Apparatus and communication method for ai / ml operation

Abstract

A communication method for artificial intelligence (AI) / machine learning (ML) operations includes a network entity discovering trustworthiness capabilities for analytics and / or models, where a network data analytics function (NWDAF) is enabled to support a model training logic function (MTLF) with trustworthiness capabilities for ML models, the NWDAF is enabled to support an analytics logic function (AnLF) with trustworthiness capabilities for analytics, or a network repository function (NRF) includes trustworthiness capability provisioning per ML model. The network entity can be the NWDAF or a NWDAF consumer.

Description

Technical Field

[0001] This disclosure relates to the field of communication systems, and more specifically, to apparatus and communication methods for AI / ML operations such as artificial intelligence (AI) / machine learning (ML) credibility services for registration and discovery in network repository functions (NRFs). Background Technology

[0002] Currently, the 3rd Generation Partnership Project (3GPP) is conducting standardization activities on artificial intelligence (AI) / machine learning (ML) capabilities. However, in current technology, the aspects of AI / ML trusted services are not yet clearly defined in the 5G core network (5GC).

[0003] Therefore, there is a need for devices and communication methods for AI / ML operations such as AI / ML trust services that can solve these and other problems. Summary of the Invention

[0004] The purpose of this disclosure is to provide an apparatus and communication method for artificial intelligence (AI) / machine learning (ML) operational services for registration and discovery in a network repository function (NRF), which can solve these and other problems in the prior art.

[0005] In a first aspect of this disclosure, a communication method for AI / ML operations is provided, wherein the method is implemented in a network entity and includes: discovering credibility capabilities for analytics and / or models, wherein a network data analytics function (NWDAF) is enabled to support a model training logical function (MTLF) having credibility capabilities for ML models, the NWDAF being enabled to support an analytics logical function (AnLF) having credibility capabilities for analytics, or an NRF including a credibility capability supply for each ML model.

[0006] In a second aspect of this disclosure, a communication device includes a determiner for discovering credibility capabilities for analysis and / or models, wherein an NWDAF is enabled to support an MTLF with credibility capabilities for ML models, an AnLF with credibility capabilities for analysis, or an NRF includes a credibility capability supply for each ML model.

[0007] In a third aspect of this disclosure, a network device includes a memory, a transceiver, and a processor coupled to the memory and the transceiver. The network device is used to perform the methods described above.

[0008] In a fourth aspect of this disclosure, a non-transitory machine-readable storage medium is provided having instructions stored thereon that, when executed by a computer, cause the computer to perform the methods described above.

[0009] In a fifth aspect of this disclosure, a chip includes a processor for calling and running a computer program stored in a memory to cause a device on which the chip is mounted to perform the methods described above.

[0010] In a sixth aspect of this disclosure, a computer-readable storage medium stores a computer program that causes a computer to perform the above-described method.

[0011] In a seventh aspect of this disclosure, a computer program product includes a computer program that causes a computer to perform the methods described above.

[0012] In the eighth aspect of this disclosure, a computer program causes a computer to perform the above-described method. Attached Figure Description

[0013] To more clearly illustrate the embodiments or related technologies of this disclosure, the following drawings, which will be described in the embodiments, will be briefly introduced. Obviously, the drawings are only some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without giving any premise.

[0014] Figure 1 This is a block diagram of a non-roaming 5G system architecture for implementing some of the embodiments presented herein.

[0015] Figure 2 This is a block diagram of a data collection architecture from any 5GC network function (NF) for implementing some of the embodiments presented herein.

[0016] Figure 3 This is a block diagram of a data collection architecture using data collection coordination for implementing some of the embodiments presented herein.

[0017] Figure 4This is a block diagram of an open architecture for network data analytics used to implement some of the embodiments presented herein.

[0018] Figure 5 This is a block diagram of an open architecture for network data analytics that uses data collection coordination to implement some of the embodiments presented herein.

[0019] Figure 6 This is a block diagram of an ML model provisioning architecture for implementing some of the embodiments presented herein.

[0020] Figure 7 This is a block diagram of a network device according to an embodiment of the present disclosure.

[0021] Figure 8 This is a flowchart of a communication method for artificial intelligence (AI) / machine learning (ML) operations according to embodiments of this disclosure.

[0022] Figure 9 This is a block diagram of a communication device according to an embodiment of the present disclosure.

[0023] Figure 10 This is a block diagram of an example computing device according to an embodiment of the present disclosure.

[0024] Figure 11 This is a block diagram of a communication system according to an embodiment of the present disclosure. Detailed Implementation

[0025] The technical problems, structural features, objectives, and effects of the embodiments of this disclosure will now be described in detail with reference to the accompanying drawings. Specifically, the terminology used in the embodiments of this disclosure is only for describing a particular embodiment and is not intended to limit the disclosure.

[0026] Figure 1 A non-roaming 5G system architecture for implementing some of the embodiments presented herein is illustrated. In this 5G non-roaming system architecture, network functions communicate with each other through service-based interfaces in the core network (CN). User equipment (UE) can communicate with the core network to establish control signaling and enable the UE to use services from the CN. Examples of control signaling functions include registration, connection and mobility management, authentication and authorization, session management, etc. After control signaling has been established, the UE can then use user plane functions to send data to and receive data from a data network (DN), such as the Internet.

[0027] The 5G system architecture includes the following network functions (NFs): Authentication Server Function (AUSF), Access and Mobility Management Function (AMF), Data Networking (DN) such as operator services, internet access, or third-party services, Unstructured Data Storage Function (UDSF), Network Exposure Function (NEF), Network Repository Function (NRF), Network Slice Admission Control Function (NSACF), Network Slice-Specific and SNPN Authentication and Authorization Function (NSSAAF), Network Slice Selection Function (NSSF), Policy Control Function (PCF), Session Management Function (SMF), Unified Data Management (UDM), Unified Data Repository (UDR), User Plane Function (UPF), UE Radio Capability Management Function (UCMF), and Application Function (NFs). Functions, AF, UE, (radio)access network ((R)AN), 5G-equipment identity register (5G-EIR), network data analytics function (NWDAF), charging function (CHarging function, CHF), time-sensitive networking AF,The system includes TSN AF, Time-Sensitive Communication and Time Synchronization Function (TSCTSF), Data Collection Coordination Function (DCCF), Analytics Data Repository Function (ADRF), Messaging Framework Adaptor Function (MFAF), and Non-Seamless WLAN Offload Function (NSWOF).

[0028] The following description emphasizes Figure 1 Some capabilities of network functions (NFs) related to control signaling.

[0029] Access and Mobility Functions (AMF): The UE sends N1 messages to the AMF through the RAN node to perform control plane signaling, such as registration, connection management, mobility management, access authentication and authorization.

[0030] Session Management Function (SMF): The SMF is responsible for session management related to establishing PDU sessions, enabling the UE to send data to a data network (DN) such as the Internet or to an application server and other session management-related functions.

[0031] Policy Control Function (PCF): PCF provides a policy framework for managing network behavior and accessing subscription information, in order to make policy decisions, etc.

[0032] Authentication Server Function (AUSF): AUSF supports UE authentication for both 3GPP access and untrusted non-3GPP access.

