Enabling data labels discovery and selection for transfer learning

Abstract

A network equipment for wireless communication is described. The network equipment may be configured to, be capable of, or be operable to perform, one or more operations as described herein. For example, the network equipment may be configured to, be capable of, or be operable to: receive a registration request associated with a first peer network equipment, wherein the registration request includes information comprising an identifier of the first peer network equipment and data label capability information; and store the information.

Description

ENABLING DATA LABELS DISCOVERY AND SELECTION FOR TRANSFER LEARNINGTECHNICAL FIELD

[0001] The present disclosure relates to wireless communications, and more specifically to enabling data labels discovery and selection for transfer learning.BACKGROUND

[0002] A wireless communications system may include one or multiple network communication devices, which may be otherwise known as network equipment (NE), supporting wireless communications for one or multiple user communication devices, which may be otherwise known as user equipment (UE), or other suitable terminology. The wireless communications system may support wireless communications with one or multiple user communication devices by utilizing resources of the wireless communication system (e.g., time resources (e.g., symbols, slots, subframes, frames, or the like) or frequency resources (e.g., subcarriers, carriers, or the like). Additionally, the wireless communications system may support wireless communications across various radio access technologies including third generation (3G) radio access technology, fourth generation (4G) radio access technology, fifth generation (5G) radio access technology, among other suitable radio access technologies beyond 5G (e.g., sixth generation (6G)).SUMMARY

[0003] An article “a” before an element is unrestricted and understood to refer to “at least one” of those elements or “one or more” of those elements. The terms “a,” “at least one,” “one or more,” and “at least one of one or more” may be interchangeable. As used herein, including in the claims, “or” as used in a list of items (e.g., a list of items prefaced by a phrase such as “at least one of’ or “one or more of’ or “one or both of’) indicates an inclusive list such that, for example, a list of at least one of A, B, or C means A or B or C or AB or AC orBC or ABC (i.e., A and B and C). Also, as used herein, the phrase “based on” shall not be construed as a reference to a closed set of conditions. For example, an example step that is described as “based on condition A” may be based on both a condition A and a condition B without departing from the scope of the present disclosure. In other words, as used herein, the phrase “based on” shall be constmed in the same manner as the phrase “based at least in part on. Further, as used herein, including in the claims, a “set” may include one or more elements.

[0004] A network equipment for wireless communication is described. The network equipment may be configured to be capable of, or be operable to perform, one or more operations as described herein. For example, the network equipment may include at least one memory and at least one processor coupled with the at least one memory and configured to cause the network equipment to receive a registration request associated with a first peer network equipment, wherein the registration request includes information comprising an identifier of the first peer network equipment and data label capability information, and store the information in the at least one memory as stored data label capability information

[0005] In some implementations of the network equipment and method described herein, said network equipment may be a network repository function and wherein the at least one processor is configured to cause the NRF to receive the registration request.

[0006] In some implementations of the network equipment and method described herein, said network equipment operates in a 5G System.

[0007] In some implementations of the network equipment and method described herein, said first peer network equipment may be a data label producer or data label repository.

[0008] In some implementations of the network equipment and method described herein, said first peer network equipment may be one of an analytics training model function or a data repository or both.

[0009] In some implementations of the network equipment and method described herein, said data label capability information may be information relating to machine learning.

[0010] In some implementations of the network equipment and method described herein, the data label capability information included in the registration information for the first peer network equipment may include one or more of:a flag that indicates whether the first peer network equipment is capable of providing data labels; a source identifier, or an owner identifier, or both associated with a data label capability; an analytics service identifier or a machine learning model identifier, or both, wherein the analytics service identifier or the machine learning model identifier or both are indicative of one or more analytic types applicable for the data labels; a data set identifier; an indication of whether the data labels can be used; a network slice identifier or an edge data network identifier, or both associated with the first peer network equipment; an allowance of permission of a data label capability; or an indication of interoperability of a data label capability.

[0011] In some implementations of the network equipment and method described herein, the processor may be configured to cause the network equipment to: receive a discovery request from a second peer network equipment associated with a data label capability, wherein the discovery request includes requested data label capability information; authorize the second peer network equipment to access the stored information associated with the first peer network equipment; determine whether a condition is satisfied based at least in part on the requested data label capability information and the stored data label capability information; and transmit, to the second peer network equipment, the identifier of the first peer network equipment based at least in part on whether the condition is satisfied.

[0012] In some implementations of the network equipment and method described herein, the processor may be configured to receive and store data label capability information and respective identities for a plurality of first peer network entities as stored data label capability information and stored identities.

[0013] In some implementations of the network equipment and method described herein, the processor may be configured to authorize the second peer network equipment to access the stored information associated with the plurality of first peer network equipments, determinewhether a condition is satisfied based at least in part on the requested data label capability information and the stored data label capability information for each of the first peer network entities, transmit the stored identities of the plurality of first peer network entities to the second peer network equipment.

[0014] In some implementations of the network equipment and method described herein, wherein the condition comprises a determination of whether the requested data label capability information includes information that matches one or a plurality of elements of the data label capability information.

[0015] According to another example, the network equipment may comprise at least one memory and at least one processor coupled with the at least one memory and configured to cause the network equipment to operate as an analytics entity for machine learning model training or as a data repository entity for storing one or more of data, models, and data labels; and transmit a registration request comprising information associated with the network equipment, wherein the information includes an identifier of the network equipment and data label capability information of the network equipment.

[0016] In some implementations of the network equipment and method described herein, the network equipment may be a data label producer or data label repository.

[0017] In some implementations of the network equipment and method described herein the network equipment may be one of an analytics training model function and a data repository.

[0018] In some implementations of the network equipment and method described herein, said data label capability information may be information relating to machine learning model.

[0019] According to another example, the network equipment may comprise at least one memory and at least one processor coupled with the at least one memory and configured to cause the network equipment to transmit a discovery request associated with a data label capability, wherein the discovery request includes requested data label capability information, receive an identifier of a first peer network equipment supporting the requested data label capability, and exchange data label information with the first peer network equipment.

