Techniques for managing artificial intelligence models and datasets

Abstract

The application relates to devices and components including apparatuses, systems, and methods for managing artificial intelligence models and datasets.

Description

TECHNICAL FIELD

[0001] The application relates to wireless networks, and more specifically, to techniques for managing artificial intelligence models and datasets within the networks. BACKGROUND

[0002] Third Generation Partnership Project (3GPP) Technical Specifications (TSs) define standards for wireless networks. These TSs describe aspects related to communications between nodes of radio access networks within these wireless networks. BRIEF DESCRIPTION OF DRAWINGS

[0003] Figure 1 A network environment is illustrated in accordance with some embodiments.

[0004] Figure 2 An intelligence framework is illustrated in accordance with some embodiments.

[0005] Figure 3 A signaling procedure is illustrated in accordance with some embodiments.

[0006] Figure 4 Another signaling procedure is illustrated in accordance with some embodiments.

[0007] Figure 5 Another signaling procedure is illustrated in accordance with some embodiments.

[0008] Figure 6 Another signaling procedure is illustrated in accordance with some embodiments.

[0009] Figure 7 A message is illustrated in accordance with some embodiments.

[0010] Figure 8 A message is illustrated in accordance with some embodiments.

[0011] Figure 9 An operational flow / algorithmic structure is illustrated in accordance with some embodiments.

[0012] Figure 10 Another operational flow / algorithmic structure is illustrated in accordance with some embodiments.

[0013] Figure 11 Another operational flow / algorithmic structure is illustrated in accordance with some embodiments.

[0014] Figure 12 User equipment is illustrated in accordance with some embodiments.

[0015] Figure 13 A network node is illustrated in accordance with some embodiments. DETAILED DESCRIPTION

[0016] The following detailed description references the drawings, wherein like numerals indicate the same or similar elements and features. In the following description, for the purposes of explanation and not limitation, specific details are set forth, such as particular structures, architectures, interfaces, and techniques, in order to provide a thorough understanding of the various aspects of various implementations. However, it will be apparent to those skilled in the art having the benefit of the present disclosure that the various aspects of various implementations can be practiced in other examples that depart from these specific details. In certain instances, descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of the various implementations. For the purposes of the present document, the phrases “A / B” and “A or B” mean (A), (B), or (A and B); and the phrase “based on A” means “based at least in part on A,” such that it can be “based on A alone” or it can be “based on A and at least one other thing.”

[0017] The following is a glossary of terms that can be found within the present disclosure.

[0018] As used herein, the term “circuitry” refers to hardware components such as an electronic circuit, a logic circuit, a processor (shared, dedicated, or group) or memory (shared, dedicated, or group), an Application Specific Integrated Circuit (ASIC), a field-programmable device (FPD) (e.g., a field-programmable gate array (FPGA), a programmable logic device (PLD), a complex PLD (CPLD), a high-capacity PLD (HCPLD), a structured ASIC, or a programmable SoC) or digital signal processors (DSPs), as well as the combinations of hardware components and / or software components for providing the described functionality that are collectively referred to as a “circuit.” In some embodiments, the circuitry can execute one or more software or firmware programs to provide at least some of the described functionality. The term “circuitry” can also refer to the combination of hardware components (or combinations of circuitry) with program code that work together to provide the described functionality. In these embodiments, the combination of hardware and program code can be referred to as a “circuit” or a “circuitry.”

[0019] As used herein, the term “processor circuitry” refers to a circuit capable of sequentially and automatically carrying out a sequence of arithmetic or logical operations, or recording, storing, or transferring digital data, is part of, or includes, such a circuit. The term “processor circuitry” can refer to an application processor, a baseband processor, a central processing unit (CPU), a graphics processing unit, a single-core processor, a dual-core processor, a triple-core processor, a quad-core processor, or any other device capable of executing or otherwise operating computer-executable instructions, such as program code, software modules, and / or functional processes.

[0020] As used herein, the term “interface circuitry” refers to, is part of, or includes circuitry that enables the exchange of information between two or more components or devices. The term “interface circuitry” can refer to one or more hardware interfaces, such as a bus, I / O interface, peripheral component interface, and network interface card.

[0021] As used herein, the term “user equipment” or “UE” refers to a device with radio communication capabilities that can allow a user to access network resources in a communication network. The term “user equipment” or “UE” can be considered synonymous, and can be referred to as a client, mobile, mobile device, mobile terminal, user terminal, mobile unit, mobile station, mobile user, subscriber, user, remote station, access agent, user agent, receiver, radio equipment, reconfigurable radio equipment, or reconfigurable mobile device. Moreover, the term “user equipment” or “UE” can include any type of wireless / wired device or any computing device including a wireless communication interface.

[0022] As used herein, the term “computer system” refers to any type of interconnected electronic devices, computer devices, or components thereof. Additionally, the term “computer system” or “system” can refer to various components of a computer that are communicatively coupled to each other. Moreover, the term “computer system” or “system” can refer to multiple computer devices or multiple computing systems that are communicatively coupled to each other and configured to share computing resources or networking resources.

[0023] As used herein, the term “resource” refers to a physical or virtual device, a physical or virtual component within a computing environment, or a physical or virtual component within a particular device, such as a computer device, a mechanical device, memory space, processor / CPU time, processor / CPU usage, processor and accelerator load, hardware time or usage, power, input / output operations, port or network socket, channel / link allocation, throughput, memory usage, storage, network, database, and application or workload unit. A “hardware resource” can refer to a computing resource, storage resource, or network resource provided by a physical hardware element. A “virtualized resource” can refer to a computing resource, storage resource, or network resource provided by a virtualization infrastructure to an application, device, or system. The term “network resource” or “communication resource” can refer to a resource that is accessible by a computer device / system via a communication network. The term “system resource” can refer to any kind of shared entity that provides a service and can include a computing resource or a network resource. A system resource can be considered as a set of coherent functionalities, network data objects, or services that are accessible through a server, where such system resources reside on a single host or multiple hosts and are clearly identifiable.

[0024] As used herein, the term "channel" refers to any tangible or intangible means of conveying a data signal or data stream. The term "channel" can be synonymous with, or equivalent to, "communication channel," "data communication channel," "transmission channel," "data transmission channel," "access channel," "data access channel," "link," "data link," "carrier," "radio frequency carrier," or any other similar terminology denoting a path or medium through which data is conveyed. Additionally, the term "link" as used herein refers to a connection made between two devices for transmitting and receiving information.

[0025] As used herein, the terms "instantiate," "instantiation," and the like mean the creation of an instance. An "instance" refers to a concrete occurrence of an object, which can occur, for example, during execution of program code.

