model from the selected MTLF acting as a VFL training client, providing its analysis ID, feature ID, and VFL session identifier.

[0075] Each MTLF404B acting as a VFL training client can retrieve the requested model from the ADRF at step 528, thereby providing the analysis ID, feature ID, and VFL session identifier. At step 529, each MTLF acting as a VFL training client provides information related to the requested model it has trained (the model or information on how to retrieve it). If the MTLF acting as a VFL training client provides information on how to retrieve the model, then the AnLF402B acting as a VFL inference client retrieves the model, as shown in step 530.

[0076] In step 531, each AnLF402B acting as a VFL inference client retrieves the input data and uses the retrieved model of the feature to compute output data associated with the corresponding feature and analysis ID. In step 532, each AnLF402B acting as a VFL inference client provides the output data to the AnLF402A acting as a VFL inference server.

[0077] For each AF312 acting as a VFL client selected in step 523, the AnLF402A acting as a VFL inference server sends a VFL inference request or subscription in step 533, providing the analytics ID, VFL session ID, and feature ID within the request. At step 534, each AF acting as a VFL client selects an applicable model for inference based on the analytics ID, VFL session ID, and feature ID, retrieves the input data, and uses the selected model to compute the output data. At step 535, each AF acting as a VFL client provides the output data to the AnLF acting as a VFL inference server.

[0078] In step 536, the AnLF402A, acting as the VFL inference server, uses the retrieved model and the received feature-related output data to perform calculations, analyses, or predictions.

[0079] Figure 6 Signaling diagram 600 illustrates VFL sample alignment of user equipment samples according to some example implementations. As shown in step 601, before starting vertical federated learning, the MTLF 404A, acting as the VFL training server, determines an initial list of target UEs to be used as samples for vertical federated learning. For each candidate UE, in step 602, the MTLF acting as the VFL training server queries at UDM 316 whether an AMF has been assigned to the UE. In step 603, the UDM returns the assigned AMFID (if any). If no AMF has been assigned to the UE, then in step 604, the MTLF acting as the VFL training server removes the UE from the list of target UEs.

[0080] In steps 605 and 606, the MTLF uses NRF314 to discover the AMF address. And in steps 607 and 608, the MTLF queries the UE location at the AMF.

[0081] If the UE is outside the desired target area, then at step 609, the MTLF404A, which acts as the VFL training server, removes the UE from the list of target UEs.

[0082] For each AF312 acting as a VFL training client, in step 610, the MTLF404A acting as a VFL training server sends a request to the AF to check the availability of the target UE and provide a list of the target UEs.

[0083] In step 611, each AF 312 checks whether each target UE can retrieve relevant input data; otherwise, the UE is removed from the target UE list. At step 612, each AF responds with a simplified list of target UEs. As shown in step 613, the MTLF404A, acting as the VFL training server, calculates the intersection of the simplified lists of target UEs provided by the AFs and uses the resulting intersection list as the target UEs to be used as samples for training the VFL model.

[0084] Figure 7A , 7B Figure 700 illustrates a signaling diagram for VFL training and inference according to other example implementations. Figure 7A As shown, at step 701, AnLF402A, acting as an active VFL inference participant, registers its ability to act as an active VFL inference participant, along with its supported analytics ID and vendor ID, in NRF314. At step 702, AnLF402B, acting as a passive VFL inference participant, registers its ability to act as a passive VFL inference participant, along with its supported analytics ID and associated feature ID, in NRF314. At step 703, MTLF404B, acting as a passive VFL training participant, registers its ability to act as a passive VFL training participant, along with its supported analytics ID and associated feature ID, in NRF314. And based on the configured information, NEF308, on behalf of AF312, registers at step 704 the ability of AF to act as a passive VFL participant in NRF314, along with the analytics ID and associated feature ID supported by AF.

[0085] In step 705, the MTLF404A, acting as an active participant in VFL training, registers its ability to act as an active participant in VFL training and the multiple analytics IDs it supports in the NRF314. At step 706, the NWDAF310, acting as a VFL coordinator, receives or observes a trigger to begin VFL model training for the analytics ID.

[0086] At step 707, the NWDAF 310, acting as the VFL coordinator, selects the desired features for model training based on the implementation. At step 708, the NWDAF, acting as the VFL coordinator, sends a discovery request to the NRF 314 for MTLF404A, which acts as an active participant in VFL training, providing its analytics ID and vendor ID. At step 709, the NRF provides profiles of candidate NFs matching the discovery request, and the NWDAF, acting as the VFL coordinator server, selects one of these NFs.

[0087] For each feature, in step 710, the NWDAF310, acting as the VFL coordinator, sends a discovery request to the NRF314 for either the MTLF404B, which acts as a passive participant in VFL training, or the AF312, which acts as a passive participant in VFL, providing the analysis ID and feature ID. For each feature, in step 711, the NRF provides profiles of candidate NFs that match the discovery request, and the NWDAF, acting as the VFL coordinator server, selects one such NF.

[0088] The NWDAF310, acting as the VFL coordinator, can perform sample alignment with the VFL training client at step 712. If the UE is used as a sample, then... Figure 8 The process in.

[0089] In step 713, the NWDAF310, acting as the VFL coordinator, assigns a unique VFL session ID. In step 714, the NWDAF, acting as the VFL coordinator, sends a request to the active VFL training participants to begin VFL training, providing the analytics ID, feature ID, and VFL session ID. It may also provide the VFL training passive participant (multiple) IDs and / or the AF312 VFL passive participant ID.

[0090] In step 715, the NWDAF310, acting as the VFL coordinator, sends a request to the passive VFL training participant for each feature, providing the analysis ID, feature ID, and VFL session ID. In an alternative example implementation, the active VFL participant sends this request. In step 716, the NWDAF, acting as the VFL coordinator, sends a request to the passive VFL participant AF312 for each feature, providing the analysis ID, feature ID, and VFL session ID. In an alternative example implementation, the active VFL participant sends this request.

[0091] like Figure 7B As shown, at step 717, MTLF404A, acting as an active participant in VFL training, and (multiple) MTLF404B and / or AF312VFL (multiple) passive participants, perform VFL model training. The VFL training server trains a model capable of combining information related to several features. Each VFL training client trains a model related to a specific feature. These models are trained together over several epochs. After each epoch, each VFL training client provides a so-called "gradient" to the VFL training server as an intermediate result, which the server uses to update its model. The server also provides intermediate results to each VFL training client, which the client uses to update its feature-related model.

[0092] In step 718, the MTLF404A acting as an active participant in VFL training stores its model trained in ADRF along with the analytics ID, VFL session ID, and feature ID of all features used to train the model, as well as an indication that the model is for a VFL active participant. At step 719, each MTLF404B acting as a passive participant in VFL training stores its model trained in ADRF along with the analytics ID, VFL session ID, feature ID of the features involved in the model, and an indication that the model is used by a VFL client.

[0093] In step 720, the NWDAF310, acting as the VFL coordinator, receives requests or subscriptions for analysis or statistics of the analysis ID. In step 721, NWDAF310, acting as the VFL coordinator, queries NRF314 for active VFL inference participants using the analytics ID and its own vendor ID. In step 722, NRF provides profiles of candidate NFs that match the discovery request, and NWDAF, acting as the VFL coordinator, selects one of these NFs.

[0094] For each feature, in step 723, the NWDAF 310, acting as the VFL coordinator, queries the NRF 314 for passive participants in the VFL inference process using the analysis ID and feature ID. In step 724, the NRF provides profiles of candidate NFs that match the discovery request, and the NWDAF, acting as the VFL coordinator, selects one such NF for each feature.

[0095] In step 725, the NWDAF310, acting as the VFL coordinator, sends a request or subscription for VFL inference to the AnLF402A, which acts as an active participant in VFL inference, providing the analytics ID and VFL session ID. It may also provide the VFL training active participant ID, VFL inference passive participant (multiple) IDs, and / or the AFVFL passive participant ID.

[0096] In step 726, AnLF402A, acting as an active participant in VFL inference, requests a model from MTLF404A, which acts as an active participant in VFL training, and provides an analytics ID. MTLF, acting as an active participant in VFL training, can retrieve the requested model from the ADRF in step 727. The ADRF provides a VFL session identifier and feature IDs for all features used to train the model, along with the model. In step 728, MTLF, acting as an active participant in VFL training, provides information related to the requested model it is training (the model or information on how to retrieve it). The active participant in VFL training also provides a VFL session identifier and feature IDs for all features used to train the model.

[0097] like Figure 7CAs shown, if MTLF404A, which acts as an active participant in VFL training, provides information on how to retrieve the model, then in step 729, AnLF402A, which acts as an active participant in VFL inference, retrieves the model.

[0098] For each feature, in step 730, the NWDAF310, acting as the VFL coordinator, sends a request or subscription for VFL inference to the AnLF402B, which acts as a passive participant in VFL inference, providing the analytics ID, VFL session ID, and feature ID. It may also provide the associated VFL training passive participant ID and / or VFL inference active participant ID. In an alternative example implementation, the AnLF402A, acting as an active participant in VFL inference, sends this request.

[0099] For each feature, if no VFL training passive participant is received in step 730, then in step 731, AnLF402B, acting as a VFL inference passive participant, queries NRF314 for MTLF404B, which is also acting as a VFL training passive participant, using the analysis ID and the relevant feature ID. For each feature, NRF provides a profile of a candidate NF that matches the discovery request in step 732, and AnLF402B, acting as a VFL inference passive participant, selects one of such NFs.

[0100] At step 733, each AnLF402B acting as a passive participant in VFL inference requests a model from the selected MTLF404B acting as a passive participant in VFL training, providing the analytics ID, feature ID, and VFL session identifier. Each MTLF acting as a passive participant in VFL training may retrieve the requested model from the ADRF at step 734, providing the analytics ID, feature ID, and VFL session identifier. At step 735, each MTLF acting as a passive participant in VFL training provides information related to the requested model it trained (the model or information on how to retrieve it).