[0020] In some implementations of the network equipment and method described herein the processor may be configured to cause the network equipment to receive the identifier of thefirst peer network equipment as one of a plurality of peer network identifiers and selecting the identifier of the first peer network identity from the plurality of peer network identifiers.

[0021] In some implementations of the network equipment and method described herein, the network equipment may be an analytics training model function.

[0022] In some implementations of the network equipment and method described herein, wherein data labels may be stored according to at least one of the following: together with data set information that relating data labels with a data set; together with analytics model information that relates data labels with an analytics model.

[0023] A method performed by a network equipment is described. The method may include receiving a registration request associated with a first peer network equipment, wherein the registration request includes information comprising an identifier of the first peer network equipment and data label capability information, and storing the information in the at least one memory as stored data label capability information.

[0024] In some implementations of the network equipment and method described herein, the method is repeated for multiple first peer network equipment and respective registration requests.

[0025] In the context of this disclosure a data label capability information may be information that identifies a capability, of a given NE, to provide data labels associated with a given AL / ML analytics (model).BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 illustrates schematically a network comprising a number of Network Data Analytics Function variants and their respective input data sources and output result consumers;

[0027] Figure 2 illustrates schematically a Transfer Learning (TL) architecture involving a knowledge transfer, i. e. , the transfer of an Artificial Intelligence / Machine Learning Model X or related parameters, among two different Analytics Tasks that use different input data sets to produce different outputs;

[0028] Figure 3 illustrates a procedure for triggering and creating new data labels;

[0029] Figure 4 illustrates a procedure for registering and discovering data labels;

[0030] Figure 5 illustrates a data label synchronization or detail exchange procedure, to allow a Data Label Service Consumer and potential Data Label Service Producers to exchange information;

[0031] Figure 6 illustrates an example of a wireless communications system in accordance with aspects of the present disclosure;

[0032] Figure 7 illustrates an example of a NE in accordance with aspects of the present disclosure; and

[0033] Figure 8 illustrate a flowchart of a method performed by a NE in accordance with aspects of the present disclosure.DETAILED DESCRIPTION

[0034] Network analytics and Artificial Intelligence / Machine Learning (AI / ML) are deployed in the 5G core network via the introduction of a network data analytics function (NWDAF). Support for various analytics types is elaborated in TS 23.288. Each deployed NWDAF may support one or more Analytics IDs and may have the role of inference called “NWDAF AnLF”, or training called “NWDAF MTLF” or both. An AnLF that supports a specific Analytics ID inference subscribes to a corresponding model training logical function (MTLF) that is responsible for training.

[0035] Whilst it is advantageous to make use of Transfer Learning (TL) is such networks, there is currently no mechanism to facilitate the transfer of associate data labels between NEs such as NWDAF MTLFs.

[0036] Figure 1 illustrates a number of NWDAF variants and their respective input data sources and output result consumers, which may include 5G core NFs, AFs, 5G core repositories, e.g., Network Repository Function (NRF), User Data Manager (UDM), etc., and the Operations, Administration and Maintenance (0AM) (MnS Consumer or MF). The MTLF and AnLF may exchange AI / ML models, e.g., via the means of serialization or containerization. Optionally, a Data Collection Coordination Function (DCCF) and Messaging Framework Adapter Function (MFAF) may be involved to distribute and collect repeated data towards or from various data sources.

[0037] Current AI / ML mechanisms adopt an isolated training approach, where each AI / ML model is independently trained for a specific usage without considering prior training knowledge or experience. In other words, the instantiation of an AI / ML model requires training from “scratch”, assuming of course that the corresponding NWDAF MTLF is able to collect sufficient input data. In cases where the collected input data is not sufficient, this can be reflected in the confidence degree parameter that is provided to the analytics consumer when the requested output data is supplied to that consumer. The confidence degree indicates the “quality” of the data output (i.e., mainly for a predictive data output) that the NWDAF MTLF has produced, considering the amount of collected input data that is ideally needed.

[0038] Collecting sufficient input data can prove to be challenging, especially when dealing with a newly installed analytics service; for example it can be too expensive considering the required network and computing resources. Training an ML model from scratch may take time and can impact the analytics service performance, especially when the requested time for AI / ML model training is limited.

[0039] In a network and service architecture, knowledge is an asset, and hence it is desirable to preserve and transfer knowledge when applicable. This is the objective of so-called Transfer Learning (TL). TL is a technique that aims to share knowledge by reusing a pretrained AI / ML model or parameters, related to an AI / ML model and which was prepared with respect to a certain task, for a new task, hence exploiting the knowledge gained a priori. An overview of the TL concept is illustrated in Figure 2 which shows a knowledge transfer, e.g., the transfer of an AI / ML Model X or related parameters, among two different Analytics Tasks that use different input data sets to produce different outputs.

[0040] TL can reuse a pre-trained AI / ML model: (i) as trained or partially trained, i.e., some of its parameters, (ii) as a starting point for further training with respect to a new task, e.g., using new data, or (iii) modified with respect to a new task or environment. AI / ML knowledge-based TL refers to a technique where the knowledge gained from training of one or more AI / ML models is applied or adapted to improve or develop another AI / ML model. A challenge with TL is to decide which part of the knowledge is beneficial to be transferred to assist a target task; this decision can be based on: (i) the relationship between the source and target tasks considering, e.g., the same or similar tasks, if they share the same features ornot, (ii) the relationship of the environments in which the source and target tasks exist, e.g., are the domain of the source and target the same, share similar characteristics or are different, and (iii) the similarity and characteristics of data used for training in the source and target environments.

[0041] The task and environment relations among the source and target tasks can reveal whether some knowledge may be common such that knowledge sharing may improve the performance of a target domain or task, and identify which knowledge can be transferred and under which circumstances. In other words, TL aims to extract the knowledge from one or more source tasks and apply the knowledge to a target task. Knowledge transfer may be contained in: (i) a context transfer, which describes how to apply the knowledge gained, (ii) a shared AI / ML model characteristic, e.g., a feature or parameter, (iii) a transformed AI / ML model characteristic, (iv) a relation mapping with respect to the applied network environment or usage.

[0042] The concept of knowledge reflects a developed “skill” or ability that can be adopted to a target task. In other words, knowledge is a quality that can be derived initially before being applied, considering commonality and the type of skills that can be re-used. Knowledge is distinct from, for example, AI / ML model parameters and model input and output data.