[0026] The term "connect" can mean that two or more elements have an established signaling relationship with each other over a communication channel, link, interface, or reference point at a common protocol layer.

[0027] As used herein, the term "network element" refers to physical or virtualized equipment or infrastructure used to provide wired or wireless communication network services. The term "network element" can be considered synonymous with or otherwise referred to as a networked computer, networked hardware, network equipment, network node, or virtualized network function.

[0028] The term "information element" refers to a structural element containing one or more fields. The term "field" refers to individual content of an information element, or a data element containing content. An information element can include one or more additional information elements.

[0029] Figure 1 A network environment 100 is illustrated in accordance with some embodiments. The network environment 100 can include a user equipment (UE) 104 communicatively coupled with a base station 108 of a radio access network (RAN) 110. In some instances, the base station 108 can be a next generation (NG) RAN node, such as a gNB or ng-eNB. The UE 104 and the base station 108 can communicate over an air interface compatible with 3GPP TS, such as TS defining a fifth generation (5G) New Radio (NR) system or higher system (e.g., a sixth generation (6G) radio system). The base station 108 can provide user plane and control plane protocol terminations towards the UE 104.

[0030] The network environment 100 can also include a core network 112. For example, the core network 112 can include a 5thGeneration Core Network (5GC) or an updated generation core network (e.g., a 6thGeneration Core Network (6GC)). The core network 112 can be coupled to the base stations 108 via fiber or wireless backhaul. The core network 112 can provide functionality to the UEs 104 via the base stations 108. These functionalities can include managing subscriber profile information, subscriber location, service authentication, switching functions for voice and data sessions, and routing and forwarding of user-plane packets between the RAN 110 and external data networks 120.

[0031] In some embodiments, one or more nodes of the network environment 100 can act as an agent that trains an AI model. As used herein, an AI model can include a machine learning (ML) model, a neural network (NN), or a deep learning network.

[0032] In some embodiments, an AI model can play a role in improving network functionality. For example, an AI model can be trained by an AI agent in the network environment 100 and can be used to facilitate decisions made in the RAN 110 or the CN 112. These decisions can be related to beam management, positioning, resource allocation, network management (e.g., operations, administration, and maintenance (OAM) aspects), routing selection, energy saving, load balancing, etc.

[0033] In some embodiments, an AI model can play a role in an AI-as-a-Service (AIaaS) platform. In an AIaaS platform, AI services can be consumed by applications initiated at a user level or a network level, and service providers can be available to any AI agent reachable in the network environment 100.

[0034] In some embodiments, an AI model can be used for one or more use case functionalities, including, for example, channel state information (CSI) feedback enhancement, beam management, or positioning accuracy enhancement. Various protocol aspects can be developed to support these or other use case functionalities While specific AI models can be left to specific implementations, various 3GPP TSs can provide protocol aspects related to AI functionality over the air interface and user data privacy. These protocol aspects can include, for example, aspects related to capability indication, configuration and control procedures (e.g., AI model training and inference), management of data, and management of AI models. AI operations for the air interface can be based on current RAN architectures without the need to introduce new interfaces.

[0035] According to one or more of the following examples, the AI model can be communicated or delivered between different entities of the network environment 100. In a first example, the base station 108 can communicate or deliver the AI model to the UE 104 via radio resource control (RRC) signaling or user plane (UP) data. In a second example, a function of the core network 112, such as, for example, an OAM node, can communicate or deliver the AI model to the UE 104 via non-access stratum (NAS) signaling or UP data. In a third example, a location management function (LMF) of the core network 112 can communicate or deliver the AI model to the UE 104 via a positioning protocol (e.g., long term evolution (LTE) positioning protocol (LPP)) or UP data. In a fourth example, a server of the external data network 120 can communicate or deliver the AI model to the UE 104 in a manner that is transparent to the 3GPP network components.

[0036] Figure 2 An intelligent framework 200 that can be implemented by the network environment 100 is illustrated in accordance with some embodiments. The intelligent framework 200 can include a data collection function 204 that collects data sets that can be model training data or model inference data. The data can include or be based on radio-related measurements, application-related measurements, sensor inputs, feedback from actors 216, and the like. In some embodiments, the data collection 204 can be performed by a component that has access to the air interface (e.g., the base station 108 or the UE 104). The data sets collected by one of the components can be communicated to another component.

[0037] The data collection aspect can be different for different lifecycle management (LCM) functions (e.g., model training, model inference, model monitoring, or model activation / deactivation / selection / switching / updating). For example, relatively large data size and relatively relaxed latency can be used for offline training, while relatively strict latency requirements can be used for model monitoring or inference. Data collection techniques with sufficient security and UE privacy are also needed.

[0038] Embodiments of the present disclosure adapt various quality of experience (QoE) framework principles for the purpose of data collection, as will be described.

[0039] The intelligent framework 200 can include a model training function 208 that receives training data from the data collection function 204. The model training function 208 can be implemented by a component of the network environment (e.g., an OAM node). The model training function 208 can use the training data to perform AI model training, validation, and testing. In some instances, the model training function 208 can perform data preparation (e.g., data pre-processing and cleaning, formatting, or transformation) for a particular AI algorithm.

[0040] In one example, the model training function 208 can train an AI model by determining a plurality of weights to use within layers of a neural network. For example, consider a neural network having an input layer with dimensions matching those of an input matrix constructed from a dataset. The neural network can include one or more hidden layers and an output layer having M x 1 dimensions that outputs an M x 1 codeword. Each of the layers of the neural network can have a different number of nodes, each connected to nodes of an adjacent layer or nodes of a non-adjacent layer. Generally, at some layers, the nodes can produce outputs that are a non-linear function of a sum of their inputs, and provide these outputs to nodes of an adjacent layer over corresponding connections. The set of weights, which can also be referred to as an AI model in this example, can adjust the strength of connections between nodes of adjacent layers. The weights can be set based on a training process with training inputs (generated from the dataset) and desired outputs. The training data can be provided to the AI model, and the weights can be adjusted using a difference between the outputs and the desired outputs. In other implementations, the model training function 208 can train an AI model in other ways. For example, the UE 104 can use training data to determine parameter values of an AI model (such as a decision tree or a simple linear function) in other ways.

[0041] In training an AI model, the model training 208 can use model deployment updates to provide a trained, validated, tested, or updated AI model to a model inference function 212 of the intelligence framework 200. The model inference 212 can also receive inference data from the data collection 204 and generate an output (e.g., a prediction or a decision). The output can be provided to the actor 216. The model inference function 212 can also provide model performance feedback to the model training function 208. In some instances, the model inference function 212 can perform data preparation (e.g., data pre-processing and cleaning, formatting, or transformation) for a particular AI algorithm.