[0101] If MTLF404B, acting as a passive participant in VFL training, provides information on how to retrieve the model, then AnLF402B, acting as a passive participant in VFL inference, retrieves the model in step 736.

[0102] In step 737, each AnLF402B acting as a passive participant in VFL inference retrieves input data and uses the retrieval model of the feature to compute output data associated with the corresponding feature and analysis ID.

[0103] In step 738, each AnLF402B acting as a passive participant in VFL inference provides output data to the AnLF402A acting as an active participant in VFL inference.

[0104] For each AF312 acting as a passive VFL participant, in step 739, the NWDAF310 acting as the VFL coordinator sends a VFL inference request or subscription, providing the analytics ID, VFL session ID, and feature ID within the request. It may also provide the ID of the AnLF402A acting as an active VFL inference participant. In an alternative example implementation, the AnLF acting as an active VFL inference participant sends this request.

[0105] At step 740, each AF312 acting as a passive participant in the VFL selects an applicable model for inference based on the analytics ID, VFL session ID, and feature ID, retrieves input data, and uses the selected model to compute output data. At step 741, each AF acting as a passive participant in the VFL provides the output data to AnLF, which acts as the VFL inference server.

[0106] In step 742, AnLF, acting as the VFL inference server, uses the retrieved model and the received feature-related output data to compute an analysis or prediction. And in step 743, AnLF, acting as the VFL inference server, provides the computed analysis or prediction to NWDAF310, which acts as the VFL coordinator.

[0107] Figure 8 Signaling diagram 800 shows VFL sample alignment of user equipment samples according to other example implementations. Before starting vertical federated learning, at step 801, the NWDAF310, acting as the VFL coordinator, determines an initial list of target UEs 208 to be used as samples for vertical federated learning.

[0108] For each candidate UE, at step 802, NWDAF 310, acting as the VFL coordinator, queries UDM 316 to determine if AMF 302 has been assigned to the UE. At step 803, UDM returns the assigned AMFID (if any). If no AMF has been assigned to UE 208, at step 804, NWDAF, acting as the VFL coordinator, removes the UE from the list of target UEs.

[0109] In steps 805 and 806, the NWDAF 310, acting as the VFL coordinator, discovers the AMF address using the NRF 314. In steps 807 and 808, the NWDAF acting as the VFL coordinator queries the location of UE208 at AMF 302. If the UE is outside the desired target area, the NWDAF acting as the VFL coordinator removes the UE from the list of target UEs in step 809.

[0110] For each AF312 acting as a VFL training client, in step 810, the NWDAF310 acting as the VFL coordinator sends a request to the AF to check the availability of the target UE208 and provide a list of target UEs. In step 811, each AF checks whether each target UE can retrieve the relevant input data; otherwise, it removes the UE from the target UE list. In step 812, each AF replies with a simplified list of target UEs.

[0111] At step 813, NWDAF310, acting as the VFL coordinator, calculates the intersection of the reduced list of target UEs 208 provided by AF312, and uses the resulting intersection list as the target UEs to be used as samples for training the VFL model.

[0112] Figure 9A-9G This is a flowchart illustrating various steps in a method 900 according to various example implementations. The method includes assigning a vertical federated learning session identifier to a vertical federated learning model training session, such as... Figure 9A As shown in box 902. The method includes selecting network functions of a telecommunications system to perform model training on machine learning models that are in a common sample space but have different feature spaces, as shown in box 904. The method also includes sending a request to each of the network functions to perform model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces. In some of these examples, the request includes a vertical federated learning session identifier to store the machine learning model in association with the vertical federated learning session identifier, as shown in box 906.

[0113] In some examples, method 900 also includes receiving intermediate output data from network functions based on model training, such as Figure 9B As shown in box 908. In some of these examples, the method also includes training a federated machine learning model to compute predictions based on intermediate output data, as shown in box 910. And the method includes storing the federated machine learning model in association with a vertical federated learning session identifier, as shown in box 912.

[0114] In some examples, different feature spaces are associated with corresponding feature identifiers. In some of these examples, at box 912, the federated machine learning model is also associated with the corresponding feature identifiers stored in relation to different feature spaces.

[0115] In some examples, storing the federated machine learning model at box 912 involves sending a request to the analytics data repository function to store the federated machine learning model within that function, such as... Figure 9C As shown in box 914.

[0116] In some examples, method 900 also includes receiving requests for federated machine learning models, such as Figure 9D As shown in box 916. In some of these examples, the method also includes sending at least one of the following: a federated machine learning model, information for retrieving the federated machine learning model, a vertical federated learning session identifier, an analysis identifier, or a corresponding feature identifier associated with a different feature space, as shown in box 918.

[0117] In some examples, different feature spaces are associated with corresponding feature identifiers; and in some of these examples, the request sent to each network function at box 906 also includes the feature identifier in the corresponding feature identifier.

[0118] In some examples, the federated learning session identifier is associated with the analysis identifier; and in some of these examples, the request sent to each network function at box 906 also includes the analysis identifier.

[0119] In some examples, method 900 also includes sending a request to a network repository function for information related to at least one network function that supports model training for feature identifiers and analytics identifiers, such as... Figure 9E As shown in box 920. In some of these examples, the method also includes receiving information from a network repository function, wherein the network function within the network function is selected based on the information received from the network repository function, as shown in box 922.

[0120] In some examples, method 900 is performed by a network function. In some of these examples, the method also includes sending a request to a network repository function to store information related to the network function, wherein the request includes at least one of the following: the network function can act as an instruction for the operation of a vertical federated learning training server; the network function can act as an instruction for the operation of an active participant in vertical federated learning training; an analytics identifier; or an identifier of the network function's vendor, such as... Figure 9F As shown in box 924.

[0121] In some examples, method 900 also includes selecting network features to act as active participants in the training of machine learning models, such as... Figure 9G As shown in box 926. In some of these examples, the method also includes sending a request to the network function to utilize the network function as an active participant to perform model training, the request including a vertical federated learning session identifier and an analytics identifier, as shown in box 928.

[0122] In some examples, the method is executed by the network data analysis function, model training logic function, or application function of the telecommunications system.

[0123] Figures 10A to 10D This is a flowchart illustrating various steps in method 1000 according to various example implementations. The method includes receiving a request from a network function of a telecommunications system for performing model training on a machine learning model, which will be one of machine learning models trained in a common sample space but with different feature spaces. The request includes a vertical federated learning session identifier, such as... Figure 10A As shown in box 1002. The method includes training a machine learning model in a common sample space using corresponding feature spaces in different feature spaces, the machine learning model being trained to determine intermediate output data, as shown in box 1004. The method includes sending the intermediate data to a network function to train a federated machine learning model based on the intermediate output data to compute predictions, as shown in box 1006. Furthermore, the method includes storing the machine learning model in association with a vertical federated learning session identifier, as shown in box 1008.

[0124] In some examples, the request received from the network function at box 1002 includes feature identifiers associated with the corresponding feature space, which the machine learning model is trained using.

[0125] In some examples, the federated machine learning model at box 1008 is further stored in association with feature identifiers.

[0126] In some examples, storing the machine learning model at box 1008 involves sending a request to the analytics data repository function to store the machine learning model at the analytics data repository function, such as... Figure 10B The message is sent at box 1010 as shown.

[0127] In some examples, the federated learning session identifier is associated with the analysis identifier, and in some of these examples, the request also includes the analysis identifier.

[0128] In some examples, method 1000 also includes receiving a request for a machine learning model, which includes a vertical federated learning session identifier and feature identifiers associated with the corresponding feature space, such as... Figure 10C As shown in box 1012. In some of these examples, the method also includes providing a response that includes at least one of a machine learning model or information for retrieving a machine learning model, as shown in box 1014.

[0129] In some examples, method 1000 is performed by a second network function, and in some of these examples, the method further includes sending a request to a network repository function to store information related to the second network function, wherein the request includes at least one of the following: instructions regarding the second network function's operability as a vertical federated learning training client, instructions regarding the second network function's operability as a vertical federated learning training passive participant, an analysis identifier, or a feature identifier associated with a corresponding feature space in which a machine learning model is trained, such as... Figure 1 As shown in box 1016 of D.

[0130] In some examples, the method is implemented by the network data analysis function, model training logic function, or application function of the telecommunications system.

[0131] Figures 11A to 11D This is a flowchart illustrating various steps in method 1100 according to various example implementations. The method includes receiving a request for at least one of the following: providing analysis or one or more predictions of a data analysis process associated with an analysis identifier, such as... Figure 11A As shown in box 1102. The method includes selecting a network function to provide information about a vertical federated learning session, which is associated with an analytics identifier, as shown in box 1104. The method includes sending a request to the network function, the request including the analytics identifier, to the network function for use with a federated machine learning model, as shown in box 1106. The method includes receiving a response from the network function, the response including a vertical federated learning session identifier and information about the federated machine learning model, as shown in box 1108.

[0132] Method 1100 includes accessing a federated machine learning model associated with a vertical federated learning session identifier, as shown in box 1110. The method includes sending a request to a network function of a telecommunications system for performing inference using a machine learning model trained in a common sample space but with a different feature space. The request sent to the network function includes a vertical federated learning session identifier and an analytics identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analytics identifier, as shown in box 1112. The method includes receiving intermediate output data from the network function based on the inference, as shown in box 1114. Furthermore, the method includes performing inference based on the intermediate output data and using the federated machine learning model to compute a prediction, as shown in box 1116.

[0133] In some examples, accessing the federated machine learning model in box 1110 includes: accessing the federated machine learning model and the corresponding feature identifiers associated with different feature spaces, such as Figure 11B As shown in box 1118. In some of these examples, method 1100 also includes selecting network functions based on corresponding feature identifiers, as shown in box 1120.

[0134] In some examples, each request to the corresponding network function sent to the network function at box 1122 also includes a feature identifier in the corresponding feature identifier.