[0043] TL is currently used in 5G systems in several scenarios. Registering an AI / ML model in a repository for further use and employing a model profile can accelerate the adoption of TL. The AI / ML model profile may carry information related to the AI / ML model identifier, the preparation phase, i.e., considering which data sources were used, the time schedule and duration of training, the geographical area and objects used as well as some post training information including how to validate and test the AI / ML model. With respect to TL, the AI / ML model profile may carry information reflecting how a trained AI / ML model can be transformed to be re-used for another specific task.

[0044] In the application layer, TL has been considered in TR 23.700-82 where an AI / ML model can be exchanged directly among the source and target analytics entities or stored and retrieved from a model registry. Registering an AI / ML model for TL may include the following information: (i) a task identifier, (ii) UE or group of UEs, (iii) service filter information, e.g., geographical area, time, environment, usage, and charging, (iv) modelcontext information, e.g., set of features, training data requirements, required computing power and (v) permissions, i.e., access, exposure, and vendor.

[0045] In addition, the notion of a TL enabler, i.e., to detect the need for using TL for a certain AI / ML service or task, is already in place. A subscriber may provide the necessary model context information including the model type, features, dataset requirements, desired confidence level, UE(s), service profile, location, and time of interest. The TL enabler can discover, fetch a model from the repository, and determine whether such model shall be used as a pre-trained model considering the similarity in its capabilities, input data and previous usage experience.

[0046] To date, TL considers as knowledge the AI / ML model transfer, parameters or feature transfer or other model related information or capabilities transfer. Although the notion of experience contains the idea of knowledge transferred to a new task, this knowledge does not contain data labels that can be used to characterize data related to AI / ML model training.

[0047] Data labels are meta-data that can be attached to data assets to help classify data. They are tags or annotations added to raw data to help machine learning (ML) and artificial intelligence (Al) models understand and learn from it. Alternatively or additionally, data labels provide context and categorization for AI / ML models to interpret data effectively.

[0048] Data labelling is the activity of assigning context or meaning to data that may vary with respect to the analytics service, e.g., in the case of network load, different labels may be used for analytics related to service experience where the high load may mean congestion, or energy saving where the high load may mean no energy saving is applicable. A data label in the context of a model for predicting a future network load may for example indicate that data is associated with network overload or failure condition.

[0049] Currently there is no mechanism for making data labels visible or available to be discovered for the purpose of sharing the data labels among analytics functions responsible for AI / ML model training.

[0050] Lor the purpose of the following discussion, the term Artificial Intelligence (Al) may be used to define systems that are capable of learning from data to perform tasks without being explicitly programmed. Machine Learning (ML) may be considered a class of Al, although the terms Al ad ML may be used alternately and together.

[0051] A solution presented here relates to an apparatus and method that introduces a data label transfer capability among analytic services that perform AI / ML model training, e.g., among NWDAF MTLFs. Specifically, a data label transfer capability may be introduced by allowing the data label owner, i.e., the NWDAF MTLF that owns data labels or the source that produces data labels or the entity that stores data labels, to register its identity and associated data label capability with a data label repository, that is a NE having memory to store data label capabilities, e.g., a NRF.

[0052] In general, the task of labelling data may take time, resources, and may need an experienced human or specialized code that assures the desired outcome when input data sets are used for AI / ML model training. Labels can be shared considering specific strategies, e.g., among data samples that share common data sets and / or environmental conditions or among the same analytics service types.

[0053] Transferring specific labels may prove useful when a problem, e.g., a trained AI / ML model is not performing as expected with inaccuracy, arises not from a lack of raw data but rather from having correct data labels. Issues with data labels can be caused either due to outdated data labels because of changes in the network conditions, e.g., topology or user movement habits (for instance towards a new training station), or simply because of a lack of data labels, i.e., having no data labels to characterize newly collected input data. The exchange of data labels can also be more efficient since the amount of the data related to data labels transfer can be small.

[0054] New data labels can be created or developed by a Data Label Producer that creates or supplies labels. Such a Data Label Producer can be realized as an analytics function or an analytics service. When a new condition arises, an analytics function or analytics service can characterize data based on “ground truth data”, i.e. real- world data that represents the observed outcome of a system being modeled. When an AI / ML model is characterized as inaccurate and the cause behind this is not a lack of raw data but rather the interpretation of data, i.e., the data labelling, then the ground truth data can assist in the creation of new data labels. The assumption is that the analytics function or analytics service is equipped with a logic / algorithm or function to create new data labels by considering the deviation of new input data from the old input data, as well as the impact of the deviation on analytics results.

[0055] Once new labels are obtained, we need to consider how to make new data labels available to be discovered and selected, when required, by interested analytics functions or analytics services. This may be achieved by registering data labels in a repository, e.g., in the NRF, either together with the NF that owns the data labels or with the NF that stored the data and / or AI / ML models, i.e., in ADRF. NFs that can offer data labels shall introduce a flag in their NF profile registered in the NRF that indicates their capability to provide data labels. The data label owners, e.g., NWDAF MTLF, can also be responsible for authorizing data label consumers, providing: (i) the requested labels either directly if stored locally or (ii) an address and pointer to retrieve data labels from the repository, e.g., ADRF.

[0056] We also need to consider how to store data labels, which may include the options of: (i) storing this data labels together with a specific set of data or data statistics (e.g., data range, max, min, deviation) that describe a data set, (ii) storing it together with the AI / ML model, i.e., including the AI / ML Model ID and / or model type, in where this data label is expected to be used.