[0042] Unless otherwise described herein, the intelligence framework 200 can operate in line with principles described in 3GPP Technical Report (TR) 37.817 v17.0.0 (2022-04-06).

[0043] Embodiments of the present disclosure provide a control plane solution for managing AI models and datasets. Some embodiments describe various processes including: model identification; offline model training and dataset reporting; and online model training / inference / monitoring. Offline model training can be hosted and managed by an OAM node. Online training can include model training, model inference, and model monitoring processes, which can be hosted by a base station 108 or LMF in cases where the use case function is associated with relatively stricter latency requirements.

[0044] Some embodiments provide signaling aspects to support a control plane solution for managing AI models and datasets. New signaling radio bearers (SRBs) with configurable priority can be used for including AI models, AI model datasets, or QoE datasets. Application layer segmentation can also be described for RRC messages carrying AI related information.

[0045] Some embodiments also provide UE assistance information to provide signaling related to dynamic UE capabilities. For example, the UE assistance information can be used to indicate the presence or absence of a temporary capability reduction at the UE 104.

[0046] Figure 3 A signaling procedure 300 is illustrated that illustrates aspects of AI model identification and registration, according to some embodiments. The signaling procedure 300 can include signals between and operations performed by the UE 104, the base station 108, and the OAM 302. In some embodiments, some or all of the operations of the signaling procedure 300 performed by the base station 108 can be performed by a base station central unit (CU). The OAM 302 can be a node of the network environment 100.

[0047] The signaling procedure 300 can include a capability report to convey access stratum (AS) capability information about AI models supported by the UE 104 to the base station 108. Specifically, at 304, the signaling procedure 300 can include the base station 108 transmitting a UE capability enquiry (RRC UECapabilityInquiry ) message to the UE 104. UECapabilityInquiry The message can request a report of AI models supported by the UE 104. The granularity of the requested capabilities can be per use case function or per AI model. For example, per use case function granularity can request an indication of which AI models are supported for a given use case function; and per model granularity can request an indication of which use case functions are supported for a given AI model. The use case functions can include, but are not limited to, CSI feedback enhancement, beam management, and positioning accuracy enhancement. The AI models can include, but are not limited to, a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, or a reinforcement learning model.

[0048] In response to receiving the UECapabilityInquiry message, the UE 104 can transmit a UE capability information (RRC UECapabilityInformation ) message to the base station 108 that includes a list of supported use case functions or AI models. UECapabilityInformation The message can include the UE capability information with the requested granularity, e.g., per use case function or per AI model.

[0049] In some embodiments, UECapabilityInquiry The message can not specify a particular granularity. In these embodiments, the UE 104 can generate the UECapabilityInformation message in a manner determined by the UE 104 to be appropriate.

[0050] Signaling procedure 300 may also include UE 104 sending a model identifier at 312 ( ModelIdentification Message. Can be sent from base station 108 to OAM 302. ModelIdentification The message is sent to determine whether OAM 302 is capable of training the labeled AI model. ModelIdentification The message may be a new RRC message, comprising a container with a model ID associated with an AI model or a functional ID associated with a use case function supported by UE 104. The container may include model / functional IDs that can be used by UE 104 but are not supported by base station 108. According to some implementations, ModelIdentification The message can be similar to Figure 7 The RRC message 700 shows ModelIdentification 704.

[0051] The model / functional ID can be a globally unique identifier pre-assigned by the manufacturer or the serving public land mobile network (PLMN). The ID can be unique across the PLMN and the vendor.

[0052] In some implementations, a functional ID may be associated with one or more model IDs.

[0053] Received ModelIdentification At point 316, base station 108 can access a list of model / functional IDs from the container and forward that list to OAM 302. In some implementations, instead of accessing the list of model / functional IDs from the container, base station 108 can transparently forward the container itself.

[0054] At position 320, OAM 302 can verify the model ID. When verifying the model ID, OAM 302 can determine that the corresponding AI model is supported by the network. For example, the AI ​​model can be obtained by OAM 302 and trained offline or online by OAM 302 or other components, as described elsewhere in this document. In the case of reported functional IDs, OAM can identify the model IDs associated with the reported functional IDs and verify those model IDs.

[0055] At point 324, OAM 302 can send a list of confirmed model IDs to base station 108. At point 328, base station 108 can generate a container with the list of confirmed model IDs and, in a new RRC message (e.g., model identification response),... ModelIdentificationResponse The container is transmitted to UE 104 in a message. In some implementations, OAM302 can generate a list of containers with confirmed model IDs, and base station 108 can transparently forward the containers to UE 104. According to some implementations, ModelIdentificationResponse The message can be similar toFigure 7 the RRC message 700 shown in ModelIdentificationResponse 708.

[0056] Figure 4 A signaling procedure 400 illustrating aspects of offline model training is exemplified in accordance with some embodiments. The signaling procedure 400 can include signals among and operations performed by the UE 104, the base station 108, an access and mobility management function (AMF) 404, a unified data management function (UDM) 408, and the OAM 302. The AMF 404 and the UDM 408 can be part of the core network 112.

[0057] The signaling procedure 400 can include the OAM 302 downloading the confirmed AI models. The confirmed AI models can be similar to those described above with respect to the signaling procedure 300. The confirmed AI models can be downloaded from, for example, an application server in the external data network 120. Each AI model can require a particular data set for training purposes. Different models can require different data sets. To facilitate offline training, the OAM 302 can perform the following data set reporting configuration procedure based on the confirmed AI models.

[0058] At 416, the OAM 302 can send data collection (DC) configurations to the UDM 408. Each DC configuration can correspond to a confirmed AI model. In some embodiments, the DC configuration can be indicated by a model ID in the message sent to the UDM 408. In other embodiments, the DC configuration can be indicated by a DC configuration ID that maps to the model ID. The DC configuration can include a configuration of how to generate a data set with training or inference data for a particular AI model. The format of the DC configuration can be an application layer format, which is a proprietary format or an open format.

[0059] At 420, the UDM 408 can check UE consent. The UE consent can indicate whether the UE 104 authorizes sharing of a data set corresponding to a particular DC configuration. The UE consent can be defined in user subscription information. The UE consent can be checked for each of the DC configurations. In the event that the UE 104 does not consent to sharing of data sets for one or more DC configurations, at 424, the UDM 408 can send a message to the OAM 302 with an indication of the rejected DC configurations. In the event that the UE 104 does consent to sharing of data sets for one or more DC configurations, at 428, the UDM 408 can send a message to the AMF 404 with an indication of the approved DC configurations.