[0135] In some examples, accessing a federated machine learning model at box 1110 includes sending a request to at least one of the following: a vertical federated learning training server, a vertical federated learning training active participant, a model training logic function, or an analytics data repository function, for retrieving the federated machine learning model, such as... Figure 11C As shown in box 1124.

[0136] In some examples, method 1100 also includes sending a request to a network repository function for information related to at least one network function that provides information about a vertical federated learning session. The request includes at least one of an analytics identifier or a vendor identifier for at least one network function, such as... Figure 11D As shown in box 1126. In some of these examples, the method also includes receiving information from a network repository function, whereby the network function used to provide information about the vertical federated learning session is selected based on the information received from the network repository function, as shown in box 1128.

[0137] In some examples, the method is executed by the network data analysis function, analysis logic function, or application function of the telecommunications system.

[0138] Figures 12A to 12D This is a flowchart illustrating various steps in method 1200 according to various example implementations. The method includes receiving a request from a network function of a telecommunications system for performing inference using a machine learning model, one of several machine learning models trained in a common sample space but with different feature spaces. The request includes a vertical federated learning session identifier, such as... Figure 12A As shown in box 1202. The method includes accessing a machine learning model associated with a vertical federated learning session identifier, as shown in box 1204. The method includes performing inference using the machine learning model to compute intermediate output data, as shown in box 1206. And the method includes sending the intermediate data to a network function to perform inference using the federated machine learning model to compute predictions based on the intermediate output data, as shown in box 1208.

[0139] In some examples, the request received from the network function at box 1202 includes a feature identifier, which is associated with one of the different feature spaces used to train a machine learning model.

[0140] In some examples, the machine learning model is further accessed based on feature identifiers.

[0141] In some examples, accessing a machine learning model at box 1204 includes sending a request to at least one of a vertical federated learning training client, a vertical federated learning training passive participant, a model training logic function, or an analytics data repository function, for storing the machine learning model, such as... Figure 12B As shown in box 1210.

[0142] In some examples, the vertical federated learning session identifier is associated with the analysis identifier, and the request received from the network function at box 1202 also includes the analysis identifier.

[0143] In some examples, at box 1204, the machine learning model is accessed from a vertical federated learning training client. In some of these examples, method 1200 also includes sending a request to a network repository function for information related to at least one network function that supports model training for feature identifiers and analytics identifiers, such as... Figure 12C As shown in box 1212. The method includes a component for receiving information from a network repository function, as shown in box 1214. Furthermore, the method includes selecting a vertical federated learning training client based on the information received from the network repository function, as shown in box 1216.

[0144] In some examples, method 1200 is performed by a second network function. In some of these examples, the method further includes sending a request to a network repository function to store information related to the second network function, wherein the request includes at least one of the following: instructions regarding the second network function's operability as a vertical federated learning inference client, an analytics identifier, or a feature identifier associated with a feature space in a different feature space used to train a machine learning model, such as... Figure 12D As shown in box 1218.

[0145] In some examples, the method is executed by the network data analysis function, analysis logic function, or application function of the telecommunications system.

[0146] According to the example implementations of this disclosure, the telecommunications system 100 or PLMN 102 and its components (such as UE 110, CN 106, RAN 108, 5GC 202, NG-RAN 204, gNB 206, UE 208, AMF 302, SMF 304, UPF 306, NEF 308, NWDAF 310, AF 312, NRF 314, AnLF 402 and / or MTLF 404) can be implemented by various means. Means for implementing the system and its components may include hardware, firmware, software, or combinations thereof. In some instances, one or more devices may be configured to act as or otherwise implement the system and its components shown and described herein. In examples involving more than one device, the respective devices may be connected to or otherwise communicate with each other in a variety of different ways, such as directly or indirectly via wired or wireless networks.

[0147] According to some example implementations, regarding Figure 9A-9G At least some of the described methods 900 can be performed by means including components for performing functions corresponding to the steps of the method. Similarly, regarding Figure 10A-10D At least some of the described methods 1000 can be performed by means including components for performing functions corresponding to the steps of the method. Regarding Figure 11A-11D At least some of the methods 1100 described can be performed by means including components for performing functions corresponding to the steps of the method; and regarding Figure 12A-12D At least some of the methods 1200 described can be performed by means including components for performing functions corresponding to the steps of the method. Examples of suitable means may include NF (e.g., AMF, SMF, UPF, NEF, NWDAF, AF, AS), AnLF, MTLF, or any suitable means such as a server, host, or node.

[0148] Figure 13 An apparatus 1300 according to some exemplary embodiments of the present disclosure is shown, wherein components for performing various functions include separate hardware or hardware directed by one or more computer programs from a computer-readable storage medium or other memory, such as computer memory. The apparatus may include one or more of each of a plurality of components, such as processing circuitry 1302 connected to a computer-readable storage medium or other memory 1304.

[0149] The processing circuitry 1302 may consist of one or more processors individually or in combination with one or more computer-readable storage media. The processing circuitry is generally any computer hardware capable of processing information such as data, computer programs, and / or other suitable electronic information. The processing circuitry consists of a collection of electronic circuits, some of which may be packaged as integrated circuits or multiple interconnected integrated circuits (integrated circuits are sometimes more commonly referred to as "chips"). The processing circuitry may be configured to execute computer programs, which may be stored on the processing circuitry or otherwise stored in memory 1304 (of the same or another device).

[0150] Depending on the specific implementation, the processing circuit 1302 may be multiple processors, a multi-core processor, or some other type of processor. Furthermore, the processing circuit may be implemented using multiple heterogeneous processor systems, where a main processor and one or more auxiliary processors reside on a single chip. As another illustrative example, the processing circuit may be a symmetric multiprocessor system comprising multiple processors of the same type. In yet another example, the processing circuit may be embodied as or otherwise include one or more ASICs, FPGAs, etc. Therefore, while the processing circuit may be able to execute a computer program to perform one or more functions, various examples of processing circuits may be able to perform one or more functions without the assistance of a computer program. In any case, the processing circuit may be appropriately programmed to perform functions or operations according to exemplary embodiments of this disclosure.

[0151] Memory 1304 is typically any computer hardware capable of temporarily and / or permanently storing information (e.g., data, computer programs, instructions 1306 (e.g., computer-readable program code), and / or other suitable information). Memory may include volatile and / or non-volatile memory and may be fixed or removable. Examples of suitable memory include recording media, random access memory (RAM), read-only memory (ROM), hard disk drives, flash memory, thumb drives, removable computer disks, optical disks, or some combination thereof.

[0152] Memory 1304 is a non-transitory device capable of storing information. An example of a suitable memory is a computer-readable storage medium, distinguishable from a computer-readable transport medium capable of carrying information from one location to another. Examples of suitable computer-readable transport media include electronic carrier signals, telecommunication signals, software distribution packages, or some combination thereof. As used herein, the term "non-transitory" is a limitation on the medium itself (i.e., tangible, not signaling), not a limitation on the persistence of data storage (e.g., RAM and ROM). As used herein, a computer-readable medium generally refers to a computer-readable storage medium or a computer-readable transport medium. A computer-readable medium is any entity or device capable of storing and carrying information such as one or more computer programs or portions thereof.

[0153] In addition to memory 1304 (e.g., a computer-readable storage medium), processing circuitry 1302 may also be connected to one or more interfaces for displaying, transmitting, and / or receiving information. Interfaces may include communication interface 1308 and / or one or more user interfaces (e.g., a display, a user input interface). Communication interfaces may be configured to transmit and / or receive information, such as to and / or from other devices(s), networks(s), etc. Communication interfaces may be configured to transmit and / or receive information via physical (wired) and / or wireless communication links. Examples of suitable communication interfaces include network interface controllers (NICs), wireless NICs (WNICs), etc.

[0154] The combination of operations supporting the implementation of exemplary embodiments of the present disclosure is supported by the processing circuitry 1302 executing instruction 1306 or storing instructions in memory 1304. In this way, apparatus 1300 may include at least one processing circuit and at least one memory coupled to the at least one processing circuit, wherein the at least one processing circuit is configured to execute instructions stored in the at least one memory. It will also be understood that one or more functions, and combinations thereof, may be implemented by a dedicated hardware-based computer system and / or processing circuitry, or a combination of dedicated hardware and program code instructions, performing the specified functions.

[0155] Some exemplary embodiments of this disclosure can also be performed as a computer process defined by one or more computer programs or portions thereof. Exemplary embodiments of this disclosure can be performed by executing at least one portion of a computer program including instructions. The computer program can be in source code form, object code form, or some intermediate form. The computer program can be stored in a computer-readable medium readable by a computer, processing circuitry, or other suitable means. As mentioned above, for example, the computer program can be stored in memory such as a computer-readable storage medium. Additionally or alternatively, for example, the computer program can be stored in a computer-readable transmission medium. The coding of software used to perform exemplary embodiments of this disclosure is entirely within the scope of those skilled in the art.

[0156] As will be understood, any suitable instructions may be loaded from memory or a computer-readable medium (e.g., a computer-readable storage medium, a computer-readable transmission medium) onto a computer, processing circuitry, or other programmable means to produce a particular machine, such that the particular machine becomes a component for implementing the functions specified herein. Instructions may also be stored in a computer-readable medium that can direct a computer, processing circuitry, or other programmable means to operate in a particular manner to produce a particular machine or a particular article of manufacture. In some examples, instructions stored in a computer-readable medium can produce an article of manufacture, wherein the article of manufacture becomes a component for implementing the functions described herein. Instructions may be retrieved from a computer-readable medium and loaded onto a computer, processing circuitry, or other programmable means to configure the computer, processing circuitry, or other programmable means to perform operations to be performed on or by the computer, processing circuitry, or other programmable means.

[0157] The retrieval, loading, and execution of instructions, including program code instructions, can be performed sequentially, such that one instruction is retrieved, loaded, and executed at a time. In some example implementations, retrieval, loading, and / or execution can be performed in parallel, such that multiple instructions are retrieved, loaded, and / or executed together. The execution of program code instructions can produce computer-implemented processes, such that the instructions, executed by a computer, processing circuitry, or other programmable device, provide operations for implementing the functions described herein.