[0033] Unified Data Management / Repository (UDM / UDR): UDM / UDR supports 3GPP AKA authentication credential generation, user identification processing, subscription management, and storage. Network Slice Selection Function (NSSF): NSSF relates to various aspects of network slice management, such as selecting network slice instances for UEs and managing NSSAI.

[0034] Network Repository Function (NRF): NRF supports service discovery in 5G networks.

[0035] Network Open Function (NEF): NEF supports exposing core network capabilities and events to third parties, application functions (AF), edge computing, etc.

[0036] To enable communication between the control plane and the user plane, the RAN node provides communication access from the UE to the core network. The UE establishes a PDU session with the CN to transmit data services on the user plane through the (R)AN and UPF nodes of the 5G system (5GS). Uplink services are transmitted by the UE, while downlink services are received by the UE using the established PDU session. Data services flow between the UE and the DN through intermediate nodes (R)AN and UPF.

[0037] In some embodiments, the storage device for trusted services may be such as Figure 1 As defined in the non-roaming 5G system architecture shown, if in such Figure 1 The non-roaming 5G system architecture shown has a trust requirement. For appropriate network functions (NFs) that support trust, the method of accessing the storage device can be selected for this operation.

[0038] With the increasing use of artificial intelligence (AI) / machine learning (ML) in mobile networks, the need to ensure the trustworthiness of AIML-supported services is also growing. The following terms may be included in proposals for AI-related rules: Trustworthy machine learning: This involves a set of seven key requirements that a machine learning system must meet to be considered trustworthy. Details of each requirement are as follows: Human agency and oversight: AI systems can empower humans to make informed decisions and promote their fundamental rights. Simultaneously, it is necessary to ensure the establishment of appropriate oversight mechanisms, which can be achieved through human-in-the-loop, human-on-the-loop, and human-in-command approaches.

[0039] Technical robustness and security: AI systems need to be resilient and secure. They must be secure, ensuring fallback mechanisms in case of failure, and possess accuracy, reliability, and reproducibility. This may be the only way to minimize and prevent unintended damage.

[0040] Privacy and data management: In addition to ensuring full respect for privacy and data protection, it is also necessary to ensure appropriate data management mechanisms that take into account the quality and integrity of the data and ensure lawful access to the data.

[0041] Transparency: Data, systems, and AI business models are transparent. Traceability mechanisms help achieve this. Furthermore, AI systems and their decisions can be explained in a way that is accessible to relevant stakeholders. Humans need to be aware that they are interacting with AI systems and can understand the system's capabilities and limitations.

[0042] Diversity, non-discrimination, and fairness: Avoiding unfair biases can have a range of negative impacts, from marginalizing vulnerable groups to exacerbating prejudice and discrimination. To promote diversity, AI systems should be accessible to everyone, regardless of their flaws, and involve relevant stakeholders throughout the system's lifecycle.

[0043] Accountability: Mechanisms can be established to ensure accountability and responsibility for AI systems and their outcomes. Auditability plays a crucial role in enabling the evaluation of algorithms, data, and design processes, especially in critical applications. Furthermore, it ensures adequate and readily available remedial measures.

[0044] Social and environmental well-being: AI systems can benefit all of humanity, including future generations. Therefore, it is possible to ensure that AI systems are sustainable and environmentally friendly. Furthermore, AI systems can take into account the environment, including other living beings, and their social and societal impacts can be carefully assessed.

[0045] As 3GPP is standardizing the definition of AI / ML functions in mobile networks, it is necessary to extend this standardization and define trusted aspects related to AI / ML functions in mobile networks, including corresponding parameters, parameter storage, network functions and services, and parameter processing. Some embodiments of this disclosure relate to defining storage devices for trusted services in a network. If a trust requirement exists in the network, the appropriate network function supporting the trust can select the method of accessing the storage device for that operation. Some embodiments of this disclosure analyze 5G architecture and requirements and propose how to embed trust aspects into 5G architecture and requirements.

[0046] Figure 2 A data collection architecture from any 5GC network function (NF) is illustrated for implementing some embodiments proposed herein. For example... Figure 2As shown, the 5G system architecture allows the NWDAF to collect data from any 5GC NF. The NWDAF belongs to the same public land mobile network (PLMN) as the 5GC NF providing the data. The NNF interface is defined for the NWDAF and is used to request subscriptions to data deliveries for a specific context, unsubscribe from data deliveries, and request specific reports for data in a specific context. The 5G system architecture allows the NWDAF to retrieve management data from the OAM by invoking the network management function (OAM) service. The 5G system architecture allows the NWDAF to collect data from any 5GC NF or OAM using the Data Collection Coordination Function (DCCF) through the associated Ndccf service. The 5G system architecture allows the NWDAF and DCCF to collect data from the NWDAF through the associated Nnwdaf_DataManagement service. The 5G system architecture allows the MFAF to extract data from the NWDAF through the associated Nnwdaf_DataManagement service.

[0047] Figure 3 A data collection architecture using data collection coordination is illustrated for implementing some embodiments proposed herein. For example... Figure 3 As shown, the Ndccf interface is defined for NWDAF and is used to support subscription requests, unsubscriptions to data deliveries, and specific reports requesting data from DCCF. If data has not yet been collected, DCCF uses the NNF service to request data from the data source. DCCF can collect data and pass it to NWDAF, or DCCF can rely on a messaging framework to collect data from NF and pass it to NWDAF.

[0048] Figure 4 An open architecture for network data analytics used to implement some of the embodiments proposed herein is illustrated. For example... Figure 4As shown, the 5G system architecture allows any 5GC NF to request network analytics information from an NWDAF that supports the AnLF (Anti-Analysis Logic Function). An AnLF-supporting NWDAF can refer to an NWDAF that includes the AnLF. The NWDAF belongs to the same PLMN as the 5GC NF consuming analytics information. The Nnwdaf interface is defined for 5GC NFs and is used to request subscriptions to, unsubscribe from, and request specific reports for network analytics in a specific context. The 5G system architecture also allows other consumers, such as OAM and charging enablement functions (CEF), to request network analytics information from the NWDAF. The 5G system architecture allows any NF to obtain analytics from the NWDAF using the DCCF (Data Management Function) through the associated Ndccf service. The 5G system architecture allows both the NWDAF and DCCF to request historical analytics from the NWDAF through the associated Nnwdaf_DataManagement service. The 5G system architecture allows the MFAF to extract historical analytics from the NWDAF through the associated Nnwdaf_DataManagement service.

[0049] Figure 5 An open architecture for network data analytics using data collection coordination is illustrated for implementing some embodiments proposed herein. For example... Figure 5 As shown, the Ndccf interface is defined for any NF and is used to support subscribing to, unsubscribing from, and requesting specific reports from network analytics. If analytics have not yet been collected, DCCF requests analytics from NWDAF using the Nnwdaf service. DCCF can collect analytics and pass them to the NF, or DCCF can rely on a messaging framework to collect analytics and pass them to the NF.

[0050] Figure 6 An ML model provisioning architecture for implementing some of the embodiments presented herein is shown. Figure 6 As shown, the 5G system architecture allows an NWDAF supporting an Analysis Logic Function (AnLF) to use a trained ML model provisioning service from another NWDAF supporting a Model Training Logic Function (MTLF). The NWDAF interface is used by the AnLF-supporting NWDAF to request and subscribe to the trained ML model provisioning service. The AnLF-supporting NWDAF can be the sole consumer of the trained ML model provisioning service. An AnLF-supporting NWDAF can refer to an NWDAF that contains an AnLF. An MTLF-supporting NWDAF can refer to an NWDAF that contains an MTLF.