[0060] Upon receiving the approved DC configuration, the AMF 404 can initiate activation of the data set management / reporting for the approved DC configuration. This can be done by the AMF 404 sending an AI model configuration information element (IE) with an indication of the DC configuration to be activated at 432. The AI model configuration IE can be sent over the N2 interface. Each DC configuration can be uniquely identified by a model ID (or DC configuration ID). Data set management / reporting can be activated for only one DC configuration or for multiple DC configurations at the same time.

[0061] In some embodiments, the AI model configuration IE can be included in an initial context setup request (INITIAL CONTEXT SETUP REQUEST) for establishing a UE context or a UE context modification request (UE CONTEXT MODIFICATION REQUEST) for modifying an established UE context. If the AI model configuration IE is included in the initial context setup request or the UE context modification request, the base station 108 can use it for AI model management if supported.

[0062] The base station 108 can generate a container with the DC configuration to be activated and include the container in a new RRC message (e.g., a data measurement configuration (DATA dataMeasConfig ) message) at 434. The dataMeasConfig message can be sent to the UE 104 at 436.

[0063] DataMeasConfig The message can also include an RRC ID corresponding to each DC configuration. The RRC ID can be uniquely mapped to the model ID. Since the RRC ID does not need to be globally unique like the model ID, it can be significantly shorter than the model ID.

[0064] In some instances, DataMeasConfig The message can also include the type of data set, the configuration reporting (e.g., periodicity or triggers associated with the reporting of the data set), and whether the data set is RAN visible. The data set type can be, for example, a measurement, a live ground tag, or even a different use case (e.g., a data set for CSI compression, a data set for beam management, or a data set for positioning). If the data set is indicated as RAN visible, the base station 108 can be able to use the data set directly. For example, a data set corresponding to a CSI or beamforming use case functionality can be visible to the base station 108; however, a data set corresponding to a positioning use case functionality can not be visible to the base station 108. The indication of which data sets are RAN visible can be provided by the OAM 302 or the AMF 404.

[0065] According to some embodiments, dataMeasConfig The Figure 7The RRC message 700 shows dataMeasConfig 712.

[0066] Received DataMeasConfig When a message is sent, UE 104 can begin collecting a dataset associated with the active DC configuration at 440. The application layer of UE 104 can collect the dataset and then encapsulate it in a transparent container for transmission to the network. UE 104 can then send this dataset in a new RRC message (e.g., a measurement report dataset). MeasurementReportDataset The container is sent to base station 108 in the message. MeasurementReportDataset The message may also include the RRC ID of the corresponding dataset. MeasurementReportDataset The message may include one dataset / RRC ID or multiple dataset / RRC IDs. At 444, the UE may send to base station 108. MeasurementReportDataset Message. According to some implementation plans, MeasurementReportDataset The message can be similar to Figure 7 The RRC message 700 shows MeasurementReportDataset 716.

[0067] Received MeasurementReportDataset At location 448, base station 108 can forward the dataset and associated model ID to OAM 302. The dataset can then be used for offline training of the associated AI model.

[0068] In some implementations, base station 108 can configure UE 104 to adjust reporting for a specific DC configuration. For example, base station 108 can configure UE 104 to suspend or resume reporting for a specific DC configuration. This can be done via RRC signaling.

[0069] When receiving the dataset from base station 108, OAM 302 can perform offline training at 452.

[0070] In the case where OAM 302 expects a dataset from base station 108 for offline training, it can obtain the dataset directly from base station 108 without performing signaling procedure 400.

[0071] Figure 5 Signaling procedure 500 according to some implementation schemes is illustrated, which illustrates various aspects of online model training. Signaling procedure 500 may include signals between UE 104, network node 504, and OAM 302, and the operations performed by them. Network node 504 may host / manage AI model training. Depending on the use case function, network node 504 may be base station 108 or LMF of core network 112.

[0072] Online training associated with signaling procedure 500 may include more stringent latency requirements than offline training associated with signaling procedure 400. Therefore, online training may be performed, for example, by network node 504 closer to the data collection point.

[0073] Signaling procedure 500 may include OAM 302 transmitting one or more AI models to network node 504. In some implementations, the model may be a model trained via offline training (such as the offline training described with respect to signaling procedure 400). In other implementations, the model may be an untrained model. If the model's use case function is for CSI compression or beam management, then at 512, OAM 302 may transmit the model to base station 108, and base station 108 may transmit the model in a new RRC message (e.g., model delivery). ModelTransfer The model is forwarded within a container of messages. If the use case function is for location, OAM 302 can transmit the model to the LMF in core network 112. At 512, the LMF can then use NAS signaling to forward the model to UE 104, which can be incorporated into a new RRC message (e.g., ModelTransfer The message is in a container. ModelTransfer The message may also include the RRC ID mapped to the model.

[0074] If the data from UE 104 is intended for online training, signaling procedure 500 may include network node 504 configuring data collection by transmitting DC configuration to UE 104. At 516, [further details are needed]. DataMeasureConfig The message transmits the DC configuration and the corresponding RRC ID. According to some implementations, this is at position 516. dataMeasConfig The message can be similar to Figure 7 The RRC message 700 shows dataMeasConfig 712.

[0075] For online training, the DC configuration can be generated by network node 504, unlike the offline training discussed above which was generated by OAM 302. At 518, UE 104 can generate the DC configuration using a container that includes the RRC ID corresponding to the DC configuration and a dataset. MeasureReportDatasetOnline The message reports a dataset collected based on the DC configuration. This is similar to the dataset reporting described above regarding signaling procedure 400; however, in signaling procedure 500, dataset delivery can be terminated by network node 504 instead of OAM 302. According to some implementations, MeasureReportDatasetOnline The message can be similar to Figure 7 The RRC message 700 shows MeasureReportDatasetOnline 720.

[0076] The signaling procedure 500 can include online training 520 by the UE 104 or the network node 504. If the model is trained by the UE 104, the signaling procedure 500 can include sending the trained model in a message at 524. ModelTransfer ModelTransfer The message can be an RRC message that includes an RRC ID and a container that includes the model.

[0077] According to some embodiments, the message at 512 or 524 can include an indication of the model type. ModelTransfer The message can be similar to the RRC message 700 shown in Figure 7 ModelTransfer 724.

[0078] In some embodiments, the signaling procedure 500 can include the network node 504 sending the online trained model to the OAM 302 at 528. At 532, the OAM 302 can aggregate trained models received from multiple nodes of the network.