[0158] As stated above and reiterated below, this disclosure includes, but is not limited to, the following example implementations.

[0159] Clause 1. An apparatus comprising: at least one memory configured to store instructions; and at least one processing circuit configured to access the at least one memory and execute the instructions such that the apparatus at least: assigns a vertical federated learning session identifier to a vertical federated learning model training session; selects a network function of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space; and sends a request to each of the network functions for performing model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier such that the machine learning model is stored in association with the vertical federated learning session identifier.

[0160] Clause 2. The apparatus according to Clause 1, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further at least: receive intermediate output data from network functions based on model training; train a federated machine learning model to compute predictions based on the intermediate output data; and store the federated machine learning model in association with a vertical federated learning session identifier.

[0161] Clause 3. The apparatus according to Clause 2, wherein different feature spaces are associated with corresponding feature identifiers, and wherein the federated machine learning model is also stored in association with the corresponding feature identifiers associated with the different feature spaces.

[0162] Clause 4. The apparatus according to Clause 2 or Clause 3, wherein storing a federated machine learning model is caused by: causing the apparatus to send a request to an analytics data repository function for storing the federated machine learning model at the analytics data repository function.

[0163] Clause 5. An apparatus according to any one of Clauses 2 to 4, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further at least: receive a request for a federated machine learning model; and send at least one of the following: a federated machine learning model, information for retrieving the federated machine learning model, a vertical federated learning session identifier, an analysis identifier, or a corresponding feature identifier associated with a different feature space.

[0164] Clause 6. An apparatus according to any one of Clauses 1 to 5, wherein different feature spaces are associated with corresponding feature identifiers, and wherein a request sent to each network function in the network functions also includes a feature identifier in the corresponding feature identifier.

[0165] Clause 7. An apparatus according to any one of Clauses 1 to 6, wherein a federated learning session identifier is associated with an analysis identifier, and wherein a request sent to each of the network functions also includes the analysis identifier.

[0166] Clause 8. An apparatus according to any one of Clauses 1 to 7, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further at least: send a request to a network repository function requesting information relating to at least one network function, the at least one network function supporting model training for feature identifiers and analysis identifiers; and receive information from the network repository function, wherein the network function among the network functions is selected based on the information received from the network repository function.

[0167] Clause 9. An apparatus according to any one of Clauses 1 to 8, wherein the apparatus is implemented by a network function, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further send a request to a network repository function to store information related to the network function, and wherein the request includes at least one of the following: an instruction that the network function is operable as a vertical federated learning training server, an instruction that the network function is operable as an active participant in vertical federated learning training, an analytics identifier, or an identifier of the network function's vendor.

[0168] Clause 10. An apparatus according to any one of Clauses 1 to 9, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further at least: select a network function to operate as an active participant in model training of a machine learning model; and send a request to the network function to perform model training using the network function as an active participant, the request including a vertical federated learning session identifier and an analytics identifier.

[0169] Clause 11. An apparatus according to any one of Clauses 1 to 10, wherein the apparatus is implemented by network data analysis functions, model training logic functions, or application functions of a telecommunications system.

[0170] Clause 12. An apparatus comprising: components for assigning a vertical federated learning session identifier to a vertical federated learning model training session; components for selecting a network function of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space; and components for sending a request to each network function for performing model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier such that the machine learning model is stored in association with the vertical federated learning session identifier.

[0171] Clause 13. The apparatus according to Clause 12, wherein the apparatus further comprises: a component for receiving intermediate output data from network functions based on model training; a component for training a federated machine learning model to compute predictions based on the intermediate output data; and a component for storing the federated machine learning model in association with a vertical federated learning session identifier.

[0172] Clause 14. The apparatus according to Clause 13, wherein different feature spaces are associated with corresponding feature identifiers, and wherein the federated machine learning model is also stored in association with the corresponding feature identifiers associated with the different feature spaces.

[0173] Clause 15. The apparatus pursuant to Clause 13 or Clause 14, wherein the component for storing the federated machine learning model includes a component for: sending a request to the analytics data repository function for storing the federated machine learning model at the analytics data repository function.

[0174] Clause 16. An apparatus according to any one of Clauses 13 to 15, wherein the apparatus further comprises: a component for receiving a request for a federated machine learning model; and a component for transmitting at least one of: a federated machine learning model, information for retrieving the federated machine learning model, a vertical federated learning session identifier, an analysis identifier, or a corresponding feature identifier associated with a different feature space.

[0175] Clause 17. An apparatus according to any one of Clauses 12 to 16, wherein different feature spaces are associated with corresponding feature identifiers, and wherein a request sent to each network function in the network functions also includes a feature identifier in the corresponding feature identifier.

[0176] Clause 18. An apparatus according to any one of Clauses 12 to 17, wherein a federated learning session identifier is associated with an analysis identifier, and wherein a request sent to each of the network functions also includes the analysis identifier.

[0177] Clause 19. An apparatus according to any one of Clauses 12 to 18, wherein the apparatus further comprises: a component for sending a request to a network repository function for information relating to at least one network function, the at least one network function supporting model training for feature identifiers and analysis identifiers; and a component for receiving information from the network repository function, wherein the network function among the network functions is selected based on the information received from the network repository function.

[0178] Clause 20. An apparatus according to any one of Clauses 12 to 19, wherein the apparatus is implemented by a network function, wherein the apparatus further includes components for: sending a request to a network repository function to store information related to the network function, and wherein the request includes at least one of: an instruction that the network function is operable as a vertical federated learning training server, an instruction that the network function is operable as an active participant in vertical federated learning training, an analytics identifier, or an identifier of the network function's vendor.

[0179] Clause 21. An apparatus according to any one of Clauses 12 to 20, wherein the apparatus further comprises: components for selecting a network function to operate as an active participant in model training of a machine learning model; and components for sending a request to the network function to perform model training using the network function as an active participant, the request including a vertical federated learning session identifier and an analytics identifier.

[0180] Clause 22. An apparatus according to any one of Clauses 12 to 21, wherein the apparatus is implemented by network data analysis functions, model training logic functions, or application functions of a telecommunications system.

[0181] Clause 23. A method comprising: assigning a vertical federated learning session identifier to a vertical federated learning model training session; selecting a network function of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space; and sending a request to each network function for performing model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier such that the machine learning model is stored in association with the vertical federated learning session identifier.

[0182] Clause 24. The method according to Clause 23, wherein the method further includes: receiving intermediate output data from network functions based on model training; training a federated machine learning model to compute predictions based on the intermediate output data; and storing the federated machine learning model in association with a vertical federated learning session identifier.

[0183] Clause 25. The method according to Clause 24, wherein different feature spaces are associated with corresponding feature identifiers, and wherein the federated machine learning model is also stored in association with the corresponding feature identifiers associated with the different feature spaces.

[0184] Clause 26. The method according to Clause 24 or Clause 25, wherein storing the federated machine learning model includes: causing the device to send a request to the analytics data repository function for storing the federated machine learning model at the analytics data repository function.

[0185] Clause 27. The method of any one of Clauses 24 to 26, wherein the method further comprises: receiving a request for a federated machine learning model; and sending at least one of the following: a federated machine learning model, information for retrieving the federated machine learning model, a vertical federated learning session identifier, an analysis identifier, or a corresponding feature identifier associated with a different feature space.

[0186] Clause 28. The method according to any one of Clauses 23 to 27, wherein different feature spaces are associated with corresponding feature identifiers, and wherein a request sent to each network function in the network functions also includes a feature identifier in the corresponding feature identifier.

[0187] Clause 29. The method according to any one of Clauses 23 to 28, wherein the federated learning session identifier is associated with an analysis identifier, and wherein requests sent to each network function in the network functions also include the analysis identifier.

[0188] Clause 30. The method according to any one of Clauses 23 to 29, wherein the method further comprises: sending a request to a network repository function for information relating to at least one network function, the at least one network function supporting model training for feature identifiers and analysis identifiers; and receiving information from the network repository function, wherein the network function among the network functions is selected based on the information received from the network repository function.

[0189] Clause 31. The method according to any one of Clauses 23 to 30, wherein the method is performed by a network function, wherein the method further includes sending a request to a network repository function to store information related to the network function, and wherein the request includes at least one of the following: instructions that the network function is operable as a vertical federated learning training server, instructions that the network function is operable as an active participant in vertical federated learning training, an analytics identifier, or an identifier of the network function's vendor.

[0190] Clause 32. The method according to any one of Clauses 23 to 31, wherein the method further comprises: selecting a network function to operate as an active participant in model training of a machine learning model; and sending a request to the network function to perform model training using the network function as an active participant, the request including a vertical federated learning session identifier and an analytics identifier.

[0191] Clause 33. The method according to any one of Clauses 23 to 32, wherein the method is implemented by the network data analysis function, model training logic function, or application function of the telecommunications system.

[0192] Clause 34. A computer-readable storage medium, the computer-readable storage medium being non-transitory and having instructions stored therein, the instructions being responsive to execution by at least one processing circuitry to cause the means to at least: assign a vertical federated learning session identifier to a vertical federated learning model training session; select a network function of a telecommunications system to perform model training on machine learning models having different feature spaces in a common sample space; and send a request to each of the network functions for performing model training on a corresponding machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, wherein the request includes a vertical federated learning session identifier such that the machine learning model is stored in association with the vertical federated learning session identifier.

[0193] Clause 35. A computer-readable storage medium pursuant to Clause 34, wherein the computer-readable storage medium further has instructions stored therein, the instructions being responsive to execution by at least one processing circuit, such that the apparatus further: based on model training, receives intermediate output data from a network function; based on the intermediate output data, trains a federated machine learning model to compute predictions; and stores the federated machine learning model in association with a vertical federated learning session identifier.