[0057] Each data label may characterize one data unit, e.g., a single data value, or a set of data, e.g., specified between a minimum and maximum, or be based on a distribution or on another data statistics parameter. Furthermore, each data label can optionally be associated with at least a single AI / ML Model ID and / or Analytics ID. Each data label or data label set may carry or be associated with metadata to characterize data labels and assist in the discovery, selection and usage. This metadata may contain one or more of the following pieces of information:(i) A data label identity and the version of the label.(ii) A data label owner identity identifying the holder of the label rights.(iii) A data label source ID that produced the data label, which can be the same as the data label owner or different.(iv) An analytics service or task, e.g., NWDAF Analytics ID.(v) The AI / ML model details including an AI / ML model ID or type, feature ID, AI / ML model depth level, etc.(vi) A data set ID or description, e.g., DataSetTag, or including data statistics (e.g., range, distribution, etc.).(vii) A use case or network context or a purpose or intent, e.g., network load, network performance, energy state, network faults, where the label is used.(viii) A target object(s) involved when creating data labels, e.g., UE types, NF type, application type.(ix) A network context, i.e., geographical area, network load, energy conditions, network domain, PLMN ID.(x) A network slice, i.e., Single - Network Slice Selection Assistance Information (S-NSSAI), or network slice type, where the label was and / or is suggested to be used.(xi) An edge data network, e.g., Data Network Name (DDN), where the label was and / or is suggested to be used.(xii) An event or set of events associated with the adoption of a data label, e.g., upon a UE movement, or for stationary UEs.(xiii) A data label usage rating in terms of the performance accuracy that can be modelled as a score percentage including optionally deviation or as true positive / negative or false positive / negative.(xiv) Details related to data label transfer for an inter-related task, i.e., re-weighting with respect to a specified analytics service or task.(xv) Permission related to data label sharing, i.e., allowed consumers, e.g., NF, AF, vendor information or mobile operator information.(xvi) An AI / ML model interoperability indicator to show compatibility of data labels with AI / ML model types or AI / ML model IDs.

[0058] A Data Label Service Consumer, i.e., a network equipment without data labels or with invalid data labels or partially invalid data labels, can be an analytics service or analytics function, e.g., an NWDAF MTLF. A Data Label Service Consumer may need to discover data labels upon realizing that the AI / ML model performance is inaccurate, with the cause not being a lack of raw data but rather the interpretation of data, i.e., the data labelling.

[0059] The Data Label Service Consumer can discover and select data labels from the repository, e.g., NRF or ADRF, by providing a request which contains one or more of the following types of information:(i) a flag that indicates the need to identify an NF that is capable of providing data labels,(ii) a data label identity and a version of the label if this is known or, alternatively(iii) an Analytics ID or AI / ML Model ID or other AI / ML model details, or(iv) a network slice (S-NSSAI) identifier and / or edge data network or DNN identifier, or(v) at least one of the remaining metadata types to characterize data labels and which can assist in identifying, discovering, and selecting the desired data label or data labels set.

[0060] Embodiments may relate to procedures for triggering creation of new data labels, registering data labels as NF profile capability, and discovering and selecting data labels. These are exemplified as follows.

[0061] Triggering and Creating new Data Labels

[0062] Assuming that data labels are: (i) pre-configured by the AI / ML model owner, e.g. , equipment vendor or application provider, or (ii) configured by the operator during a data preparation process, this embodiment introduces the process of new data label creation to anticipate AI / ML model inaccuracy.

[0063] Once the AI / ML model training function 11, i.e., NWDAF MTLF, determines that certain current data labels provide inaccurate results, e.g., the interpretation or meaning of a data label has changed, the function triggers a new data label creation procedure. The process of triggering and creating new data labels is illustrated in Figure 3.

[0064] Steps 1-4 of Figure 3 describe the procedure for MTLF -based AI / ML model accuracy monitoring as per TS 23.288 clause 6.2E.2.

[0065] After obtaining an AI / ML model from the respective NWDAF MTLF 10, the NWDAF AnLF 11 indicates its capability to send Analytics Feedback Information and / or AI / ML Model Accuracy Information from the analytics consumers for the AI / ML Model using Nnwdaf_MLModelMonitoring_Register service operation. The NWDAF MTLF may subscribe to the NWDAF AnLF to get Analytics Feedback Information and / or ML Model Accuracy Information using Nnwdaf_MLModelMonitoring_Subscribe. The NWDAF AnLF may collect Analytics Feedback Information from an Analytics consumer as per TS 23.288 clause 6.1.1 and can then send such Analytics Feedback Information and / or ML ModelAccuracy Information received by invoking Nnwdaf_MLModelMonitor_Notify service operation.

[0066] The NWDAF MTLF considers the feedback from one or more NW DF AnLF or based-on its local policy to determine whether to perform ML Model Accuracy Monitoring and re-training / re-provisioning of AI / ML Model by collecting new data, i.e., ground truth data, from various data sources. The NWDAF MTLF can then compute the accuracy, and hence can check accuracy, using the methods described in TS 23.288 clause 5C.1.

[0067] Steps 5-11 of Figure 3 describe an alternative procedure for AI / ML model accuracy monitoring in which NWDAF MTLF relies on AnLF -assistance to determine AI / ML model accuracy as per TS 23.288 clause 6.2E.3. The NWDAF MTLF sends an Nnwdaf MLModelMonitor Subscribe request to an NWDAF AnLF subscription endpoint, which responds with an acceptance acknowledgment. The NWDAF AnLF may then start monitoring the analytics accuracy of AI / ML Model(s), e.g., by comparing predictions to ground truth data, or comparing changes in internal configuration for the Analytics ID generation or considering previous existent records.

[0068] The NWDAF AnLF determines whether the analytics accuracy of the ML Model is insufficient and send an Nnwdaf_MLModelMonitor_Notify request that includes either Analytics Feedback Information, or the monitored accuracy information of the AI / ML Model. The NWDAF MTLF determines whether the AI / ML Model is degraded or not, based on the notification.

[0069] At a newly introduce step 12, the NWDAF MTLF can then determine the cause of inaccuracy and whether or not any inaccuracy is due to invalid data labels and not a lack of raw data, and whether or not it needs to update data labels that are identified as invalid. NWDAF MTLFs equipped with a logic / algorithm or function can create new data labels by considering the deviation of new input data from the old input data and the impact this may have on analytics results.

[0070] Data Label Service Producer Registration and Discovery

[0071] A Data Label Service Producer 22 that holds and / or creates data labels can register its data label capability with the NRF 21 to allow interested Data Label Service Consumers 20 to discover and select data label capabilities. This procedure is illustrated in Figure 4.