[0079] Figure 6 A signaling procedure 600 is illustrated that illustrates aspects of model inference and monitoring, according to some embodiments. The signaling procedure 600 can include signals between and operations performed by the UE 104 and the network node 504 and the OAM 302. According to some embodiments, the network node 504 can host / manage model inference and monitoring.

[0080] The signaling procedure 600 can include the network node 504 transmitting an RRC message, e.g., a model LCM configuration ( ModelLCMConfig ) message, to the UE 104 at 604. ModelLCMConfig The message can include an RRC ID (or model / functional ID), an LCM configuration, and a reporting configuration. The LCM configuration can provide an indication of how often the UE 104 is to monitor a particular model or DC configuration, and provide feedback. The LCM configuration can provide instructions regarding model inference, monitoring, activation, deactivation, fallback, or switching. The reporting configuration can provide instructions regarding how and when to feedback monitoring information. For example, the reporting configuration can configure periodic reporting by providing a reporting interval, or can configure event-triggered reporting. With respect to event-triggered reporting, the reporting configuration can configure or otherwise activate various thresholds that the UE 104 can use to detect whether an event exists that will trigger reporting. According to some embodiments, the message at 604 can include an indication of the model type. ModelLCMConfig The message can be similar to the RRC message 700 of Figure 7 ModelLCMConfig 728.

[0081] The signaling procedure 600 can also include the UE 104 performing LCM operations based on the LCM configuration at 608. In some embodiments, some or all of the LCM operations in 608 can be performed by the network node 504.​​​ The signaling procedure 600 can also include the UE 104 transmitting a report to the network node 504 at 612. The report can be based on the report configuration. In some embodiments, the LCM operation can be performed by the network node 504 and the report can not be needed.

[0082] ModelLCMConfig The RRC ID of the RRC message 700 can correspond to a model ID or a functionality ID. Whether a model ID or a functionality ID is used depends on whether a model-based LCM or a functionality-based LCM is applied. Model-based LCM means that the NW and the UE 104 exchange model information via their model IDs. The UE 104 applies a model according to the model ID provided by the NW. Functionality-based LCM means that the NW and the UE 104 exchange function information (e.g., CSI or beam management) about AI. The UE 104 can select which model to use in a model for one function (e.g., beam management).

[0083] The signaling procedure 600 can also include the network node 504 detecting a need for a model change based on the LCM operation 608 at 616. Upon detecting the need, the network node 504 can send an indication for a model change to the UE 104 at 620. The model change can be a switch or fallback from a first model to a second model, activation of a model, or deactivation of a model.

[0084] In some embodiments, AS level delta configuration can be applied. For example, ModelLCMConfig The message can include multiple RRC IDs / LCM configurations. In one example, the UE 104 can switch between models corresponding to the configured RRC IDs when certain events are detected. In this way, models can be added or removed based on predetermined events. In another example, the UE 104 can switch between models corresponding to the configured RRC IDs indicated by the base station 108. In this way, models can be added or removed according to RRC reconfiguration from the base station 108.

[0085] In some embodiments, a new SRB can be used for various RRC messages described herein. The RRC message can be Figure 7 any one of the RRC messages 700.

[0086] The new SRB can benefit from a higher priority than the legacy SRB4. If the new SRB has a lower priority, it can be prevented from being successfully transmitted due to potentially large payload size of the new SRB. However, the priority of the new SRB can be lower than SRB0 / 1 to prevent it from preventing transmission of data radio bearers (DRBs).

[0087] The priority of the new SRB can be determined based on one or more of the following options. In a first option, the priority can be configurable. For example, a configurable priority similar to DRBs can be provided by the base station 108 sending a configuration to the UE 104 via RRC signaling. In a second option, a separate radio link control (RLC) / logical channel (LCH) in the existing SRB4 can be used for the new SRB. A relatively high priority (e.g., highest priority) can be provided to the new RLC / LCH, while the existing RLC / LCH can be associated with a relatively low priority (e.g., lowest priority).

[0088] Some embodiments can utilize segmentation capabilities to allow transmission of large payload sizes that can be associated with one or more of the RRC messages 700. Existing downlink RRC message segmentation is AS segmentation, which has a limit of up to 5 segments (e.g., 45 KB) due to the limitation of the AS layer buffer size. However, the application layer buffer is much larger. Thus, according to some embodiments, one or more of the following options can be used for segmentation.

[0089] Figure 8 A segmented message 800 including an illustration of segmentation according to some embodiments. The segmented message 800 can include a first message 804 segmented according to a first option and a second message 808 segmented according to a second option.

[0090] In the first option, the application layer at the UE 104 or the OAM 302 can perform segmentation and include the segment ID as part of the application ID (e.g., application layer segment ID) along with the model / data set in the container 816. The message 804 can also include the RRC ID 812 as discussed elsewhere herein.

[0091] In the second option, the application layer at the UE 104 or the OAM 302 can perform segmentation and can provide the application layer segment ID to the AS layer (at the UE 104 or the base station 108). The AS layer can then generate the message 808 to include the application layer segment ID 824 in an IE of the payload of the message 808. The message can also include the RRC ID 820 and the container 828 with the model / data set.

[0092] Application segmentation can be utilized in conjunction with RRC segmentation.

[0093] In some embodiments, data compression can be applied to RRC messages transmitted by the new SRB to reduce the payload size. This can be particularly useful in cases where the RRC message is used to transmit large data sets. In some embodiments, a device can use the Request For Comments (RFC) 1951 DEFLATE Compressed Data FormatSpecification Consistent with the DEFLATE-based solution to reduce payload size in May 1996.

[0094] In some embodiments, the UE 104 can send UE assistance information to provide an indication of a temporary reduction in AI model capabilities of the UE. The AI model capabilities can correspond to reporting, monitoring, data collection, or any other operation associated with training or using an AI model. For example, the UE 104 can indicate whether it is experiencing (or no longer experiencing) an overheating, memory shortage, computational resource shortage, or low battery condition. Whether to take any action in response to receiving the UE assistance information can be up to the network. In some embodiments, the network can reduce computational operations to be performed by the UE 104 in the case that the UE has reduced AI model capabilities.

[0095] Figure 9 An operational flow / algorithmic structure 900 is provided, in accordance with some embodiments. The operational flow / algorithmic structure 900 can be executed by a UE, such as the UE 104, the UE 1200, or a component thereof (e.g., the processor 1204).

[0096] The operational flow / algorithmic structure 900 can include receiving a capability query, at 904. The capability query can be received from a base station and can request information about AI model capabilities of the UE.

[0097] The operational flow / algorithmic structure 900 can also include generating a capability response, at 908. The capability response can indicate AI models supported by the UE for use case functionality.

[0098] The operational flow / algorithmic structure 900 can also include sending the capability response to the base station, at 912.