[0194] Clause 36. A computer-readable storage medium pursuant to Clause 35, wherein different feature spaces are associated with corresponding feature identifiers, and wherein the federated machine learning model is also stored in association with the corresponding feature identifiers associated with the different feature spaces.

[0195] Clause 37. A computer-readable storage medium pursuant to Clause 35 or Clause 36, wherein the means of storing a federated machine learning model comprises: causing the means to send a request to an analytics data repository function for storing the federated machine learning model at the analytics data repository function.

[0196] Clause 38. A computer-readable storage medium as described in any of Clauses 35 to 37, wherein the computer-readable storage medium stores additional instructions that, in response to execution by at least one processing circuit, cause the apparatus to further at least: receive a request for a federated machine learning model; and send at least one of the following: a federated machine learning model, information for retrieving the federated machine learning model, a vertical federated learning session identifier, an analysis identifier, or a corresponding feature identifier associated with a different feature space.

[0197] Clause 39. A computer-readable storage medium pursuant to any of Clauses 34 to 38, wherein different feature spaces are associated with corresponding feature identifiers, and wherein a request sent to each network function in the network functions also includes a feature identifier in the corresponding feature identifier.

[0198] Clause 40. A computer-readable storage medium pursuant to any of Clauses 34 to 39, wherein different feature spaces are associated with corresponding feature identifiers, and wherein a request sent to each of the network functions further includes a feature identifier in the corresponding feature identifier.

[0199] Clause 41. A computer-readable storage medium according to any one of Clauses 34 to 40, wherein the computer-readable storage medium stores additional instructions that, in response to execution by at least one processing circuit, cause the apparatus to further perform at least the following operations: sending a request to a network repository function for information relating to at least one network function, the at least one network function supporting model training for feature identifiers and analysis identifiers; and receiving information from the network repository function, wherein the network function among the network functions is selected based on the information received from the network repository function.

[0200] Clause 42. A computer-readable storage medium according to any one of Clauses 34 to 41, wherein the apparatus is implemented by a network function, wherein the computer-readable storage medium has additional instructions stored therein, the additional instructions being responsive to execution by at least one processing circuitry to cause the apparatus to further send a request to a network repository function to store information related to the network function, and wherein the request includes at least one of the following: instructions that the network function is operable as a vertical federated learning training server, instructions that the network function is operable as an active participant in vertical federated learning training, an analytics identifier, or an identifier of the network function's vendor.

[0201] Clause 43. A computer-readable storage medium according to any one of Clauses 34 to 42, wherein the computer-readable storage medium has additional instructions stored therein, the additional instructions being responsive to execution by at least one processing circuitry to cause the apparatus to at least: select a network function to operate as an active participant in model training of a machine learning model; and send a request to the network function to perform model training using the network function as an active participant, the request including a vertical federated learning session identifier and an analytics identifier.

[0202] Clause 44. A computer-readable storage medium pursuant to any of Clauses 34 to 43, wherein the apparatus is implemented by network data analysis functions, model training logic functions, or application functions of a telecommunications system.

[0203] Clause 45. An apparatus comprising components for performing the method pursuant to any one of Clauses 23 to 33.

[0204] Clause 46. A computer-readable medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 23 to 33.

[0205] Clause 47. A computer-readable storage medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 23 to 33.

[0206] Clause 48. A computer program comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 23 to 33.

[0207] Clause 49. An apparatus comprising: at least one memory configured to store instructions; and at least one processing circuitry configured to access the at least one memory and execute the instructions to cause the apparatus to at least: receive a request from a network function of a telecommunications system for performing model training of a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having different feature spaces, the request including a vertical federated learning session identifier; train the machine learning model in the common sample space using corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; send the intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and store the machine learning model in association with the vertical federated learning session identifier.

[0208] Clause 50. The apparatus according to Clause 49, wherein the request received from the network function includes a feature identifier associated with a corresponding feature space, and the machine learning model is trained using the corresponding feature space.

[0209] Clause 51. The apparatus according to Clause 50, wherein the federated machine learning model is further stored in association with feature identifiers.

[0210] Clause 52. An apparatus according to any one of Clauses 49 to 51, wherein causing the apparatus to store a machine learning model includes causing the apparatus to send a request to an analytics data repository function to store the machine learning model at the analytics data repository function.

[0211] Clause 53. An apparatus pursuant to any one of Clauses 49 to 52, wherein a federated learning session identifier is associated with an analysis identifier, and wherein requests sent to each of the network functions also include the analysis identifier.

[0212] Clause 54. An apparatus according to any one of Clauses 49 to 53, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further at least: receive a request for a machine learning model, the request including a vertical federated learning session identifier and a feature identifier associated with a corresponding feature space; and provide a response including at least one of a machine learning model or information for retrieving a machine learning model.

[0213] Clause 55. An apparatus according to any one of Clauses 49 to 54, wherein the apparatus is implemented by a second network function, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further send a request to a network repository function to store information related to the network function, and wherein the request includes at least one of the following: an instruction that the network function is operable as a vertical federated learning training server, an instruction that the network function is operable as an active participant in vertical federated learning training, an analytics identifier, or an identifier of the network function's vendor.

[0214] Clause 56. An apparatus according to any one of Clauses 49 to 55, wherein the apparatus is implemented by network data analysis functions, model training logic functions, or application functions of a telecommunications system.

[0215] Clause 57. An apparatus comprising: components for receiving a request from a network function of a telecommunications system for performing model training of a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having different feature spaces, the request including a vertical federated learning session identifier; components for training the machine learning model in the common sample space using corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; components for sending intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and components for storing the machine learning model in association with the vertical federated learning session identifier.

[0216] Clause 58. The apparatus according to Clause 57, wherein the request received from the network function includes a feature identifier associated with a corresponding feature space, and the machine learning model is trained using the corresponding feature space.

[0217] Clause 59. The apparatus pursuant to Clause 58, wherein the federated machine learning model is further stored in association with feature identifiers.

[0218] Clause 60. An apparatus according to any one of Clauses 57 to 59, wherein the component for storing the machine learning model includes a component for: sending a request to an analytics data repository function to store the machine learning model at the analytics data repository function.

[0219] Clause 61. An apparatus pursuant to any of Clauses 57 to 60, wherein a federated learning session identifier is associated with an analysis identifier, and wherein the request further includes the analysis identifier.

[0220] Clause 62. An apparatus according to any one of Clauses 57 to 61, wherein the apparatus further comprises: a component for receiving a request for a machine learning model, the request including a vertical federated learning session identifier and a feature identifier associated with a corresponding feature space; and a component for providing a response, the response including at least one of a machine learning model or information for retrieving a machine learning model. Clause 63. An apparatus according to any one of Clauses 57 to 62, wherein the apparatus is implemented by a second network function, wherein the apparatus further includes components for: sending a request to a network repository function to store information related to the second network function, and wherein the request includes at least one of the following: instructions regarding the second network function being operable as a vertical federated learning training client, instructions regarding the second network function being operable as a vertical federated learning training passive participant, an analysis identifier, or a feature identifier associated with a corresponding feature space in which a machine learning model is trained.

[0221] Clause 64. An apparatus according to any one of Clauses 57 to 63, wherein the apparatus is implemented by network data analysis functions, model training logic functions, or application functions of a telecommunications system.

[0222] Clause 65. A method comprising: receiving a request from a network function of a telecommunications system for performing model training of a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having different feature spaces, the request including a vertical federated learning session identifier; training the machine learning model in the common sample space using the corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; sending the intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and storing the machine learning model in association with the vertical federated learning session identifier.

[0223] Clause 66. The method according to Clause 65, wherein the request received from the network function includes a feature identifier associated with a corresponding feature space, and the machine learning model is trained using the corresponding feature space.

[0224] Clause 67. The method according to Clause 66, wherein the federated machine learning model is further stored in association with feature identifiers.

[0225] Clause 68. The method according to any one of Clauses 65 to 67, wherein storing the machine learning model includes causing the device to send a request to the analytics data repository function to store the machine learning model at the analytics data repository function.

[0226] Clause 69. The method according to any one of Clauses 65 to 68, wherein the federated learning session identifier is associated with an analysis identifier, and wherein requests sent to each network function in the network functions also include the analysis identifier.

[0227] Clause 70. The method of any one of Clauses 65 to 69, wherein the method further comprises: receiving a request for a machine learning model, the request including a vertical federated learning session identifier and a feature identifier associated with a corresponding feature space; and providing a response including at least one of a machine learning model or information for retrieving a machine learning model.

[0228] Clause 71. The method according to any one of Clauses 65 to 70, wherein the method is performed by a second network function, wherein the method further comprises: sending a request to a network repository function to store information related to the second network function, and wherein the request includes at least one of the following: an instruction that the second network function is operable as a vertical federated learning training client, an instruction that the second network function is operable as a vertical federated learning training passive participant, an analysis identifier, or a feature identifier associated with a corresponding feature space used to train a machine learning model.

[0229] Clause 72. The method according to any one of Clauses 65 to 71, wherein the method is implemented by the network data analysis function, or model training logic function, or application function of the telecommunications system.

[0230] Clause 73. A computer-readable storage medium, the computer-readable storage medium being non-transitory and having instructions stored therein, the instructions being responsive to execution by at least one processing circuitry to cause a means to at least: receive a request from a network function of a telecommunications system for performing model training of a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having different feature spaces, the request including a vertical federated learning session identifier; train the machine learning model in the common sample space using corresponding feature spaces in the different feature spaces, the machine learning model being trained to determine intermediate output data; send the intermediate data to the network function to train the federated machine learning model based on the intermediate output data to compute predictions; and store the machine learning model in association with the vertical federated learning session identifier.

[0231] Clause 74. A computer-readable storage medium pursuant to Clause 73, wherein a request received from a network function includes a feature identifier associated with a corresponding feature space, and a machine learning model is trained using the corresponding feature space.