[0072] Data Label Service Producers can be: (i) an NWDAF MTLF that creates (data label owner) or holds data labels that are not restricted from being re-distributed, or that needs to authenticate requests for data labels that are owned but stored into the ADRF, (ii) an ADRF which holds data labels on behalf of data label owners inside and / or outside the PLMN.

[0073] The illustrated data label capability registration process includes the following steps:1. The Data Label Service Producer sends a Nnrf_NFManagement_NFRegister Request message to the NRF to inform the NRF of its NF profile, including its data label capability, when the Data Label Service Producer becomes operative for the first time or updates its data labels capability if this has been modified.The data label capability shall be a flag registered in the NRF. When the flag indicates that an NF has the capability to provide data labels it may include minimum details, which are static, i.e., do not change frequently. Such minimum details may contain at least one of the following: a Data Label ID (to address a request targeting a known Data Label ID) or a specific Analytics ID or AI / ML Model ID, Feature ID, and optionally permissions and interoperability information.Alternatively, permissions may be derived by considering the Data Label Service Producers and the Data Label Service Consumer information and interoperability by examining the desired AI / ML Model ID and Feature ID.2. The NRF stores the NF profile of the Data Label Service Producer including its data label capability flag and optionally the additional details and marks the Data Label Service Producer as available.3. The NRF acknowledge that Data Label Service Producer NF profile Registration is accepted including its data label capability flag via Nnrf_NFManagement_NFRegister response.The illustrated data label capability discovery process includes the following steps:4. The Data Label Service Consumer that needs to discover data label services available in the network invokes Nnrf_NFDiscovery_Request (indicating a need for data label capability, Data Label ID if applicable, Analytics ID or AI / ML Model ID, Feature ID and optionally permission and interoperability) from an appropriate configured NRF in the same PLMN.5. The NRF authorizes the Nnrf_NFDiscovery_Request. The NRF determines whether the Data Label Service Consumer is allowed to discover the Data Label capability of specific Data Label Service Producer instance(s) considering the indicated permissions in the request or considering the Data Label Service Consumer type. If the Data Label Service Producer instance(s) are deployed in a certain network slice, the NRF authorizes the discovery request according to the discovery configuration of the Network Slice, e.g., the expected Data Label Service Producer instance(s) are only discoverable by the NF in the same network slice.6. If allowed, the NRF determines a set of potential Data Label Service Producer NF instance(s) capable of providing data labels matching the Nnrf NFDiscovery Request and sends them to the Data Label Service Consumer.If the Data Label Service Consumer provided a preferred geographical target location, the NRF shall not limit the set of discovered Data Label Service Producer NF instances to the target location if no other alternatives could be found.

[0074] Data Label Service Producer Selection Process

[0075] Once potential Data Label Service Producers are discovered, the Data Label Service Consumer shall select at least one Data Label Service Producer that serves, or is closest to serving, its needs, i.e., to replace or partially replace its invalid data labels. To accomplish this the Data Label Service Consumer shall consider the data sets and / or AI / ML Model information related to the data labels available in each Data Label Service Producer.

[0076] Since new data is collected continuously, data sets related to specific data labels can be renewed. Similarly, the use of data labels may be adopted into new AI / ML models. Hence it is better to consider a data label synchronization or detail exchange step as illustrated in Figure 5, to allow a Data Label Service Consumer 30 and potential Data Label Service Producers 31 to exchange this information. In addition, in this step, further metadata that characterize data labels may also be exchanged as this data can otherwise prove to be too “heavy”, i.e. the volume of data too great, to be stored in the NRF. In addition, additional parameters can make searching the data more complex. However certain metadata related to slice or edge data network can also be contained in the NF profile of each Data Label Service Producer, especially when a Data Label Service Producer represents an NWDAF MTLF.

[0077] Referring to the procedure of Figure 5, the following steps are illustrated:0. The Data Label Service Consumer receives a list of potential Data Label Service Producers as a result of the data label capability discovery process, with the data capability flag indicating the Producers’ willingness to offer data labels towards the Data Label Service Consumer.1. The Data Label Service Consumer requests the data sets associated with Data Labels, or a data set description, and further metadata details that characterize the data labels which may include at least one of the AI / ML model details, use case, target objects, network context, network slice, edge data network, rating, transfer weights.The Data Label Service Consumer can also discover how data labels are stored in the respective Data Label Service Producer, i.e., together with data sets, AI / ML models or separate. This may influence the service that can be used to request data labels.It shall be noted that the Data Label Service Consumer may request the most relevant metadata details and not all the details listed.2. The Data Label Service Consumer may select one or more Data Label Service Producers from which to request data labels.In case there is not an exact match, the Data Label Service Consumer may request data labels from the Data Label Service Producers that is closest to satisfying its (more) important requirement(s), e.g., is used in the same use case or involves the same target objects.

[0078] Figure 6 illustrates an example of a wireless communications system 100 in accordance with aspects of the present disclosure. The wireless communications system 100 may include one or more NE 102, one or more UE 104, and a core network (CN) 106. The wireless communications system 100 may support various radio access technologies. In some implementations, the wireless communications system 100 may be a 4G network, such as an LTE network or an LTE- Advanced (LTE-A) network. In some other implementations, the wireless communications system 100 may be a NR network, such as a 5G network, a 5G- Advanced (5G-A) network, or a 5G ultrawideband (5G-UWB) network. In other implementations, the wireless communications system 100 may be a combination of a 4G network and a 5G network, or other suitable radio access technology including Institute of Electrical and Electronics Engineers (IEEE) 802.11 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802.20. The wireless communications system 100 may support radio access technologiesbeyond 5G, for example, 6G. Additionally, the wireless communications system 100 may support technologies, such as time division multiple access (TDMA), frequency division multiple access (FDMA), or code division multiple access (CDMA), etc.

[0079] The one or more NE 102 may be dispersed throughout a geographic region to form the wireless communications system 100. One or more of the NE 102 described herein may be or include or may be referred to as a network node, a base station, a network element, a network function, a network equipment, a radio access network (RAN), a NodeB, an eNodeB (eNB), a next-generation NodeB (gNB), or other suitable terminology. An NE 102 and a UE 104 may communicate via a communication link, which may be a wireless or wired connection. For example, an NE 102 and a UE 104 may perform wireless communication (e.g., receive signaling, transmit signaling) over a Uu interface.