[0099] Figure 10 An operational flow / algorithmic structure 1000 is provided, in accordance with some embodiments. The operational flow / algorithmic structure 1000 can be executed by an OAM, such as the OAM 302, or a network node 1300, or a component thereof (e.g., the processor 1304).

[0100] The operational flow / algorithmic structure 1000 can include identifying AI models supported by a UE, at 1004. The AI models supported by the UE can be identified based on a capability report received from the UE.

[0101] The operational flow / algorithmic structure 1000 can also include downloading AI models from a server, at 1008. The server can reside in an external data network.

[0102] The operational flow / algorithmic structure 1000 can further include sending the DC configuration to the UDM at 1012. The DC configuration can be associated with the AI model downloaded from the server. The UDM can determine whether the UE has agreed to share data based on the DC configuration. If yes, the approved DC configuration can be forwarded to the UE. If no, the rejected DC configuration can be provided to the OAM.

[0103] Figure 11 An operational flow / algorithmic structure 1100 according to some embodiments is provided. The operational flow / algorithmic structure 1100 can be executed by a network node, such as the OAM 302, the network node 504, or the network node 1300, or a component thereof (e.g., the processor 1304).

[0104] The operational flow / algorithmic structure 1100 can include receiving an AI model supported by a UE at 1104. The AI model supported by the UE can be identified based on a capability report received from the UE.

[0105] The operational flow / algorithmic structure 1100 can further include sending a DC configuration to the UE at 1108. The DC configuration can be associated with the AI model.

[0106] The operational flow / algorithmic structure 1100 can further include receiving a data set from the UE at 1112. The data set can be collected by the UE based on the DC configuration.

[0107] The operational flow / algorithmic structure 1100 can further include training the AI model based on the data set at 1116. In some embodiments, the training can be offline training performed by the OAM. In other embodiments, the training can be online training performed by a base station or an LMF.

[0108] Figure 12 An example of an example UE 1200 according to some embodiments is illustrated. The UE 1200 can be any mobile or non-mobile computing device, such as, for example, a mobile phone, a computer, a tablet, an industrial wireless sensor (e.g., a microphone, a carbon dioxide sensor, a pressure sensor, a humidity sensor, a thermometer, a motion sensor, an accelerometer, a laser scanner, a fluid level sensor, an inventory sensor, a voltage / current meter, or an actuator), a video surveillance / monitoring device (e.g., a camera), a wearable device (e.g., a smartwatch), or an Internet of Things (IoT) device.

[0109] The UE 1200 can include a processor 1204, RF interface circuitry 1208, memory / storage 1212, user interface 1216, sensors 1220, drive circuitry 1222, power management integrated circuit (PMIC) 1224, antenna structure 1226, and battery 1228. The components of the UE 1200 can be implemented as integrated circuits (ICs), portions thereof, discrete electronic devices, or other modules, logic, hardware, software, firmware, or a combination thereof. Figure 12 The block diagram of FIG. 12 is intended to show a high-level view of some of the components of the UE 1200. However, some of the components shown can be omitted in some embodiments, additional components can be present, and different arrangements of the components shown can occur in other embodiments.

[0110] The components of the UE 1200 can be coupled through one or more interconnects 1232, which can represent any type of interface, input / output, bus (local, system, or expansion), transmission line, trace, optical connection, etc. that allows various circuit components (on common or different chips or chipsets) to interact with one another.

[0111] The processor 1204 can include processor circuitry such as, for example, baseband processor circuitry (BB) 1204A, central processor unit circuitry (CPU) 1204B, and graphics processor unit circuitry (GPU) 1204C. The processor 1204 can include any type of circuit or processor circuitry that executes or otherwise operates computer-executable instructions, such as program code, software modules, or functional processes from the memory / storage 1212, to cause the UE 1200 to perform operations of the UE with respect to AI models as described herein.

[0112] In some embodiments, the baseband processor circuitry 1204A can access a communication protocol stack 1236 in the memory / storage 1212 to communicate over a 3GPP-compatible network. Generally, the baseband processor circuitry 1204A can access the communication protocol stack to perform user plane functions at the PHY layer, MAC layer, RLC layer, PDCP layer, SDAP layer, and PDU layer; and control plane functions at the PHY layer, MAC layer, RLC layer, PDCP layer, RRC layer, and non-access stratum layer. In some embodiments, PHY layer operations can additionally / alternatively be performed by components of the RF interface circuitry 1208.

[0113] The baseband processor circuitry 1204A can generate or process baseband signals or waveforms that carry information that is transmitted in a 3GPP-compatible network. In some embodiments, waveforms for NR can be based on cyclic prefix OFDM (CP-OFDM) in the uplink or downlink, and discrete Fourier transform spread OFDM (DFT-S-OFDM) in the uplink.

[0114] The memory / storage 1212 can include one or more non-transitory computer-readable media including instructions (e.g., communication protocol stack 1236) executable by one or more of the processors 1204 to cause the UE 1200 to perform various operations described herein. Memory / storage 1212 includes any type of volatile or nonvolatile memory needed to store instructions and data accessible to processors 1204, including both local and remote memory applications. In some embodiments, some of memory / storage 1212 can reside on the processors 1204 themselves (e.g., Ll cache and L2 cache) while other memory / storage 1212 is external to the processors 1204. Each processor 1204 can include multiple cores (e.g., a dual-core processor or a quad-core processor) and / or a shared cache memory. The system 1200 can further include an external memory / storage 1212, such as a hard disk drive and / or a removable storage drive. The memory / storage 1212 can also include solid-state drives, optically readable storage media, and the like. The instructions of the communication protocol stack 1236 can be part of software that when executed by the processor(s) 1204, enable the processor(s) 1204 to perform various operations described herein.

[0115] The RF interface circuitry 1208 can include transceiver circuitry and radio frequency front module (RFEM) that allow UE 1200 to communicate with other devices over a radio access network. The RF interface circuitry 1208 can include various elements arranged in transmit or receive paths. These elements can include, for example, switches, mixers, amplifiers, filters, synthesizer circuitry, control circuitry, etc.

[0116] In the receive path, the RFEM can receive a radiated signal from the air interface via the antenna structure 1226 and continue to filter and amplify (with a low noise amplifier) the signal. The signal can be provided to a receiver of the transceiver that down-converts the RF signal to a baseband signal that is provided to the baseband processor of the processor 1204.

[0117] In the transmit path, a transmitter of the transceiver up-converts baseband signals received from the baseband processor to RF signals that are provided to the RFEM. The RFEM can amplify the signal through a power amplifier before the signal is radiated across the air interface via the antenna structure 1226.