[0232] Clause 75. A computer-readable storage medium pursuant to Clause 74, wherein the federated machine learning model causes further storage in association with feature identifiers.

[0233] Clause 76. A computer-readable storage medium as described in any of Clauses 73 to 75, wherein enabling the device to store a machine learning model includes enabling the device to send a request to an analytics data repository function to store the machine learning model at the analytics data repository function.

[0234] Clause 77. A computer-readable storage medium pursuant to any of Clauses 73 through 76, wherein a federated learning session identifier is associated with an analysis identifier, and wherein the request further includes the analysis identifier.

[0235] Clause 78. A computer-readable storage medium as described in any of Clauses 73 to 77, wherein the computer-readable storage medium has additional instructions stored therein, the additional instructions being responsive to execution by at least one processing circuitry to cause the means to further at least: receive a request for a machine learning model, the request including a vertical federated learning session identifier and a feature identifier associated with a corresponding feature space; and provide a response including at least one of a machine learning model or information for retrieving a machine learning model.

[0236] Clause 79. A computer-readable storage medium according to any one of Clauses 73 to 78, wherein the apparatus is implemented by a second network function, wherein the computer-readable storage medium has additional instructions stored therein, the additional instructions being responsive to execution by at least one processing circuitry to cause the apparatus to further send a request to a network repository function to store information related to the network function, and wherein the request includes at least one of the following: instructions that the network function is operable as a vertical federated learning training server, instructions that the network function is operable as an active participant in vertical federated learning training, an analytics identifier, or an identifier of the network function's vendor.

[0237] Clause 80. A computer-readable storage medium pursuant to any of Clauses 73 to 79, wherein the apparatus is implemented by network data analysis functions, model training logic functions, or application functions of a telecommunications system.

[0238] Clause 81. An apparatus comprising components for performing the method pursuant to any one of Clauses 65 to 72.

[0239] Clause 82. A computer-readable medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 65 to 72.

[0240] Clause 83. A computer-readable storage medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 65 to 72.

[0241] Clause 84. A computer program comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 65 to 72.

[0242] Clause 85. An apparatus comprising: at least one memory configured to store instructions; and at least one processing circuitry configured to access the at least one memory and execute the instructions to cause the apparatus to at least: receive a request for providing at least one of: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; select a network function to provide information about a vertical federated learning session associated with the analysis identifier; send a request to the network function, the request including the analysis identifier, the request being sent to the network function for use with a federated machine learning model; receive a response from the network function, the response including a vertical federated learning session identifier and information about the federated machine learning model; access a federated machine learning model associated with the vertical federated learning session identifier; send a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a common sample space with different feature spaces, the request including the vertical federated learning session identifier and the analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; receive intermediate output data from the network function based on the inference; and perform inference using the federated machine learning model based on the intermediate output data to compute predictions.

[0243] Clause 86. The apparatus according to Clause 85, wherein enabling the apparatus to access the federated machine learning model includes: enabling the apparatus to access the federated machine learning model and corresponding feature identifiers associated with different feature spaces, and wherein the apparatus further includes components for: selecting network functions based on the corresponding feature identifiers.

[0244] Clause 87. In the apparatus pursuant to Clause 86, each request sent to the corresponding network function in the network function further includes a characteristic identifier in the corresponding characteristic identifier.

[0245] Clause 88. An apparatus according to any one of Clauses 85 to 87, wherein enabling the apparatus to access a federated machine learning model includes enabling the apparatus to send a request to at least one of a vertical federated learning training server, a vertical federated learning training active participant, a model training logic function, or an analytics data repository function for retrieving the federated machine learning model.

[0246] Clause 89. An apparatus according to any one of Clauses 85 to 88, wherein at least one processing circuitry is configured to execute instructions to cause the apparatus to further at least: send a request to a network repository function requesting information relating to at least one network function, the at least one network function being used to provide information about a vertical federated learning session, the request including at least one of an analytics identifier or a vendor identifier of at least one network function; and receive information from the network repository function, wherein the network function for providing information about the vertical federated learning session is selected based on the information received from the network repository function.

[0247] Clause 90. An apparatus according to any one of Clauses 85 to 89, wherein the apparatus is implemented by network data analysis functions, analysis logic functions, or application functions of a telecommunications system.

[0248] Clause 91. An apparatus comprising: components for receiving a request for providing at least one of: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; components for selecting a network function to provide information about a vertical federated learning session associated with the analysis identifier; components for sending a request to the network function, the request including the analysis identifier, the request being sent to the network function for use with a federated machine learning model; components for receiving a response from the network function, the response including a vertical federated learning session identifier and information about the federated machine learning model; components for accessing a federated machine learning model associated with the vertical federated learning session identifier; components for sending a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a common sample space with different feature spaces, the request including the vertical federated learning session identifier and the analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; components for receiving intermediate output data from the network function based on inference; and components for performing inference using the intermediate output data and the federated machine learning model to compute predictions.

[0249] Clause 92. The apparatus according to Clause 91, wherein the component for accessing the federated machine learning model includes components for accessing the federated machine learning model and corresponding feature identifiers associated with different feature spaces, and wherein the apparatus further includes components for selecting network functions based on the corresponding feature identifiers.

[0250] Clause 93. In the apparatus of Clause 92, each request sent to the corresponding network function in the network function further includes a characteristic identifier in the corresponding characteristic identifier.

[0251] Clause 94. An apparatus according to any one of Clauses 91 to 93, wherein the component for accessing the federated machine learning model includes a component for: sending a request to at least one of a vertical federated learning training server, a vertical federated learning training active participant, a model training logic function, or an analytics data repository function for retrieving the federated machine learning model.

[0252] Clause 95. An apparatus according to any one of Clauses 91 to 94, wherein the apparatus further comprises: a component for sending a request to a network repository function, the request being for information relating to at least one network function, the at least one network function being used to provide information about a vertical federated learning session, the request including at least one of an analytics identifier or a vendor identifier of at least one network function; and a component for receiving information from the network repository function, wherein the network function for providing information about the vertical federated learning session is selected based on the information received from the network repository function.

[0253] Clause 96. An apparatus according to any one of Clauses 91 to 95, wherein the apparatus is implemented by network data analysis functions, analysis logic functions, or application functions of a telecommunications system.

[0254] Clause 97. A method comprising: receiving a request for providing at least one of: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; selecting a network function to provide information about a vertical federated learning session associated with the analysis identifier; sending a request to the network function, the request including the analysis identifier, the request being sent to the network function for use with a federated machine learning model; receiving a response from the network function including the vertical federated learning session identifier and information about the federated machine learning model; accessing a federated machine learning model associated with the vertical federated learning session identifier; sending a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a public sample space with different feature spaces, the request including the vertical federated learning session identifier and the analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; receiving intermediate output data from the network function based on the inference; and performing inference using the federated machine learning model based on the intermediate output data to compute predictions.

[0255] Clause 98. The method according to Clause 97, wherein accessing the federated machine learning model includes accessing the federated machine learning model and corresponding feature identifiers associated with different feature spaces, and wherein the apparatus further includes components for selecting network functions based on the corresponding feature identifiers.

[0256] Clause 99. According to the method of Clause 98, each request sent to the corresponding network function in the network function also includes a characteristic identifier in the corresponding characteristic identifier.

[0257] Clause 100. The method according to any one of Clauses 97 to 99, wherein accessing a federated machine learning model includes causing the device to send a request to at least one of a vertical federated learning training server, a vertical federated learning training active participant, a model training logic function, or an analytics data repository function for retrieving the federated machine learning model.

[0258] Clause 101. The method according to any one of Clauses 97 to 100, wherein the method further comprises: sending a request to a network repository function for information relating to at least one network function, the at least one network function being used to provide information about a vertical federated learning session, the request including at least one of an analytics identifier or a vendor identifier of at least one network function; and receiving information from the network repository function, wherein the network function for providing information about the vertical federated learning session is selected based on the information received from the network repository function.

[0259] Clause 102. The method according to any one of Clauses 97 to 101, wherein the method is implemented by network data analysis functions, analysis logic functions or application functions of the telecommunications system.

[0260] Clause 103. A computer-readable storage medium, which is non-transitory and has instructions stored therein, the instructions being responsive to execution by at least one processing circuitry, causing a device to at least: receive a request for providing at least one of: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; select a network function to provide information about a vertical federated learning session associated with the analysis identifier; send a request to the network function, the request including the analysis identifier, the request being sent to the network function for use with a federated machine learning model; receive a response from the network function, the response including a vertical federated learning session identifier and information about the federated machine learning model; access a federated machine learning model associated with the vertical federated learning session identifier; send a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a public sample space with different feature spaces, the request including the vertical federated learning session identifier and the analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; receive intermediate output data from the network function based on the inference; and perform inference using the federated machine learning model based on the intermediate output data to compute predictions.

[0261] Clause 104. The computer-readable storage medium pursuant to Clause 103, wherein enabling the device to access the federated machine learning model comprises: enabling the device to access the federated machine learning model and corresponding feature identifiers associated with different feature spaces, and wherein the device further comprises components for: selecting network functions based on the corresponding feature identifiers.

[0262] Clause 105. The computer-readable storage medium pursuant to Clause 104, wherein each request sent to the corresponding network function in the network function further includes a characteristic identifier in the corresponding characteristic identifier.

[0263] Clause 106. A computer-readable storage medium as described in any of Clauses 103 to 105, wherein enabling a device to access a federated machine learning model includes enabling the device to send a request to at least one of a vertical federated learning training server, a vertical federated learning training active participant, a model training logic function, or an analytics data repository function for retrieving the federated machine learning model.

[0264] Clause 107. A computer-readable storage medium according to any one of Clauses 103 to 106, wherein the computer-readable storage medium further has instructions stored therein, the instructions being responsive to execution by at least one processing circuitry, such that the apparatus further: sends a request to a network repository function requesting information relating to at least one network function, the at least one network function being used to provide information about a vertical federated learning session, the request including at least one of an analytics identifier or a vendor identifier of at least one network function; and receives information from the network repository function, wherein the network function for providing information about the vertical federated learning session is selected based on the information received from the network repository function.