[0080] An NE 102 may provide a geographic coverage area for which the NE 102 may support services for one or more UEs 104 within the geographic coverage area. For example, an NE 102 and a UE 104 may support wireless communication of signals related to services (e.g., voice, video, packet data, messaging, broadcast, etc.) according to one or multiple radio access technologies. In some implementations, an NE 102 may be moveable, for example, a satellite associated with a non-terrestrial network (NTN). In some implementations, different geographic coverage areas 112 associated with the same or different radio access technologies may overlap, but the different geographic coverage areas may be associated with different NE 102.

[0081] The one or more UE 104 may be dispersed throughout a geographic region of the wireless communications system 100. A UE 104 may include or may be referred to as a remote unit, a mobile device, a wireless device, a remote device, a subscriber device, a transmitter device, a receiver device, or some other suitable terminology. In some implementations, the UE 104 may be referred to as a unit, a station, a terminal, or a client, among other examples. Additionally, or alternatively, the UE 104 may be referred to as an Internet-of- Things (loT) device, an Internet-of-Everything (loE) device, or machine-type communication (MTC) device, among other examples.

[0082] A UE 104 may be able to support wireless communication directly with other UEs 104 over a communication link. For example, a UE 104 may support wireless communication directly with another UE 104 over a device-to-device (D2D) communicationlink. In some implementations, such as vehicle-to-vehicle (V2V) deployments, vehicle-to- everything (V2X) deployments, or cellular-V2X deployments, the communication link 114 may be referred to as a sidelink. For example, a UE 104 may support wireless communication directly with another UE 104 over a PC5 interface.

[0083] An NE 102 may support communications with the CN 106, or with another NE 102, or both. For example, an NE 102 may interface with other NE 102 or the CN 106 through one or more backhaul links (e.g., SI, N2, N2, or network interface). In some implementations, the NE 102 may communicate with each other directly. In some other implementations, the NE 102 may communicate with each other or indirectly (e.g., via the CN 106. In some implementations, one or more NE 102 may include subcomponents, such as an access network equipment, which may be an example of an access node controller (ANC). An ANC may communicate with the one or more UEs 104 through one or more other access network transmission entities, which may be referred to as a radio heads, smart radio heads, or transmission-reception points (TRPs).

[0084] The CN 106 may support user authentication, access authorization, tracking, connectivity, and other access, routing, or mobility functions. The CN 106 may be an evolved packet core (EPC), or a 5G core (5GC), which may include a control plane entity that manages access and mobility (e.g., a mobility management entity (MME), an access and mobility management functions (AMF)) and a user plane entity that routes packets or interconnects to external networks (e.g., a serving gateway (S-GW), a Packet Data Network (PDN) gateway (P-GW), or a user plane function (UPF)). In some implementations, the control plane entity may manage non-access stratum (NAS) functions, such as mobility, authentication, and bearer management (e.g., data bearers, signal bearers, etc.) for the one or more UEs 104 served by the one or more NE 102 associated with the CN 106.

[0085] The CN 106 may communicate with a packet data network over one or more backhaul links (e.g., via an SI, N2, N2, or another network interface). The packet data network may include an application server. In some implementations, one or more UEs 104 may communicate with the application server. A UE 104 may establish a session (e.g., a protocol data unit (PDU) session, or the like) with the CN 106 via an NE 102. The CN 106 may route traffic (e.g., control information, data, and the like) between the UE 104 and the application server using the established session (e.g., the established PDU session). The PDUsession may be an example of a logical connection between the UE 104 and the CN 106 (e.g., one or more network functions of the CN 106).

[0086] In the wireless communications system 100, the NEs 102 and the UEs 104 may use resources of the wireless communications system 100 (e.g., time resources (e.g., symbols, slots, subframes, frames, or the like) or frequency resources (e.g., subcarriers, carriers)) to perform various operations (e.g., wireless communications). In some implementations, the NEs 102 and the UEs 104 may support different resource structures. For example, the NEs 102 and the UEs 104 may support different frame structures. In some implementations, such as in 4G, the NEs 102 and the UEs 104 may support a single frame structure. In some other implementations, such as in 5G and among other suitable radio access technologies, the NEs 102 and the UEs 104 may support various frame structures (i.e., multiple frame structures). The NEs 102 and the UEs 104 may support various frame structures based on one or more numerologies.

[0087] One or more numerologies may be supported in the wireless communications system 100, and a numerology may include a subcarrier spacing and a cyclic prefix. A first numerology (e.g., / r=0) may be associated with a first subcarrier spacing (e.g., 15 kHz) and a normal cyclic prefix. In some implementations, the first numerology (e.g., / r=0) associated with the first subcarrier spacing (e.g., 15 kHz) may utilize one slot per subframe. A second numerology (e.g., / r=l) may be associated with a second subcarrier spacing (e.g., 30 kHz) and a normal cyclic prefix. A third numerology (e.g., / r=2) may be associated with a third subcarrier spacing (e.g., 60 kHz) and a normal cyclic prefix or an extended cyclic prefix. A fourth numerology (e.g., / r=3) may be associated with a fourth subcarrier spacing (e.g., 120 kHz) and a normal cyclic prefix. A fifth numerology (e.g., / r=4) may be associated with a fifth subcarrier spacing (e.g., 240 kHz) and a normal cyclic prefix.

[0088] A time interval of a resource (e.g., a communication resource) may be organized according to frames (also referred to as radio frames). Each frame may have a duration, for example, a 10 millisecond (ms) duration. In some implementations, each frame may include multiple subframes. For example, each frame may include 10 subframes, and each subframe may have a duration, for example, a lms duration. In some implementations, each frame may have the same duration. In some implementations, each subframe of a frame may have the same duration.