[0051] NWDAF can include the following logical functions: Analysis Logic Function (AnLF): A logic function in NWDAF that performs inference, derives analytical information (i.e., derives statistical data and / or forecasts based on analytical consumer requests), and exposes analytical services, namely Nnwdaf_AnalyticsSubscription or Nnwdaf_Analyticslnfo.

[0052] Model Training Logic Function (MTLF): A logic function in NWDAF that trains machine learning (ML) models and opens up new training services (e.g., providing ML models for training).

[0053] NWDAF can include MTLF or AnLF, or both. The analysis information is either statistical information about past events or predictive information. Different NWDAF instances can exist within 5GC, and these instances can be specialized according to the type of analysis. The capabilities of an NWDAF instance are described in the NWDAF configuration file stored in NRF. To ensure the accuracy of the analysis output for the analysis identifier (ID), NWDAF detects and can delete input data from anomalous UEs based on analysis of anomalous UE lists and observation time windows from itself or other NWDAFs. Then, if there is no input data related to the anomalous UE list within the observation time window, a new ML model and / or analysis output can be generated for that analysis ID. The ML model information and / or analysis output are then sent / updated to the subscribed NWDAF service consumers.

[0054] To support NF discovery and selection of NWDAF instances that support MTLF, AnLF, or both and can provide the required services (e.g., analytics open or ML model provisioning) for the desired type of analytics, each NWDAF instance should, in addition to other NRF registration elements in the NF configuration file, provide a list of supported analytics IDs when registering with the NRF (potentially based on the supported services). NFs that need to discover NWDAF instances that support certain services for specific types of analytics can query the NRF for NWDAFs that support the required services and analytics IDs. Consumers, i.e., 5GC NFs and OAMs, decide how to use the data analytics provided by the NWDAFs. Interaction between 5GC NFs and NWDAFs occurs within the PLMN. NWDAFs are unaware of NF application logic. NWDAFs can use subscribed data, but only for statistical purposes. The NWDAF architecture allows multiple NWDAF instances to be arranged in a hierarchy / tree with a flexible number of layers / branches. The number and organization of the hierarchical layers, as well as the capabilities of each NWDAF instance, remain deployment options.

[0055] In a tiered deployment, when DCCF and MFAF are absent in the network, NWDAF can provide open data collection capabilities for generating analyses based on data collected by other NWDAFs. To make NWDAF discoverable in some network deployments, NWDAF can be configured (e.g., for UE mobility analysis) to register in the UDM (Nudm_UECM_Registration service operation) for the UEs it serves and the associated analysis ID. Registration in the UDM can occur when the NWDAF starts serving a UE or collecting data for a UE. When the NWDAF deletes the analysis context of a UE with the associated analysis ID, it is deregistered in the UDM.

[0056] Some solutions disclosed herein propose including trusted services in a network. In some embodiments, a communication method for AI / ML operations includes: discovering trusted capabilities for analytics and / or models, wherein an NWDAF is enabled to support an MTLF with trusted capabilities for ML models, the NWDAF is enabled to support an AnLF with trusted capabilities for analytics, or an NRF includes a trusted capability provision for each ML model. Specifically, in some examples, the NWDAF supports AnLF capabilities, and / or the NWDAF supports MTLF capabilities. Some embodiments of the invention define how to discover trusted capabilities for corresponding analytics and corresponding models in the following solutions: 1. Supports AnLF-based NWDAF, whose analysis parameters include confidence level.

[0057] 2. Supports NWDAF for MTLF, with analysis parameters corresponding to the trained ML model. Optionally, existing model provisioning services are reused and extended to include confidence parameters. Optionally, a new service for confidence is introduced.

[0058] 3. In addition, for cases where NWDAF is not deployed in the network, or there is no NWDAF that supports MTLF, or there is an NWDAF that supports MTLF but does not support the relevant model ID, and the model can be directly used by different network functions, the model can be directly supplied to NRF and discovered by the corresponding network function.

[0059] NWDAF discovery and selection NWDAF service consumers use NWDAF discovery principles to select NWDAFs that support the requested analytics information, required analytics capabilities, and / or requested ML model information. In some embodiments, NWDAF may be enabled to support MTLF with trustworthiness capabilities for ML models. In some embodiments, NWDAF may be enabled to support AnLF with trustworthiness capabilities for analytics. Different deployments may require different discovery and selection parameters. The different ways in which discovery and selection mechanisms are performed depend on the different types of analytics / data (NF-related analytics / data and UE-related analytics / data). NF-related refers to analytics / data that does not require either SUPI or SUPI groups (e.g., NF load analysis). UE-related refers to analytics / data that requires either SUPI or SUPI groups (e.g., UE mobility analysis).

[0060] To use NRF to discover NWDAFs that support AnLF: If the analysis is related to NF and the NWDAF service consumer (other than the NWDAF itself) cannot provide a region of interest for the requested data analysis, the NWDAF service consumer can select an NWDAF with a large service area from the candidate NWDAFs from the discovery response. Alternatively, if the consumer receives an NWDAF with aggregation capabilities, the consumer preferably selects an NWDAF with a large service area and aggregation capabilities.

[0061] If, for example, the selected NWDAF cannot provide the requested data analysis because the NF to be contacted is outside the service area of ​​the NWDAF, the selected NWDAF may reject the analysis request / subscription, or the selected NWDAF may query the NRF to determine another target NWDAF by the service area of ​​the NF to be contacted. If the analysis is UE-related and the NWDAF service consumer (other than the NWDAF) cannot provide the region of interest for the requested data analysis, the NWDAF service consumer may select an NWDAF with a large service area from the candidate NWDAFs from the discovery response. Alternatively, if the consumer receives an NWDAF with aggregation capabilities, the consumer preferably selects an NWDAF with a large service area and aggregation capabilities.

[0062] If the selected NWDAF cannot provide analytics for the requested UE (e.g., the NWDAF serves a different service area), the selected NWDAF may reject the analytics request / subscription, or the selected NWDAF may determine the AMF serving the UE, request the UE location information from the AMF, and query the NRF through the tracking area where the UE is located to discover another target NWDAF serving the area where the UE is located. If the analytics is UE-related, and if the NWDAF instance indicates the weight of the TAI in its configuration file, the NWDAF service consumer can use the weight of the TAI to decide which NWDAF to select. If the NWDAF service consumer needs to discover an NWDAF that supports AnLF with accuracy checking capabilities, the consumer can query the NRF and provide the accuracy checking capabilities in the discovery request. If the NWDAF service consumer needs to discover an NWDAF that can collect data from a specific data source identified by an NF set ID or NF type of a specific data source, the consumer can query the NRF and provide the NF set ID or NF type in the discovery request.

[0063] To discover NWDAFs that have already been registered in the UDM for a given UE: NWDAF service consumers, or other NWDAFs interested in UE-related data or analysis, can query the UDM, if supported, to discover NWDAF instances already serving a given UE. If an NWDAF service consumer needs to discover NWDAFs with open data collection capabilities, it can discover the NWDAF providing the Nnwdaf_DataManagement service and its associated data source NF type or its associated data source NF set ID via NRF.