[0118] In various embodiments, the RF interface circuitry 1208 can be configured to transmit / receive signals in a manner compatible with NR access technology.

[0119] The antenna structure 1226 can include antenna elements to convert electrical signals into radio waves to travel through the air and to convert received radio waves into electrical signals. The antenna elements can be arranged in one or more antenna panels. The antenna structure 1226 can have an antenna panel that is omnidirectional, directional, or a combination thereof to enable beamforming and multiple input multiple output communication. The antenna structure 1226 can include microstrip antennas, printed antennas fabricated on a surface of one or more printed circuit boards, patch antennas, phased array antennas, etc. The antenna structure 1226 can have one or more panels designed for a particular frequency band including bands in FR1 or FR2.

[0120] The user interface 1216 includes various input / output (I / O) devices designed to enable user interaction with the UE 1200. The user interface 1216 includes input device circuitry and output device circuitry. Input device circuitry includes any physical or virtual means for accepting an input including, inter alia, one or more physical or virtual buttons (e.g., a reset button), a physical keyboard, a keypad, a mouse, a touchpad, a touchscreen, a microphone, a scanner, or a headset, etc. The output device circuitry includes any physical or virtual means for showing information or otherwise conveying information, such as sensor readings, actuator positions, or other similar information. The output device circuitry can include any number or combination of audio or visual display including, inter alia, one or more simple visual outputs / indicators (e.g., binary status indicators (such as light emitting diodes (LEDs)) and multi-character visual outputs), or more complex outputs such as display devices or touchscreens (e.g., liquid crystal displays (LCD), LED displays, quantum dot displays, projectors, etc.), the output of which is generated under operation of the UE 1200.

[0121] The sensors 1220 can include devices that detect events or changes in its environment and transmits the information about the detected events (sensor data) to some other devices, modules, subsystems, etc. Examples of such sensors include, inter alia, an inertial measurement unit comprising an accelerometer, a gyroscope, or a magnetometer; a microelectromechanical system or nanoelectromechanical system comprising a 3-axis accelerometer, a 3-axis gyroscope, or a magnetometer; a liquid level sensor; a flow rate sensor; a temperature sensor (e.g., a thermistor); a pressure sensor; a barometric pressure sensor; a gravimeter; an altimeter; an image capture device (e.g., a camera or a lensless aperture); a light detection and ranging sensor; a proximity sensor (e.g., an infrared radiation detector, etc.); a depth sensor; an ambient light sensor; an ultrasonic transceiver; a microphone or other like audio capture device; etc.

[0122] The drive circuitry 1222 can include software and hardware elements that operate to control particular devices embedded in, or attached to, or otherwise interfaced with the UE 1200. The drive circuitry 1222 can include individual drivers to allow other components to interact with or control various input / output (I / O) devices that can be present within, or connected to, the UE 1200. For example, the drive circuitry 1222 can include a display driver to control and allow access to a display device, a touchscreen driver to control and allow access to a touchscreen interface, a sensor driver to retrieve sensor readings from, and control and allow access to, the sensors 1220, a driver to retrieve actuator positions from, or control and allow access to, electromechanical components, a camera driver to control and allow access to an embedded image capture device, an audio driver to control and allow access to one or more audio devices.

[0123] The PMIC 1224 can manage power for the various components of the UE 1200. In particular, with respect to the processor 1204, the PMIC 1224 can control power source selection, voltage scaling, battery charging, or DC-to-DC conversion.

[0124] In some embodiments, the PMIC 1224 can control, or otherwise be part of, various power saving mechanisms of the UE 1200. For example, if the platform UE is in an RRC_Connected state, in which it is still connected to the RAN node as it expects to receive traffic shortly, then the platform UE can enter a state known as Discontinuous Reception (DRX) in which it powers down for intervals. During this state, the UE 1200 can power down for intervals of time and thus save power. If there is no data traffic activity for some period of time, the UE 1200 can transition off to an RRC_Idle state, where it disconnects from the network and does not perform operations such as channel quality feedback, handover, etc. The UE 1200 goes into a very low power state and it performs paging where again it periodically wakes up to listen to the network and then powers down. The UE 1200 can not receive data in this state; in order to receive data, it must transition back to RRC_Connected state. An additional power saving mode can be a deep sleep state where the device powers down for an extended period of time, e.g., 30 minutes, 60 minutes, 90 minutes, etc. During this time, it can not connect to the network at all and it can not receive data. Any data incoming for the device while in the deep sleep state is buffered and the device can be notified of the data once it connects back to the network.

[0125] The battery 1228 can power the UE 1200, although in some examples the UE 1200 can be installed or deployed in a fixed location, and can have an electrical power source coupled to an electrical grid. The battery 1228 can be a lithium ion battery, or a metal-air battery, such as a zinc-air battery, an aluminum-air battery, a lithium-air battery, and so forth. In some implementations, such as in vehicle-based applications, the battery 1228 can be a typical lead-acid automotive battery.

[0126] Figure 13 An example network node 1300 is illustrated in accordance with some embodiments. The network node 1300 can be a base station, an LMF, an OAM, an AMF, or a UDM as described elsewhere herein. The network node 1300 can include a processor 1304, RF interface circuitry 1308, core network (CN) interface circuitry 1312, memory / storage circuitry 1316, and antenna structure 1326. When the network node 1300 is an AMF, the RF interface circuitry 1308 and the antenna structure 1326 can not be included.

[0127] The components of the network node 1300 can be coupled with various other components by one or more interconnects 1328.

[0128] The processor 1304, RF interface circuit 1308, memory / storage circuit 1316 (including communication protocol stack 1310), antenna structure 1326, and interconnect 1328 can be similar to those described with respect to Figure 12 Like-named components have been described with like-naming convention.

[0129] The processor 1304 can execute instructions for the network node 1300 to perform operations such as those described with respect to AI model operations of a base station, LMF, OAM, AMF, or UDM described elsewhere herein.

[0130] The CN interface circuit 1312 can provide connectivity to a core network (e.g., a 5thGeneration Core Network (5GC) using a 5GC-compatible network interface protocol such as carrier Ethernet protocol or some other suitable protocol). Network connectivity to / from the network node 1300 can be provided via optical or wireless backhaul. The CN interface circuit 1312 can include one or more dedicated processors or FPGAs to communicate using one or more of the aforementioned protocols. In some implementations, the CN interface circuit 1312 can include multiple controllers to provide connectivity to other networks using the same or different protocols.