[0089] Additionally or alternatively, a time interval of a resource (e.g., a communication resource) may be organized according to slots. For example, a subframe may include a number (e.g., quantity) of slots. The number of slots in each subframe may also depend on the one or more numerologies supported in the wireless communications system 100. For instance, the first, second, third, fourth, and fifth numerologies (i.e., / r=0, jU=l , / r=2, jU=3, / r=4) associated with respective subcarrier spacings of 15 kHz, 30 kHz, 60 kHz, 120 kHz, and 240 kHz may utilize a single slot per subframe, two slots per subframe, four slots per subframe, eight slots per subframe, and 16 slots per subframe, respectively. Each slot may include a number (e.g., quantity) of symbols (e.g., OFDM symbols). In some implementations, the number (e.g., quantity) of slots for a subframe may depend on a numerology. For a normal cyclic prefix, a slot may include 14 symbols. For an extended cyclic prefix (e.g., applicable for 60 kHz subcarrier spacing), a slot may include 12 symbols. The relationship between the number of symbols per slot, the number of slots per subframe, and the number of slots per frame for a normal cyclic prefix and an extended cyclic prefix may depend on a numerology. It should be understood that reference to a first numerology (e.g., / i =0) associated with a first subcarrier spacing (e.g., 15 kHz) may be used interchangeably between subframes and slots.

[0090] In the wireless communications system 100, an electromagnetic (EM) spectrum may be split, based on frequency or wavelength, into various classes, frequency bands, frequency channels, etc. By way of example, the wireless communications system 100 may support one or multiple operating frequency bands, such as frequency range designations FR1 (410 MHz - 7.125 GHz), FR2 (24.25 GHz - 52.6 GHz), FR3 (7.125 GHz - 24.25 GHz), FR4 (52.6 GHz - 114.25 GHz), FR4a or FR4-1 (52.6 GHz - 71 GHz), and FR5 (114.25 GHz - 300 GHz). In some implementations, the NEs 102 and the UEs 104 may perform wireless communications over one or more of the operating frequency bands. In some implementations, FR1 may be used by the NEs 102 and the UEs 104, among other equipment or devices for cellular communications traffic (e.g., control information, data). In some implementations, FR2 may be used by the NEs 102 and the UEs 104, among other equipment or devices for short-range, high data rate capabilities.

[0091] FR1 may be associated with one or multiple numerologies (e.g., at least three numerologies). For example, FR1 may be associated with a first numerology (e.g., / r=0),which includes 15 kHz subcarrier spacing; a second numerology (e.g., / r=l), which includes 30 kHz subcarrier spacing; and a third numerology (e.g., / r=2), which includes 60 kHz subcarrier spacing. FR2 may be associated with one or multiple numerologies (e.g., at least 2 numerologies). For example, FR2 may be associated with a third numerology (e.g., / r=2), which includes 60 kHz subcarrier spacing; and a fourth numerology (e.g., / r=3), which includes 120 kHz subcarrier spacing.

[0092] Figure 7 illustrates an example of a NE 200 in accordance with aspects of the present disclosure. The NE 200 may include a processor 202, a memory 204, a controller 206, and a transceiver 208. The processor 202, the memory 204, the controller 206, or the transceiver 208, or various combinations thereof or various components thereof may be examples of means for performing various aspects of the present disclosure as described herein. These components may be coupled (e.g., operatively, communicatively, functionally, electronically, electrically) via one or more interfaces.

[0093] The processor 202, the memory 204, the controller 206, or the transceiver 208, or various combinations or components thereof may be implemented in hardware (e.g., circuitry). The hardware may include a processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), or other programmable logic device, or any combination thereof configured as or otherwise supporting a means for performing the functions described in the present disclosure.

[0094] The processor 202 may include an intelligent hardware device (e.g., a general- purpose processor, a DSP, a CPU, an ASIC, an FPGA, or any combination thereof). In some implementations, the processor 202 may be configured to operate the memory 204. In some other implementations, the memory 204 may be integrated into the processor 202. The processor 202 may be configured to execute computer-readable instructions stored in the memory 204 to cause the NE 200 to perform various functions of the present disclosure.

[0095] The memory 204 may include volatile or non-volatile memory. The memory 204 may store computer-readable, computer-executable code including instructions when executed by the processor 202 cause the NE 200 to perform various functions described herein. The code may be stored in a non-transitory computer-readable medium such the memory 204 or another type of memory. Computer-readable media includes both non- transitory computer storage media and communication media including any medium thatfacilitates transfer of a computer program from one place to another. A non-transitory storage medium may be any available medium that may be accessed by a general-purpose or specialpurpose computer.

[0096] In some implementations, the processor 202 and the memory 204 coupled with the processor 202 may be configured to cause the NE 200 to perform one or more of the functions described herein (e.g., executing, by the processor 202, instructions stored in the memory 204). For example, the processor 202 may support wireless communication at the NE 200 in accordance with examples as disclosed herein. The NE 200 may be configured to support a means for receiving a registration request associated with a first peer network equipment, wherein the registration request includes information comprising an identifier of the first peer network equipment and data label capability information, and storing the information in the at least one memory as stored data label capability information

[0097] The controller 206 may manage input and output signals for the NE 200. The controller 206 may also manage peripherals not integrated into the NE 200. In some implementations, the controller 206 may utilize an operating system such as iOS®, ANDROID®, WINDOWS®, or other operating systems. In some implementations, the controller 206 may be implemented as part of the processor 202.

[0098] In some implementations, the NE 200 may include at least one transceiver 208. In some other implementations, the NE 200 may have more than one transceiver 208. The transceiver 208 may represent a wireless transceiver. The transceiver 208 may include one or more receiver chains 210, one or more transmitter chains 212, or a combination thereof.

[0099] A receiver chain 210 may be configured to receive signals (e.g., control information, data, packets) over a wireless medium. For example, the receiver chain 210 may include one or more antennas for receive the signal over the air or wireless medium. The receiver chain 210 may include at least one amplifier (e.g., a low-noise amplifier (LN A)) configured to amplify the received signal. The receiver chain 210 may include at least one demodulator configured to demodulate the receive signal and obtain the transmitted data by reversing the modulation technique applied during transmission of the signal. The receiver chain 210 may include at least one decoder for decoding the processing the demodulated signal to receive the transmitted data.