[0064] To discover NWDAFs that support MTLF via NRF: When a trained ML model is available for one or more analytics IDs, an MTLF-enabled NWDAF should include ML model provisioning services (i.e., Nnwdaf_MLModelProvision, Nnwdaf_MLModelInfo) as one of the services supported during registration with the NRF. An MTLF-enabled NWDAF can provide the NRF with a list of analytics IDs corresponding to the trained ML model, and, if possible, ML model filter information for the trained ML model for each analytics ID. In this version of the specification, during MTLF-enabled NWDAF registration, only the S-NSSAI and region of interest from the ML model filter information for the trained ML model for each analytics ID can be registered with the NRF. For each analytics ID, if the MTLF-enabled NWDAF supports ML model interoperability, it may also include an ML model interoperability indicator when registering with the NRF.

[0065] The ML model interoperability indicator includes a list of NWDAF providers (vendors) that are permitted to retrieve ML models from the NWDAF that supports MTLF. The ML model interoperability indicator also indicates the interoperable ML models requested by the NWDAFs of vendors in the MTLF-supporting NWDAF support list.

[0066] The S-NSSAI and region of interest from the ML model filter information are located within the S-NSSAI and NWDAF service area information indicated in the NF configuration file of the NWDAF that supports MTLF.

[0067] In the process of discovering NWDAFs that support MTLF, the consumer (i.e., the NWDAF that supports AnLF) can include the target NF type (i.e., NWDAF), analysis ID, S-NSSAI, region of interest of the desired ML model to be trained, ML model interoperability indicator, and NF consumer information in the request. The NRF returns one or more candidate instances of NWDAFs that support MTLF to the NF consumer. Each candidate instance of an MTLF-supporting NWDAF includes the analysis ID and, if possible, may include ML model filter information for an available ML model to be trained.

[0068] If an NWDAF service consumer needs to discover NWDAFs that support MTLF with accuracy checking capabilities, the consumer can query the NRF and provide the accuracy checking capability in the discovery request.

[0069] To discover NWDAFs that support MTLF with federated learning (FL) capabilities via NRF: During the registration period in the NRF, an NWDAF supporting MTLF as a server should also include the FL capability type (i.e., FL server) and the time interval for supporting FL as FL capability information. Similarly, an NWDAF supporting MTLF as a client should also include the FL capability type (i.e., FL client) and the time interval for supporting FL as FL capability information during the NRF registration period. This NWDAF may also include an NF type, where data can be collected as input for local model training. An NWDAF supporting MTLF can indicate simultaneous support for both FL servers and FL clients within the FL capabilities of a specific analytics ID.

[0070] In the process of discovering NWDAFs that support MTLF as FL servers, the consumer (e.g., the NWDAF supporting MTLF) includes in its request the FL capability type as an FL server, the time period of interest, and, if possible, ML model filter information for the ML model trained for each analysis ID. The NRF returns one or more candidate instances of NWDAFs that support MTLF as FL servers to the consumer. In the process of discovering NWDAFs that support MTLF as FL clients, the consumer (e.g., the FL server) includes in its request the FL capability type as an FL client, the time period of interest, ML model filter information for the ML model trained for each analysis ID, and a list of NF types. The NRF returns one or more candidate instances of NWDAFs that support MTLF as FL clients to the consumer. The service consumer for discovering NWDAFs that support MTLF with FL capabilities is limited to NWDAFs that support MTLF.

[0071] The PCF can learn through signaling which NWDAFs are being used by the AMF, SMF, and UPF for a specific UE. This allows the PCF to select the same NWDAF instance already used for a specific UE. In the roaming architecture, NWDAFs with roaming switching capabilities to request analytics or input data (RE-NWDAFs) are discovered via NRF. Consumers in the same PLMN as the RE-NWDAF discover the RE-NWDAF by querying the NWDAF indicated in its NRF profile by the roaming switching capability. Consumers in the peer PLMN (i.e., the RE-NWDAF) discover the RE-NWDAF by querying the NWDAF in the target PLMN that supports the specific services defined for roaming. RE-NWDAFs (if not using delegated discovery) are discovered in different PLMNs (i.e., HPLMN or VPLMN) using a procedure where detailed parameters are determined based on analytics requests or subscriptions from the consumer's 5GC NF, operator policies, user consent, and / or local configuration.

[0072] Solution 1: NWDAF can be enabled to support MTLF with credibility capabilities for ML models.

[0073] NWDAFs supporting MTLF with credibility capability can be locally configured with a set of IDs of NWDAFs supporting MTLF with credibility capability and the analysis ID supported by each NWDAF supporting MTLF with credibility capability to retrieve trained ML models and corresponding credibility parameters, or NWDAFs supporting MTLF with credibility capability can be discovered using a further specified NWDAF discovery procedure.

[0074] To discover NWDAFs that support MTLFs with trustworthiness capabilities via NRF: NWDAFs that support MTLFs with trustworthiness capabilities may include at least one of the following.

[0075] The ML model provisioning service (i.e., Nnwdaf_MLModelProvision, Nnwdaf_MLModelInfo) is one of the supported services: when a trained ML model is available for one or more analytics IDs, the NWDAF with MTLF support for credibility can provide this ML model provisioning service during registration in the NRF.

[0076] In some examples, one or more ML model provisioning services and / or one or more ML trust services include: a priority of trusted AI solutions, untrusted AI solutions, and fallback mechanisms between non-AI solutions to ensure security; and / or a list of one or more analytics IDs corresponding to the trained ML models and ML model filter information for the ML models trained for each analytics ID.

[0077] In some examples, the ML model filter information includes at least one of the following confidence-related parameters: data type; version; sampling frequency; sampling weight; model weights in the case of linking / merging with another model; labeled / unlabeled data; interpretability level; risk level (e.g., unacceptable, high, limited); fairness; robustness; privacy; security; safety; reliability; traceability; ML decision confidence score (a numerical value representing the reliability / quality of a given decision generated by the AI / ML inference function); and / or data value quality score, which is a numerical value representing the reliability / quality of a given type of observation and measurement. These parameters can be added to the ML model filter information on top of existing parameters.

[0078] The ML Trustworthiness Service is one of the supported services: During registration in the NRF, an NWDAF supporting MTLF with trustworthiness capabilities can provide this service when the trained ML model is available for one or more analytics IDs. The ML Trustworthiness Service may contain information on various trustworthiness parameters generated by the corresponding model training of the MTLF, such as: the priority of trusted AI solutions, untrusted AI solutions, and fallback mechanisms between non-AI solutions used to ensure security; and / or a list of analytics IDs corresponding to the trained ML model and ML model filter information for the ML model trained for each analytics ID.

[0079] In some examples, the ML model filter information includes at least one of the following credibility-related parameters: data type; version; sampling frequency; sampling weight; model weight in the case of linking / merging with another model; labeled / unlabeled data; interpretability level; risk level (e.g., unacceptable, high, limited); fairness; robustness; privacy; security; safety; reliability; traceability; ML decision confidence score (a numerical value representing the credibility / quality of a given decision generated by the AI / ML inference function); and / or data value quality score, which is a numerical value representing the credibility / quality of a given type of observation and measurement. These parameters can be added to the ML model filter information on top of existing parameters. In some examples, a list of analysis IDs is associated between the ML model provisioning service and the ML credibility service.