[0265] Clause 108. A computer-readable storage medium pursuant to any of Clauses 103 to 107, wherein the apparatus is implemented by network data analysis functions, analysis logic functions, or application functions of a telecommunications system.

[0266] Clause 109. An apparatus comprising components for performing the method pursuant to any one of Clauses 97 to 102.

[0267] Clause 110. A computer-readable medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 97 to 102.

[0268] Clause 111. A computer-readable storage medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 97 to 102.

[0269] Clause 112. A computer program comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 97 to 102.

[0270] Clause 113. An apparatus comprising: at least one memory configured to store instructions; and at least one processing circuitry configured to access the at least one memory and execute instructions to cause the apparatus to at least: receive a request from a network function of a telecommunications system for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having different feature spaces, the request including a vertical federated learning session identifier; access the machine learning model associated with the vertical federated learning session identifier; perform inference using the machine learning model to compute intermediate output data; and send the intermediate data to the network function for inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0271] Clause 114. The apparatus according to Clause 113, wherein the request received from the network function includes a feature identifier, the feature identifier being associated with one of the different feature spaces in which it trains the machine learning model.

[0272] Clause 115. The apparatus pursuant to Clause 114, wherein the machine learning model is further accessed based on feature identifiers.

[0273] Clause 116. An apparatus as described in any of Clauses 113 to 115, wherein enabling the apparatus to access a machine learning model comprises enabling the apparatus to send a request to at least one of a vertical federated learning training client, a vertical federated learning training passive participant, a model training logic function, or an analytics data repository function, for storing the machine learning model.

[0274] Clause 117. An apparatus according to any one of Clauses 113 to 116, wherein a vertical federated learning session identifier is associated with an analysis identifier, and a request received from a network function also includes the analysis identifier.

[0275] Clause 118. The apparatus according to Clause 117, wherein the machine learning model is accessed from a vertical federated learning training client, and wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further: send a request to a network repository function requesting information relating to at least one network function supporting model training for feature identifiers and analysis identifiers; receive information from the network repository function; and select a vertical federated learning training client based on the information received from the network repository function.

[0276] Clause 119. An apparatus according to any one of Clauses 113 to 118, wherein the apparatus is implemented by a second network function, wherein at least one processing circuit is configured to execute instructions to cause the apparatus to further send a request to a network repository function for storing information related to the second network function, and wherein the request includes at least one of the following: instructions regarding the second network function being operable as a vertical federated learning inference client, an analysis identifier, or a feature identifier associated with a feature space in a different feature space using which a machine learning model is trained.

[0277] Clause 120. An apparatus according to any one of Clauses 113 to 119, wherein the apparatus is implemented by network data analysis functions, analysis logic functions or application functions of a telecommunications system.

[0278] Clause 121. An apparatus comprising: components for receiving a request from a network function of a telecommunications system, the request being for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having a different feature space, the request including a vertical federated learning session identifier; components for accessing the machine learning model associated with the vertical federated learning session identifier; components for performing inference using the machine learning model to compute intermediate output data; and components for sending the intermediate data to the network function for inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0279] Clause 122. The apparatus according to Clause 121, wherein a request received from a network function includes a feature identifier, the feature identifier being associated with one of the different feature spaces in which it trains a machine learning model.

[0280] Clause 123. The apparatus according to Clause 122, wherein the machine learning model is further accessed based on feature identifiers.

[0281] Clause 124. An apparatus according to any one of Clauses 121 to 123, wherein the component for accessing the machine learning model includes a component for: sending a request to at least one of a vertical federated learning training client, a vertical federated learning training passive participant, a model training logic function, or an analytics data repository function for storing the machine learning model.

[0282] Clause 125. An apparatus according to any one of Clauses 121 to 124, wherein a vertical federated learning session identifier is associated with an analysis identifier, and a request received from a network function also includes the analysis identifier.

[0283] Clause 126. The apparatus according to Clause 125, wherein the machine learning model is accessed from a vertical federated learning training client, and wherein the apparatus further comprises: a component for sending a request to a network repository function for information relating to at least one network function supporting model training for feature identifiers and analysis identifiers; a component for receiving information from the network repository function; and a component for selecting a vertical federated learning training client based on the information received from the network repository function.

[0284] Clause 127. An apparatus according to any one of Clauses 121 to 126, wherein the apparatus is implemented by a second network function, wherein the apparatus further includes components for: sending a request to a network repository function for storing information related to the second network function, and wherein the request includes at least one of: an instruction regarding the second network function being operable as a vertical federated learning inference client, an analysis identifier, or a feature identifier associated with a feature space in a different feature space using which a machine learning model is trained.

[0285] Clause 128. An apparatus according to any one of Clauses 121 to 127, wherein the apparatus is implemented by network data analysis functions, analysis logic functions or application functions of a telecommunications system.

[0286] Clause 129. A method comprising: receiving a request from a network function of a telecommunications system for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having a different feature space, the request including a vertical federated learning session identifier; accessing the machine learning model associated with the vertical federated learning session identifier; performing inference using the machine learning model to compute intermediate output data; and sending the intermediate data to the network function for inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0287] Clause 130. The method according to Clause 129, wherein the request received from the network function includes a feature identifier, which is associated with one of the different feature spaces in which it trains the machine learning model.

[0288] Clause 131. The method according to Clause 130, wherein the machine learning model is further accessed based on feature identifiers.

[0289] Clause 132. The method of any of Clauses 129 to 131, wherein accessing a machine learning model includes sending a request to at least one of a vertical federated learning training client, a vertical federated learning training passive participant, a model training logic function, or an analytics data repository function for storing the machine learning model.

[0290] Clause 133. The method according to any one of Clauses 129 to 132, wherein the vertical federated learning session identifier is associated with an analysis identifier, and the request received from the network function also includes the analysis identifier.

[0291] Clause 134. The method according to any one of Clause 133, wherein the machine learning model is accessed from a vertical federated learning training client, and wherein the method further comprises: sending a request to a network repository function for information relating to at least one network function supporting model training for feature identifiers and analytics identifiers; receiving information from the network repository function; and selecting a vertical federated learning training client based on the information received from the network repository function.

[0292] Clause 135. The method according to any one of Clauses 129 to 134, wherein the method is performed by a second network function, wherein the method further comprises: sending a request to a network repository function for storing information related to the second network function, and wherein the request includes at least one of the following: an instruction regarding the second network function's operability as a vertical federated learning inference client, an analysis identifier, or a feature identifier associated with a feature space in a different feature space using which a machine learning model is trained.

[0293] Clause 136. The method according to any one of Clauses 129 to 135, wherein the method is implemented by network data analysis functions, analysis logic functions or application functions of the telecommunications system.

[0294] Clause 137. A computer-readable storage medium, the computer-readable storage medium being non-transitory and having instructions stored therein, the instructions being responsive to execution by at least one processing circuitry to cause a means to at least: receive a request from a network function of a telecommunications system, the request being for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but having a different feature space, the request including a vertical federated learning session identifier; access the machine learning model associated with the vertical federated learning session identifier; perform inference using the machine learning model to compute intermediate output data; and send the intermediate data to the network function to perform inference using the federated machine learning model to compute a prediction based on the intermediate output data.

[0295] Clause 138. A computer-readable storage medium pursuant to Clause 137, wherein a request received from a network function includes a feature identifier, the feature identifier being associated with one of the different feature spaces in which it trains a machine learning model.

[0296] Clause 139. Computer-readable storage media pursuant to Clause 138, wherein the machine learning model is further accessed based on feature identifiers.

[0297] Clause 140. A computer-readable storage medium pursuant to any of Clauses 137 to 139, wherein enabling a device to access a machine learning model includes enabling the device to send a request to at least one of a vertical federated learning training client, a vertical federated learning training passive participant, a model training logic function, or an analytics data repository function, for storing a machine learning model. Clause 141. A computer-readable storage medium pursuant to any of Clauses 137 to 140, wherein a vertical federated learning session identifier is associated with an analysis identifier, and a request received from a network function also includes the analysis identifier.

[0298] Clause 142. The computer-readable storage medium according to Clause 141, wherein the machine learning model is accessed from a vertical federated learning training client, and wherein the computer-readable storage medium further has instructions stored therein, the instructions being responsive to execution by at least one processing circuitry, such that the apparatus further: sends a request to a network repository function for information relating to at least one network function supporting model training for feature identifiers and analysis identifiers; receives information from the network repository function; and selects a vertical federated learning training client based on the information received from the network repository function.

[0299] Clause 143. A computer-readable storage medium according to any one of Clauses 137 to 142, wherein the apparatus is implemented by a second network function, wherein the computer-readable storage medium has additional instructions stored therein, the additional instructions being responsive to execution by at least one processing circuitry to cause the apparatus to further send a request to a network repository function for storing information relating to the second network function, and wherein the request includes at least one of the following: instructions regarding the second network function being operable as a vertical federated learning inference client, an analysis identifier, or a feature identifier associated with a feature space in a different feature space using which a machine learning model is trained.

[0300] Clause 144. A computer-readable storage medium pursuant to any of Clauses 137 to 143, wherein the apparatus is implemented by network data analysis functions, analysis logic functions, or application functions of a telecommunications system.

[0301] Clause 145. An apparatus comprising components for performing the method pursuant to any one of Clauses 129 to 136.

[0302] Clause 146. A computer-readable medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 129 to 136.

[0303] Clause 147. A computer-readable storage medium comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 129 to 136.

[0304] Clause 148. A computer program comprising instructions that, in response to execution by at least one processing circuit, cause a device to perform a method according to any one of Clauses 129 to 136.