[0100] A transmitter chain 212 may be configured to generate and transmit signals (e.g., control information, data, packets). The transmitter chain 212 may include at least one modulator for modulating data onto a carrier signal, preparing the signal for transmission over a wireless medium. The at least one modulator may be configured to support one or more techniques such as amplitude modulation (AM), frequency modulation (FM), or digital modulation schemes like phase-shift keying (PSK) or quadrature amplitude modulation (QAM). The transmitter chain 212 may also include at least one power amplifier configured to amplify the modulated signal to an appropriate power level suitable for transmission over the wireless medium. The transmitter chain 212 may also include one or more antennas for transmitting the amplified signal into the air or wireless medium.

[0101] Figure 8 illustrates a flowchart of a method in accordance with aspects of the present disclosure. The operations of the method may be implemented by a NE as described herein. In some implementations, the NE may execute a set of instructions to control the function elements of the NE to perform the described functions.

[0102] At 300, the method may include receiving a registration request associated with a first peer network equipment, wherein the registration request includes information comprising an identifier of the first peer network equipment and data label capability information. The operations of 300 may be performed in accordance with examples as described herein. In some implementations, aspects of the operations of 300 may be performed by a NE as described with reference to Figure 7.

[0103] At 302, the method may include storing the information in the at least one memory as stored data label capability information. The operations of 302 may be performed in accordance with examples as described herein. In some implementations, aspects of the operations of 302 may be performed by a NE as described with reference to Figure 7.

[0104] It should be noted that the method described herein describes a possible implementation, and that the operations and the steps may be rearranged or otherwise modified and that other implementations are possible.

[0105] The description herein is provided to enable a person having ordinary skill in the art to make or use the disclosure. Various modifications to the disclosure will be apparent to a person having ordinary skill in the art, and the generic principles defined herein may be applied to other variations without departing from the scope of the disclosure. Thus, thedisclosure is not limited to the examples and designs described herein but is to be accorded the broadest scope consistent with the principles and novel features disclosed herein.

Claims

What is claimed is:

1. A network equipment for wireless communication, comprising: at least one memory; and at least one processor coupled with the at least one memory and configured to cause the network equipment to: receive a registration request associated with a first peer network equipment, wherein the registration request includes information comprising an identifier of the first peer network equipment and data label capability information; and store the information in the at least one memory as stored data label capability information.

2. A network equipment according to claim 1, wherein the network equipment comprises a network repository function (NRF), and wherein the at least one processor is configured to cause the NRF to receive the registration request.

3. A network equipment according to claim 1 or 2, the first peer network equipment is a data label producer or a data label repository.

4. A network equipment according to claim 1 or 2, wherein the first peer network equipment is an analytics training model function or a data repository, or both.

5. A network equipment according to any of the preceding claims, the data label capability information being information relating to machine learning.

6. A network equipment according to any of the preceding claims, wherein the data label capability information further includes one or more of: a flag that indicates whether the first peer network equipment is capable of providing data labels; a source identifier, or an owner identifier, or both associated with a data label capability;an analytics service identifier or a machine learning model identifier, or both, wherein the analytics service identifier or the machine learning model identifier or both are indicative of one or more analytic types applicable for the data labels; a data set identifier; an indication of whether the data labels can be used; a network slice identifier or an edge data network identifier, or both associated with the first peer network equipment; an allowance of permission of a data label capability; or an indication of interoperability of a data label capability.

7. A network equipment according to any of the preceding claims, wherein the at least one processor is further configured to cause the network equipment to: receive a discovery request from a second peer network equipment associated with a data label capability, wherein the discovery request includes requested data label capability information; authorize the second peer network equipment to access the stored information associated with the first peer network equipment; determine whether a condition is satisfied based at least in part on the requested data label capability information and the stored data label capability information; and transmit, to the second peer network equipment, the identifier of the first peer network equipment based at least in part on whether the condition is satisfied.

8. A network equipment according to any of the preceding claims, wherein the at least one processor is further configured to cause the network equipment to receive and store data label capability information and respective identities for a plurality of first peer network entities as stored data label capability information and stored identities.

9. A network equipment according to claim 8 when dependent upon claim 7, the at least one processor is further configured to cause the network equipment to: authorize the second peer network equipment to access the stored information associated with the plurality of first peer network equipments;determine whether a condition is satisfied based at least in part on the requested data label capability information and the stored data label capability information for each of the first peer network entities; transmit the stored identities of the plurality of first peer network entities to the second peer network equipment.

10. A network equipment according to any one of claims 7 to 9, wherein the condition comprises a determination of whether the requested data label capability information includes information that matches one or a plurality of elements of the data label capability information.

11. A network equipment for wireless communication, comprising: at least one memory; and at least one processor coupled with the at least one memory and configured to cause the network equipment to: operate as an analytics entity for machine learning model training or as a data repository entity for storing one or more of data, models, and data labels; and transmit a registration request comprising information associated with the network equipment, wherein the information includes an identifier of the network equipment and data label capability information of the network equipment.

12. A network equipment according to claim 11, the network equipment is a data label producer or data label repository.

13. A network equipment according to any of claim 12 to 14, the data label capability information being information relating to a machine learning model.

14. A network equipment for wireless communication, comprising: at least one memory; and at least one processor coupled with the at least one memory and configured to cause the network equipment to:transmit a discovery request associated with a data label capability, wherein the discovery request includes requested data label capability information; receive an identifier of a first peer network equipment supporting the requested data label capability; and exchange data label information with the first peer network equipment.

15. A network equipment according to claim 14, wherein the at least one processor is further configured to cause the network equipment to receive the identifier of the first peer network equipment as one of a plurality of peer network identifiers and selecting the identifier of the first peer network identity from the plurality of peer network identifiers.

16. A network equipment according to claim 15 or 16, the network equipment is an analytics training model function.

17. A method performed by a network equipment for wireless communication, comprising: receiving a registration request associated with a first peer network equipment, wherein the registration request includes information comprising an identifier of the first peer network equipment and data label capability information; and storing the information in the at least one memory as stored data label capability information.

18. A method according to claim 17, the network equipment being a network repository function.

19. A method according to claim 17 or 18, the data label capability information being information relating to machine learning.

20. A method according to any of claims 17 to 18, the method being repeated for multiple first peer network equipment and respective registration requests.

Images