[0080] Solution 2: NWDAF can be enabled to support AnLF with credibility capabilities for analytics.

[0081] NWDAFs supporting Trustworthy AnLF can be configured locally with a set of IDs supporting Trustworthy AnLF and an analytics ID supported by each NWDAF supporting Trustworthy AnLF, or they can be discovered using a further specified NWDAF discovery procedure. To discover NWDAFs supporting Trustworthy AnLF via NRF: based on each analytics ID, the corresponding trustworthiness-related parameters for each analytics filter information can include the following information: risk level (e.g., unacceptable, high, limited); fairness; robustness; privacy; security; safety; reliability; and / or traceability.

[0082] Solution 3: Alternative solutions (for situations where NWDAF is not deployed in the network, or where there is no NWDAF supporting MTLF, or where there is an NWDAF supporting MTLF but not the relevant model ID) and (where the model can be directly used by different network functions): NRF can include the credibility capability provision of each ML model, so the ML model filter information for each model can include the following credibility-related parameters: data type; version; sampling frequency; sampling weight; model weight in the case of linking / merging with another model; labeled / unlabeled data; interpretability level; risk level (e.g., unacceptable, high, limited); fairness; robustness; privacy; security; safety; reliability; traceability; ML decision confidence score (a numerical value representing the credibility / quality of a given decision generated by the AI / ML inference function); and / or data value quality score, which is a numerical value representing the credibility / quality of a given observation and measurement type. These parameters can be added to the ML model filter information on top of existing parameters.

[0083] Figure 7An example of a network device 300 according to an embodiment of the present disclosure is shown. The network device 300 is used to implement some embodiments of this document. Some embodiments of the present disclosure can be implemented in the network device 300 using any suitably configured hardware and / or software. The network device 300 may include a memory 301, a transceiver 302, and a processor 303 coupled to the memory 301 and the transceiver 302. The processor 303 may be used to implement the functions, programs, and / or methods described herein. Layers of wireless interface protocols may be implemented in the processor 303. The memory 301 is operatively coupled to the processor 303 and stores various information to operate the processor 303. The transceiver 302 is operatively coupled to the processor 303 and transmits and / or receives radio signals. The processor 303 may include an application-specific integrated circuit (ASIC), other chipsets, logic circuits, and / or data processing devices. Memory 301 may include read-only memory (ROM), random access memory (RAM), flash memory, memory cards, storage media, and / or other storage devices. Transceiver 302 may include baseband circuitry for processing radio frequency signals. When this embodiment is implemented in software, the techniques described herein can be implemented by modules (e.g., programs, functions, etc.) that perform the functions described herein. These modules may be stored in memory 301 and executed by processor 303. Memory 301 may be implemented within processor 303 or external to processor 303, in which case memory 301 may be communicatively coupled to processor 303 in various ways known in the art.

[0084] In some embodiments, memory 301 stores executable instructions that, when executed by a processor, cause processor 303 to perform operations including: discovering confidence capabilities for analytics and / or models, wherein NWDAF is enabled to support MTLF with confidence capabilities for ML models, the NWDAF is enabled to support AnLF with confidence capabilities for analytics, or NRF includes a confidence capability provision for each ML model.

[0085] Figure 8 A communication method for AI / ML operations according to an embodiment of this disclosure is illustrated. Figure 8This is an example of a communication method 400 for AI / ML operations according to embodiments of the present disclosure. The communication method 400 for AI / ML operations is used to implement some embodiments of the present disclosure. Some embodiments of the present disclosure can be implemented in the communication method 400 for AI / ML operations using any suitably configured hardware and / or software. In some embodiments, the communication method 400 for AI / ML operations includes: operation 402, where a network entity discovers credibility capabilities for analytics and / or models, wherein an NWDAF is enabled to support an MTLF with credibility capabilities for ML models, an AnLF with credibility capabilities for analytics, or an NRF includes a credibility capability provision for each ML model.

[0086] Figure 9 A communication device according to an embodiment of the present disclosure is shown. Figure 9 In some embodiments, the communication device 500 includes a determiner 501 for discovering credibility capabilities for analytics and / or models, wherein NWDAF is enabled to support MTLF with credibility capabilities for ML models, the NWDAF is enabled to support AnLF with credibility capabilities for analytics, or NRF includes a credibility capability supply for each ML model.

[0087] In some embodiments, discovering credibility capabilities for analytics and / or models includes: NWDAFs being enabled to support MTLFs with credibility capabilities for ML models. In some embodiments, the NWDAFs supporting MTLFs with credibility capabilities are locally configured with one or more NWDAF IDs and / or one or more analytics IDs. In some embodiments, one or more NWDAF IDs and / or one or more analytics IDs supported by each NWDAF supporting MTLFs with credibility capabilities are configured to retrieve trained ML models and corresponding credibility parameters. In some embodiments, network entities are used to discover one or more NWDAFs supporting MTLFs with credibility capabilities using an NWDAF discovery procedure. In some embodiments, network entities are used to discover NWDAFs supporting MTLFs with credibility capabilities via NRF.

[0088] In some embodiments, an NWDAF supporting MTLF with credibility capabilities includes one or more ML model provisioning services and / or one or more ML credibility services. In some embodiments, when a trained ML model is available for one or more analytics IDs, the NWDAF supporting MTLF with credibility capabilities provides one or more ML model provisioning services and / or one or more ML credibility services during registration in the NRF. In some embodiments, one or more ML model provisioning services and / or one or more ML credibility services include: priorities for trusted AI solutions, untrusted AI solutions, and fallback mechanisms between non-AI solutions to ensure security; and / or one or more analytics IDs corresponding to the trained ML model and ML model filter information for the trained ML model for each analytics ID. In some embodiments, the ML model filter information includes at least one of the following credibility-related parameters: data type; version; sampling frequency; sampling weight; model weight in the case of linking / merging with another model; labeled / unlabeled data; interpretability level; risk level; fairness; robustness; privacy; security; safety; reliability; traceability; ML decision confidence score; and / or data value quality score. In some embodiments, the list of analytics IDs is associated between one or more ML model provisioning services and one or more ML credibility services.

[0089] In some embodiments, discovering credibility capabilities for analytics and / or models includes: NWDAFs being enabled to support AnLFs with credibility capabilities for analytics. In some embodiments, the NWDAFs supporting AnLFs with credibility capabilities are locally configured with one or more NWDAF IDs and / or one or more analytics IDs. In some embodiments, one or more NWDAF IDs and / or one or more analytics IDs supported by each NWDAF supporting AnLFs with credibility capabilities are configured to retrieve trained ML models and corresponding credibility parameters. In some embodiments, network entities are used to discover one or more NWDAFs supporting AnLFs with credibility capabilities using an NWDAF discovery procedure. In some embodiments, network entities are used to discover NWDAFs supporting AnLFs with credibility capabilities via NRF. In some embodiments, credibility-related parameters for each analytics filter information include information based on each analytics ID. In some embodiments, the credibility-related parameters for each analytics filter information include information including: risk level; fairness; robustness; privacy; security; safety; reliability; and / or traceability.