[0305] Benefiting from the teachings presented in the foregoing description and associated drawings, those skilled in the art to which this disclosure pertains will conceive of numerous modifications and other embodiments of the disclosure set forth herein. Therefore, it should be understood that this disclosure is not limited to the specific embodiments disclosed, and that modifications and other embodiments are intended to be included within the scope of the appended claims. Furthermore, although the foregoing description and associated drawings describe exemplary embodiments in the context of certain example combinations of elements and / or functions, it should be understood that different combinations of elements and / or functions may be provided by alternative embodiments without departing from the scope of the appended claims. In this regard, for example, combinations of elements and / or functions different from those explicitly described above are also contemplated as being set forth in some of the appended claims. Although specific terminology is used herein, it is used only in a general and descriptive sense and not for limiting purposes.

Claims

1. An apparatus comprising: The component used to assign a vertical federated learning session identifier to a vertical federated learning model training session; A component used to select network functions of a telecommunications system to perform model training on machine learning models with different feature spaces in a common sample space. as well as A component for sending a request to each of the network functions, the request being used to perform model training on a corresponding machine learning model in the common sample space using corresponding feature spaces in the different feature spaces, wherein the request includes the vertical federated learning session identifier so that the machine learning model is stored in association with the vertical federated learning session identifier.

2. The apparatus according to claim 1, wherein the apparatus further comprises: A component for receiving intermediate output data from the network function based on the model training; Components used to train a federated machine learning model to compute predictions based on the intermediate output data; as well as A component used to store the federated machine learning model in association with the vertical federated learning session identifier.

3. The apparatus of claim 2, wherein the different feature spaces are associated with corresponding feature identifiers, and wherein the federated machine learning model is also stored in association with the corresponding feature identifiers associated with the different feature spaces.

4. The apparatus of claim 2 or 3, wherein the component for storing the federated machine learning model includes a component for: sending a request to an analytics data repository function for storing the federated machine learning model at the analytics data repository function.

5. The apparatus according to any one of claims 2 to 4, wherein the apparatus further comprises: Components for receiving requests for the federated machine learning model; as well as Components for sending at least one of the following: the federated machine learning model, information for retrieving the federated machine learning model, the vertical federated learning session identifier, the analysis identifier, or a corresponding feature identifier associated with the different feature spaces.

6. The apparatus according to any one of claims 1 to 5, wherein the different feature spaces are associated with corresponding feature identifiers, and wherein the request sent to each of the network functions further includes a feature identifier from the corresponding feature identifiers.

7. The apparatus of any one of claims 1 to 6, wherein the federated learning session identifier is associated with an analysis identifier, and wherein the request sent to each of the network functions further includes the analysis identifier.

8. The apparatus according to any one of claims 1 to 7, wherein the apparatus further comprises: A component for sending requests to network repository functions, the requests being for information related to at least one network function that supports model training for feature identifiers and analytics identifiers. as well as A component for receiving the information from the network repository function, wherein the network function is selected based on the information received from the network repository function.

9. The apparatus according to any one of claims 1 to 8, wherein the apparatus is implemented by a network function, wherein the apparatus further comprises a component for operating to send a request to a network repository function to store information related to the network function, and wherein the request includes at least one of the following: an instruction that the network function is operable as a vertical federated learning training server, an instruction that the network function is operable as an active participant in vertical federated learning training, an analysis identifier, or an identifier of the vendor of the network function.

10. The apparatus according to any one of claims 1 to 9, wherein the apparatus further comprises: A component for selecting network functions to operate as active participants in the training of the machine learning model. as well as A component for sending a request to the network function to perform model training using the network function as an active participant, the request including the vertical federated learning session identifier and the analysis identifier.

11. The apparatus according to any one of claims 1 to 10, wherein the apparatus is implemented by the network data analysis function, model training logic function, or application function of the telecommunications system.

12. An apparatus comprising: A component for receiving a request from a network function of a telecommunications system, the request being for performing model training of a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but with different feature spaces, the request including a vertical federated learning session identifier. A component for training the machine learning model in the common sample space using corresponding feature spaces in the different feature spaces, wherein the machine learning model is trained to determine intermediate output data; A component for sending the intermediate data to the network function, wherein the intermediate data is sent to the network function to train a federated machine learning model based on the intermediate output data to compute predictions; as well as A component used to store the machine learning model in association with the vertical federated learning session identifier.

13. The apparatus of claim 12, wherein the request received from the network function includes a feature identifier associated with the corresponding feature space, the machine learning model being trained using the corresponding feature space.

14. The apparatus of claim 13, wherein the federated machine learning model is further stored in association with the feature identifier.

15. The apparatus of any one of claims 12 to 14, wherein the component for storing the machine learning model includes a component for: sending a request to an analytics data repository function to store the machine learning model at the analytics data repository function.

16. The apparatus of any one of claims 12 to 15, wherein the federated learning session identifier is associated with an analysis identifier, and wherein the request further includes the analysis identifier.

17. The apparatus according to any one of claims 12 to 16, wherein the apparatus further comprises: A component for receiving a request for the machine learning model, the request including the vertical federated learning session identifier and a feature identifier associated with the corresponding feature space; as well as The component for providing a response, wherein the response includes at least one of the machine learning model or information for retrieving the machine learning model.

18. The apparatus of any one of claims 12 to 17, wherein the apparatus is implemented by a second network function, wherein the apparatus further comprises a component for operating to: send a request to a network repository function to store information related to the second network function, and wherein the request includes at least one of the following: an instruction regarding the second network function being operable as a vertical federated learning training client, an instruction regarding the second network function being operable as a vertical federated learning training passive participant, an analysis identifier, or a feature identifier associated with the corresponding feature space, wherein the machine learning model is trained using the corresponding feature space.

19. The apparatus according to any one of claims 12 to 18, wherein the apparatus is implemented by network data analysis function, model training logic function, or application function of the telecommunications system.

20. An apparatus comprising: A component for receiving requests, the requests being for providing at least one of the following: analysis of a data analysis process associated with an analysis identifier, or one or more predictions; A component for selecting network functions to provide information about a vertical federated learning session, which is associated with the analysis identifier; A component for sending a request to the network function, the request being sent to the network function including the analytics identifier, the request being sent to the network function for use in a federated machine learning model; A component for receiving a response from the network function, the response including a vertical federated learning session identifier and information about the federated machine learning model; Components for accessing the federated machine learning model, which is associated with the vertical federated learning session identifier; A component for sending a request to a network function of a telecommunications system for performing inference using the machine learning model, the machine learning model being trained in a common sample space but having a different feature space, the request sent to the network function of the telecommunications system including the vertical federated learning session identifier and the analysis identifier for retrieving the machine learning model based on the vertical federated learning session identifier and the analysis identifier; A component for receiving intermediate output data from the network function based on the inference; as well as The component used to perform inference to compute predictions based on the intermediate output data and using the federated machine learning model.

21. The apparatus of claim 20, wherein the component for accessing the federated machine learning model includes components for: accessing the federated machine learning model and corresponding feature identifiers associated with the different feature spaces, and The device further includes components for selecting the network function based on the corresponding feature identifier.

22. The apparatus of claim 21, wherein each of the requests sent to the corresponding network function in the network function further includes a feature identifier in the corresponding feature identifier.

23. The apparatus of any one of claims 20 to 22, wherein the component for accessing the federated machine learning model includes a component for: sending a request to at least one of a vertical federated learning training server, a vertical federated learning training active participant, a model training logic function, or an analytics data repository function for retrieving the federated machine learning model.

24. The apparatus according to any one of claims 20 to 23, wherein the apparatus further comprises: A component for sending a request to a network repository function, the request being for information related to at least one network function, the at least one network function being used to provide information about the vertical federated learning session, the request including at least one of the analytics identifier or the vendor identifier of the at least one network function; as well as The component for receiving the information from the network repository function, wherein the network function for providing the information about the vertical federated learning session is selected based on the information received from the network repository function.

25. The apparatus according to any one of claims 20 to 24, wherein the apparatus is implemented by network data analysis functions, analysis logic functions, or application functions of the telecommunications system.

26. An apparatus comprising: A component for receiving a request from a network function of a telecommunications system, the request being for performing inference using a machine learning model, the machine learning model being one of machine learning models trained in a common sample space but with different feature spaces, the request including a vertical federated learning session identifier. Components for accessing the machine learning model associated with the vertical federated learning session identifier; Components used to perform inference using the machine learning model to compute intermediate output data; as well as Components for sending the intermediate data to the network function to perform inference using a federated machine learning model to compute predictions based on the intermediate output data.

27. The apparatus of claim 26, wherein the request received from the network function includes a feature identifier associated with one of the different feature spaces in which the machine learning model was trained.

28. The apparatus of claim 27, wherein the machine learning model is further accessed based on the feature identifier.

29. The apparatus of any one of claims 26 to 28, wherein the component for accessing the machine learning model includes a component for: sending a request to at least one of a vertical federated learning training client, a vertical federated learning training passive participant, a model training logic function, or an analytics data repository function, the request being for storing the machine learning model.

30. The apparatus of any one of claims 26 to 29, wherein the vertical federated learning session identifier is associated with an analysis identifier, and the request received from the network function further includes the analysis identifier.

31. The apparatus of claim 30, wherein the machine learning model is accessed from a vertical federated learning training client, and wherein the apparatus further comprises: A component for sending a request to a network repository function, the request being for information related to at least one network function that supports model training for feature identifiers and analysis identifiers; Components for receiving the information from the network repository function; as well as A component for selecting the vertical federated learning training client based on the information received from the network repository function.

32. The apparatus of any one of claims 26 to 31, wherein the apparatus is implemented by a second network function, wherein the apparatus further comprises a component for operating to send a request to a network repository function for storing information related to the second network function, and wherein the request includes at least one of the following: an indication that the second network function is operable as a vertical federated learning inference client, the analysis identifier, or a feature identifier associated with one of the feature spaces in the different feature spaces used to train the machine learning model thereon.

33. The apparatus according to any one of claims 26 to 32, wherein the apparatus is implemented by network data analysis functions, analysis logic functions or application functions of the telecommunications system.

Images