[0090] In some embodiments, discovering credibility capabilities for analytics and / or models includes: the NRF including a credibility capability supply for each ML model. In some embodiments, the NRF includes a credibility capability supply for each ML model when no NWDAF is deployed in the network; when no NWDAF supports MTLF; or when an NWDAF supports MTLF but does not support the associated model ID and / or the model is used by different network functions.

[0091] Some embodiments offer the following business benefits: 1. Solving problems and other issues in the prior art. 2. Handling the credibility of analysis and ML models in 5GC. 3. One of the key mechanisms required is the discovery of corresponding models supporting credibility parameters, and some embodiments of the present invention define how this process related to NWDAF capabilities and direct model provisioning capabilities is accomplished in different situations. Therefore, some embodiments of the present invention enable credibility service discovery capabilities. Some embodiments of this disclosure can be used in many applications. Some embodiments of this disclosure are used by chipset vendors, communication system development vendors, automobile manufacturers including cars, trains, trucks, buses, bicycles, motorcycles, helmets, etc., drones (unmanned aerial vehicles), smartphone manufacturers, communication equipment for public safety, and AR / VR / MR device manufacturers for purposes such as gaming, conferences / seminars, and education. Some embodiments of this disclosure are combinations of "technologies / processes" that can be adopted in video standards to create end products. Some embodiments of this disclosure propose technical mechanisms. At least one proposed solution, method, system, and apparatus of some embodiments of this disclosure can be used with respect to current and / or new / future standards for communication systems (such as UEs, base stations, network equipment, and / or communication systems). Compatible products follow at least one proposed solution, method, system, and apparatus according to some embodiments of this disclosure. The proposed solutions, methods, systems, and apparatus are widely used in UEs, base stations, network equipment, and / or communication systems. Implementations of at least one proposed solution, method, system, and apparatus according to some embodiments of this disclosure are considered to include at least one modification / improvement to the method and apparatus for billing reports for AI / ML operations for standardization.

[0092] Furthermore, once AI / ML adoption begins in 5G / 6G networks, the demand for supporting AI / ML trustworthiness is likely to increase. Reasons may include: local regulations allowing the use of AI / ML services meeting specific levels, such as fairness, in mobile networks; and / or service providers requesting network equipment providers to support specific levels of AI / ML trustworthiness, such as robustness and interpretability. Therefore, some embodiments of this application can serve as a basis for 3GPP standardization (starting from Release 19), thereby implementing a standardized approach to such AI / ML trustworthiness and supporting it within message structures and through standardized network functions. It is not a "final product," but rather part of the network implementation for creating 5G network products.

[0093] Figure 10 This is an example of a computing device 1100 according to an embodiment of the present disclosure. Any suitable computing device can be used to perform the operations described herein. For example, Figure 10 This demonstrates that it can be implemented using any appropriately configured hardware and / or software. Figures 1 to 9 Examples of computing devices 1100 illustrating the apparatus and / or methods shown herein. In some embodiments, computing device 1100 may include processor 1112, which is communicatively coupled to memory 1114 and executes computer-executable program code and / or accesses information stored in memory 1114. Processor 1112 may include a microprocessor, application-specific integrated circuit (“ASIC”), state machine, or other processing device. Processor 1112 may include any of a plurality of processing devices, including one processing device. Such processors may include, or can communicate with, a computer-readable medium storing instructions that, when executed by processor 1112, cause the processor to perform the operations described herein.

[0094] Memory 1114 may include any suitable non-transitory computer-readable medium. Computer-readable media may include any electronic, optical, magnetic, or other storage device capable of providing computer-readable instructions or other program code to a processor. Non-limiting examples of computer-readable media include disks, memory chips, ROM, RAM, ASICs, configured processors, optical storage devices, magnetic tape or other magnetic storage devices, or any other medium from which a computer processor may read instructions. Instructions may include processor-specific instructions generated by a compiler and / or interpreter from code written in any suitable computer programming language, including, for example, C, C++, C#, Visual Basic, Java, Python, Perl, JavaScript, and ActionScript.

[0095] The computing device 1100 may also include a bus 1116. The bus 1116 may communicatively couple one or more components of the computing device 1100. The computing device 1100 may also include multiple external or internal devices, such as input or output devices. For example, the computing device 1100 is shown having an input / output (I / O) interface 1118 that can receive input from one or more input devices 1120 or provide output to one or more output devices 1122. One or more input devices 1120 and one or more output devices 1122 may be communicatively coupled to the I / O interface 1118. The communicative coupling may be implemented in any suitable manner (e.g., via a printed circuit board connection, via a cable connection, via wireless communication, etc.). Non-limiting examples of the input device 1120 include a touchscreen (e.g., one or more cameras for imaging a touch area or a pressure sensor for detecting pressure changes caused by a touch), a mouse, a keyboard, or any other device capable of generating input events in response to physical actions of a user of the computing device. Non-limiting examples of output device 1122 include a liquid crystal display (LCD) screen, an external monitor, a speaker, or any other device that can be used to display or otherwise present output generated by a computing device.

[0096] Computing device 1100 can execute program code that configures processor 1112 to perform the above-mentioned... Figures 1 to 9 One or more of the operations described in some of the embodiments shown. The program code may reside in memory 1114 or any suitable computer-readable medium and may be executed by processor 1112 or any other suitable processor.

[0097] The computing device 1100 may also include at least one network interface device 1124. The network interface device 1124 may include any device or group of devices suitable for establishing wired or wireless data connections to one or more data networks 1128. Non-limiting examples of the network interface device 1124 include Ethernet adapters, modems, etc. The computing device 1100 can transmit messages as electrical or optical signals via the network interface device 1124.

[0098] Figure 11 This is a block diagram of an example communication system 1200 according to an embodiment of the present disclosure. The embodiments described herein can be implemented in the communication system 1200 using any suitably configured hardware and / or software. Figure 11A communication system 1200 is shown, which includes a radio frequency (RF) circuit 1210, a baseband circuit 1220, an application circuit 1230, a memory / storage device 1240, a display 1250, a camera 1260, a sensor 1270, and an input / output (I / O) interface 1280, all of which are at least coupled to each other as shown.

[0099] Application circuitry 1230 may include, for example, but not limited to, circuitry of one or more single-core or multi-core processors. The processor may include any combination of general-purpose processors and special-purpose processors (such as graphics processors, application processors). The processor may be coupled to a memory / storage device and used to execute instructions stored in the memory / storage device to enable various applications and / or operating systems to run on the system. Communication system 1200 may execute program code that configures application circuitry 1230 to perform the above-mentioned... Figures 1 to 9 One or more of the operations described herein. The program code may reside in application circuit 1230 or any suitable computer-readable medium, and may be executed by application circuit 1230 or any other suitable processor.

[0100] The baseband circuit 1220 may include, for example, but not limited to, circuitry of one or more single-core or multi-core processors. The processor may include a baseband processor. The baseband circuit can handle various wireless control functions that enable communication with one or more wireless networks via RF circuitry. These wireless control functions may include, but are not limited to, signal modulation, encoding, decoding, and radio frequency shifting. In some embodiments, the baseband circuit can provide communication compatible with one or more wireless technologies. For example, in some embodiments, the baseband circuit can support communication with the evolved universal terrestrial radio access network (EUTRAN) and / or other wireless metropolitan area networks (WMAN), wireless local area networks (WLAN), and wireless personal area networks (WPAN). Embodiments where the baseband circuit is configured to support wireless communication using more than one wireless protocol may be referred to as a multi-mode baseband circuit.

[0101] In various embodiments, baseband circuit 1220 may include circuitry that operates using signals strictly considered not to be in the baseband frequency range. For example, in some embodiments, the baseband circuitry may include circuitry that operates using signals having an intermediate frequency (IF) between the baseband frequency and the radio frequency (RF). RF circuitry 1210 enables communication with a wireless network using modulated electromagnetic radiation over a non-solid-state medium. In various embodiments, RF circuitry may include switches, filters, amplifiers, etc., to facilitate communication with a wireless network. In various embodiments, RF circuitry 1210 may include circuitry that operates using signals strictly considered not to be in the radio frequency range. For example, in some embodiments, RF circuitry may include circuitry that operates using signals having an intermediate frequency (IF) between the baseband frequency and the radio frequency.

[0102] In various embodiments, the above regarding Figures 1 to 11 The transmitter circuitry, control circuitry, or receiver circuitry discussed in the apparatus and / or methods illustrated may be wholly or partially embodied in one or more of the RF circuitry, baseband circuitry, and / or application circuitry. As used herein, “circuit” may refer to, be part of, or may include: an ASIC, electronic circuitry, processor (shared, dedicated, or grouped) and / or memory (shared, dedicated, or grouped), combinational logic circuitry, and / or other suitable hardware components that provide the described functionality, executing one or more software or firmware programs. In some embodiments, the electronic device circuitry may be implemented in one or more software or firmware modules, or the functionality associated with the circuitry may be implemented by one or more software or firmware modules. In some embodiments, some or all of the components of the baseband circuitry, application circuitry, and / or memory / storage device may be implemented together on a system on chip (SOC). The memory / storage device 1240 may be used to load and store, for example, data and / or instructions for the system. The memory / storage device for one embodiment may include any combination of suitable volatile memory (such as dynamic random access memory, DRAM) and / or non-volatile memory (such as flash memory).

[0103] In various embodiments, I / O interface 1280 may include one or more user interfaces designed to enable user interaction with the system and / or peripheral component interfaces designed to enable peripheral components to interact with the system. User interfaces may include, but are not limited to, physical keyboards or keypads, touchpads, speakers, microphones, etc. Peripheral component interfaces may include, but are not limited to, non-volatile memory ports, universal serial bus (USB) ports, audio jacks, and power interfaces. In various embodiments, sensor 1270 may include one or more sensing devices to determine environmental conditions and / or location information relevant to the system. In some embodiments, sensors may include, but are not limited to, gyroscope sensors, accelerometers, proximity sensors, ambient light sensors, and positioning units. Positioning units may also be part of or interact with baseband and / or RF circuitry to communicate with components of a positioning network, such as Global Positioning System (GPS) satellites.

[0104] In various embodiments, display 1250 may include a display, such as a liquid crystal display (LCD) and a touchscreen display. In various embodiments, communication system 1200 may be a mobile computing device, such as, but not limited to, a laptop, tablet, netbook, ultrabook, smartphone, AR / VR glasses, etc. In various embodiments, the system may have more or fewer components and / or different architectures. Where appropriate, the methods described herein may be implemented as a computer program. The computer program may be stored on a storage medium, such as a non-transitory storage medium.

[0105] Those skilled in the art will understand that each unit, algorithm, and step described and disclosed in the embodiments of this disclosure is implemented using electronic hardware or a combination of software and electronic hardware for a computer. Whether a function operates in hardware or software depends on the application conditions and the design requirements of the technical solution. Those skilled in the art can implement the functionality of each specific application in different ways, and such implementations should not exceed the scope of this disclosure. Those skilled in the art will understand that since the working processes of the above-described systems, devices, and units are substantially the same, the working processes of the systems, devices, and units in the above embodiments can be referred to. For ease of description and simplicity, these working processes will not be described in detail.

[0106] It is understood that the systems, devices, and methods disclosed in the embodiments of this disclosure can be implemented in other ways. The above embodiments are merely exemplary. The division of units is based solely on logical function, and other divisions exist in the implementation. Multiple units or components may be combined or integrated in another system. It is also possible to omit or skip some features. On the other hand, the mutual coupling, direct coupling, or communication coupling shown or discussed operates indirectly or communicatively through some ports, devices, or units in an electrical, mechanical, or other kind of manner.

[0107] The units used for explanation as separate components may be physically separate or not. The units used for display may be physical units or not, i.e., located in one place or distributed across multiple network units. Some or all of the units are used depending on the purpose of the embodiment. Furthermore, each functional unit in each embodiment may be integrated into a physically independent processing unit, or it may be integrated into a processing unit with two or more units.

[0108] If software functional units are implemented, used, and sold as products, they can be stored in a readable storage medium within a computer. Based on this understanding, the technical solutions proposed in this disclosure can be implemented substantially or partially as software products. Alternatively, a portion of a technical solution beneficial to conventional technology can be implemented as a software product. The software product in the computer is stored in a storage medium that includes multiple commands for a computing device (such as a personal computer, server, or network device) to execute all or some of the steps disclosed in the embodiments of this disclosure. The storage medium includes a USB flash drive, external hard drive, ROM, RAM, floppy disk, or other types of media capable of storing program code.

[0109] While this disclosure has been described in conjunction with embodiments that are considered to be the most practical and preferred, it should be understood that this disclosure is not limited to the disclosed embodiments, but is intended to cover various arrangements made without departing from the broadest interpretation of the appended claims.

Claims

1. A communication method for artificial intelligence (AI) / machine learning (ML) operations, wherein, The method is implemented in network entities and includes: Discover the credibility of the analysis and / or model; Specifically, the Network Data Analysis Function (NWDAF) is enabled to support the Model Training Logic Function (MTLF) with the credibility capability for the ML model, the NWDAF is enabled to support the Analysis Logic Function (AnLF) with the credibility capability for analysis, or the Network Repository Function (NRF) includes the provision of credibility capability for each ML model.

2. The method according to claim 1, wherein, The network entity is the NWDAF or NWDAF consumer.

3. The method according to claim 1 or 2, wherein, The NWDAF that supports the MTLF with the aforementioned trust capability is locally configured with one or more NWDAF identifiers (IDs) and / or one or more analytics IDs.

4. The method according to claim 3, wherein, Configure the one or more NWDAF IDs and / or support the one or more analysis IDs supported by each NWDAF of the MTLF with the credibility capability to retrieve the trained ML model and the corresponding credibility parameters.

5. The method according to any one of claims 2 to 4, wherein, The network entity is used to discover one or more NWDAFs that support the MTLF with the aforementioned trustworthiness using the NWDAF discovery procedure.

6. The method according to any one of claims 2 to 5, wherein, The network entity is used to discover the NWDAF that supports the MTLF with the aforementioned trustworthiness capability through the Network Repository Function (NRF).

7. The method according to any one of claims 2 to 6, wherein, The NWDAF that supports the MTLF with the aforementioned credibility capability includes one or more ML model provisioning services and / or one or more ML credibility services.

8. The method according to claim 7, wherein, When a trained ML model is available for one or more analytics IDs, the NWDAF supporting the MTLF with the said credibility capability provides the one or more ML model provisioning services and / or the one or more ML credibility services during the registration period in the NRF.

9. The method according to claim 7 or 8, wherein, The one or more ML model provisioning services and / or the one or more ML trust services include: priorities for trusted AI solutions, untrusted AI solutions, and fallback mechanisms between non-AI solutions to ensure security; and / or one or more analytics IDs corresponding to the trained ML model and ML model filter information for the trained ML model for each analytics ID.

10. The method according to claim 9, wherein, The ML model filter information includes at least one of the following confidence-related parameters: data type; version; sampling frequency; sampling weight; Model weights in cases of linking / merging with another model; labeled / unlabeled data; interpretability level; risk level; Fairness; robustness; privacy; security; Security; reliability; traceability; ML decision confidence score; and / or data value quality score.

11. The method according to any one of claims 7 to 10, wherein, The list of analytics IDs is associated between the one or more ML model provisioning services and the one or more ML credibility services.

12. The method according to any one of claims 1 to 11, wherein, The NWDAF that supports the AnLF with the aforementioned trust capability is locally configured with one or more NWDAF IDs and / or one or more analytics IDs.

13. The method according to claim 12, wherein, Configure the one or more NWDAF IDs and / or support the one or more analysis IDs supported by each NWDAF of the AnLF with the credibility capability to retrieve the trained ML model and the corresponding credibility parameters.

14. The method according to any one of claims 1 to 13, wherein, Network entities are used to discover one or more NWDAFs that support the AnLF and have the aforementioned trustworthiness using the NWDAF discovery procedure.

15. The method according to any one of claims 1 to 14, wherein, The network entity is used to discover the NWDAF that supports the AnLF with the aforementioned trustworthiness capability via the NRF.

16. The method according to claim 15, wherein, For each analysis ID, the credibility-related parameters for each analysis filter information include information.

17. The method according to claim 16, wherein, The credibility-related parameters for each analysis filter include: risk level; fairness; robustness; privacy; security; safety; reliability; and / or traceability.

18. The method according to any one of claims 1 to 17, wherein, In the absence of NWDAF deployed in the network; in the absence of NWDAF supporting the MTLF; or in the presence of NWDAF supporting the MTLF but not supporting the associated model ID and / or the model being used by different network functions, the NRF includes the provision of credibility capabilities for each ML model.

19. A communication device, comprising: Determiner, used to discover the credibility of an analysis and / or model; Specifically, the Network Data Analysis Function (NWDAF) is enabled to support the Model Training Logic Function (MTLF) with the credibility capability for the ML model, the NWDAF is enabled to support the Analysis Logic Function (AnLF) with the credibility capability for analysis, or the Network Repository Function (NRF) includes the provision of credibility capability for each ML model.

20. The communication device according to claim 19, wherein, The determiner is the NWDAF or the NWDAF consumer.

21. The communication device according to claim 19 or 20, wherein, The NWDAF that supports the MTLF with the aforementioned trust capability is locally configured with one or more NWDAF identifiers (IDs) and / or one or more analytics IDs.

22. The communication device according to claim 21, wherein, Configure the one or more NWDAF IDs and / or support the one or more analysis IDs supported by each NWDAF of the MTLF with the credibility capability to retrieve the trained ML model and the corresponding credibility parameters.

23. The communication device according to any one of claims 19 to 22, wherein, The determiner is used to discover one or more NWDAFs that support the MTLF with the said credibility capability using the NWDAF discovery procedure.

24. The communication device according to any one of claims 19 to 23, wherein, The determiner is used to discover, via the Network Repository Function (NRF), the NWDAF that supports the MTLF with the aforementioned trustworthiness capability.

25. The communication device according to any one of claims 19 to 24, wherein, The NWDAF that supports the MTLF with the aforementioned credibility capability includes one or more ML model provisioning services and / or one or more ML credibility services.

26. The communication device according to claim 25, wherein, When a trained ML model is available for one or more analytics IDs, the NWDAF supporting the MTLF with the said credibility capability provides the one or more ML model provisioning services and / or the one or more ML credibility services during the registration period in the NRF.

27. The communication device according to claim 25 or 26, wherein, The one or more ML model provisioning services and / or the one or more ML trust services include: priorities for trusted AI solutions, untrusted AI solutions, and fallback mechanisms between non-AI solutions to ensure security; and / or one or more analytics IDs corresponding to the trained ML model and ML model filter information for the trained ML model for each analytics ID.

28. The communication device according to claim 27, wherein, The ML model filter information includes at least one of the following confidence-related parameters: data type; version; sampling frequency; sampling weight; Model weights in cases of linking / merging with another model; labeled / unlabeled data; interpretability level; risk level; Fairness; robustness; privacy; security; Security; reliability; traceability; ML decision confidence score; and / or data value quality score.

29. The communication device according to any one of claims 25 to 28, wherein, The list of analytics IDs is associated between the one or more ML model provisioning services and the one or more ML credibility services.

30. The communication device according to any one of claims 19 to 29, wherein, The NWDAF that supports the AnLF with the aforementioned trust capability is locally configured with one or more NWDAF IDs and / or one or more analytics IDs.

31. The communication device according to claim 30, wherein, Configure the one or more NWDAF IDs and / or support the one or more analysis IDs supported by each NWDAF of the AnLF with the credibility capability to retrieve the trained ML model and the corresponding credibility parameters.

32. The communication device according to any one of claims 19 to 31, wherein, The determiner is used to discover one or more NWDAFs that support the AnLF with the said credibility capability using the NWDAF discovery procedure.

33. The communication device according to any one of claims 19 to 32, wherein, The determiner is used to discover, via the NRF, the NWDAF that supports the AnLF with the credibility capability.

34. The communication device according to claim 33, wherein, For each analysis ID, the credibility-related parameters for each analysis filter information include information.

35. The communication device according to claim 34, wherein, The credibility-related parameters for each analysis filter include: risk level; fairness; robustness; privacy; security; safety; reliability; and / or traceability.

36. The communication device according to claim 35, wherein, In the absence of NWDAF deployed in the network; in the absence of NWDAF supporting the MTLF; or in the presence of NWDAF supporting the MTLF but not supporting the associated model ID and / or the model being used by different network functions, the NRF includes the provision of credibility capabilities for each ML model.

37. A network device, comprising: Memory; transceiver; as well as A processor coupled to the memory and the transceiver; The network device is used to perform the method according to any one of claims 1 to 18.

38. A non-transitory machine-readable storage medium having instructions stored thereon that, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 18.

39. A chip, comprising: A processor for calling and running a computer program stored in memory to cause a device on which the chip is mounted to perform the method according to any one of claims 1 to 18.

40. A computer-readable storage medium storing a computer program, wherein, The computer program causes the computer to perform the method according to any one of claims 1 to 18.

41. A computer program product comprising a computer program, wherein, The computer program causes the computer to perform the method according to any one of claims 1 to 18.

42. A computer program, wherein, The computer program causes the computer to perform the method according to any one of claims 1 to 18.

Images