[0131] It is well understood that the use of personally identifiable information should follow privacy policies and practices that are generally recognized as meeting or exceeding industry or governmental requirements for maintaining the privacy of users. In particular, personally identifiable information data should be managed and handled in a way to minimize risk of unauthorized or unintended access or use of the data and to minimize the perpetuation of such unauthorized or unintended access or use. Further, users should be informed of the nature of the data being collected and the uses anticipated.

[0132] For one or more embodiments, at least one of the components shown in one or more of the preceding figures can be configured to perform one or more operations, techniques, processes, or methods set forth in the Example section below. For example, the baseband circuitry described above in connection with one or more of the preceding figures can be configured to operate according to one or more of the examples described below. For another example, circuitry associated with a UE, base station, or network element described above in connection with one or more of the preceding figures can be configured to operate according to one or more of the embodiments set forth in the Example section below.

[0133] Embodiments In the following sections, additional example embodiments are provided.

[0134] Example 1 includes a method to be implemented by a user equipment (UE), the method comprising: receiving, from a base station, a capability query; generating, based on the capability query, a capability response to indicate an artificial intelligence (AI) model supported by the UE for a use case function; and transmitting, to the base station, the capability response.

[0135] Example 2 includes the method of example 1 or some other example herein, wherein the capability response is to indicate one or more AI models supported by the UE per use case function; or to indicate one or more use case functions supported by the UE per AI model.

[0136] Example 3 includes the method of example 1 or some other example herein, the method further comprising: generating a radio resource configuration (RRC) message including a container with an identifier associated with the AI model or the use case function; and transmitting, to the base station, the RRC message, wherein the identifier is to be forwarded to an operations, administration, and maintenance (OAM) node for the OAM node to train the AI model for the use case function.

[0137] Example 4 includes the method of example 3 or some other example herein, wherein the identifier is a first identifier associated with the AI model, and the RRC message further includes a second identifier associated with the use case function.

[0138] Example 5 includes the method of example 3 or some other example herein, the method further comprising: determining, based on a configuration from a manufacturer of the UE or based on signaling from a public land mobile network (PLMN), that the identifier is associated with the AI model, wherein the identifier is a globally unique identifier.

[0139] Example 6 includes the method of example 3 or some other example herein, wherein the AI model is a first AI model, a second AI model is associated with the use case function, the identifier is a first identifier associated with the first AI model, and the RRC message further includes a second identifier associated with the second AI model.

[0140] Example 7 includes the method of example 3 or some other example herein, wherein the identifier is an identifier of the AI model, the RRC message is a first RRC message, the container is a first container, and the method further comprises: receiving, from the base station, a second RRC message including a second container with the identifier of the AI model to confirm that the OAM node supports the AI model.

[0141] Example 8 includes the method of example 1 or some other example herein, further comprising: transmitting assistance information to the base station to provide an indication of reduced AI model capabilities of the UE.

[0142] Example 9 includes the method of example 8 or some other example herein, wherein the indication corresponds to whether the UE is experiencing an overheat, memory shortage, compute resource shortage, or low battery condition.

[0143] Example 10 includes the method of example 8 or some other example herein, further comprising: transmitting the assistance information in a RRC message or in uplink control information.

[0144] Example 11 includes a method to be implemented by an operations, administration, and maintenance (OAM) node, the method comprising: identifying an artificial intelligence (AI) model supported by a user equipment (UE); downloading the AI model from a server; and transmitting, to a unified data management (UDM) function of a core network, an identifier of a data collection configuration associated with the AI model to determine whether the UE authorizes sharing of a data set based on the data collection configuration.

[0145] Example 12 includes the method of example 11 or some other example herein, further comprising: confirming that the AI model is supported by the OAM node; and communicating, to the UE, a confirmed model identifier (ID) associated with the AI model.

[0146] Example 13 includes the method of example 11 or some other example herein, further comprising: a response from the UDM to indicate that the UE does not authorize sharing of the data set based on the data collection configuration.

[0147] Example 14 includes the method of example 11 or some other example herein, further comprising: generating a container to include one or more identifiers respectively associated with one or more data collection configurations, wherein the one or more identifiers include the identifier and the one or more data collection configurations include the data collection configuration, wherein the transmitting the identifier comprises transmitting the container.

[0148] Example 15 includes the method of example 11 or some other example herein, further comprising: receiving, from the UE via a base station, a data set corresponding to the data collection configuration, wherein the data set is transmitted to the base station in a container of a radio resource control (RRC) message; and training the AI model based on the data set to generate a trained AI model.

[0149] Example 16 includes the method of example 15 or some other example herein, wherein the AI model is associated with channel state information (CSI) compression or beam management, and the method further comprises transmitting, to the UE via a base station, the trained AI model.

[0150] Example 17 includes the method of example 15 or some other example herein, wherein the AI model is associated with UE positioning, and the method further comprises transmitting, to the UE via a location management function (LMF), the trained AI model.

[0151] Example 18 includes a method to be implemented by a unified data management (UDM) function, the method comprising: receiving, from an operations, administration, and maintenance (OAM) node, an identifier of a data collection configuration associated with an artificial intelligence (AI) model; determining whether a user equipment (UE) authorizes sharing of a data set associated with the data collection configuration; and sending a message to the OAM or to an access and mobility management function (AMF) based on the determination of whether the UE authorizes sharing of the data set.

[0152] Example 19 includes the method of example 18 or some other example herein, wherein determining whether the UE authorizes sharing of the data set comprises determining that the UE authorizes sharing of the data set, and the method further comprises sending the message to an AMF to activate the data collection configuration.

[0153] Example 20 includes a method to be implemented in a base station, the method comprising: receiving an AI model configuration information element (IE) having an identifier of a data collection configuration associated with an artificial intelligence (AI) model supported by a user equipment (UE); and sending, to the UE, a radio resource control (RRC) message having the AI model configuration IE to activate the data collection configuration at the UE.

[0154] Example 21 includes the method of example 20 or some other example herein, wherein the RRC message further comprises an RRC identifier associated with a globally unique identifier of the AI model.

[0155] Example 22 includes the method of example 20 or some other example herein, the method further comprising: receiving the AI model configuration IE over an N2 interface, wherein the RRC message is a data measurement configuration message.

[0156] Example 23 includes the method of example 20 or some other example herein, wherein the RRC message further indicates a type of data set, a reporting configuration, or whether the data collection configuration is visible to a radio access network (RAN).

[0157] Example 24 includes the method according to Example 20 or some other embodiment herein, the method further comprising: receiving a Measurement Report Dataset RRC message from the UE, the Measurement Report Dataset RRC message including a container having a dataset collected by the UE based on the data collection configuration.

[0158] Example 24 includes the method described according to Example 24 or some other embodiment herein, the method further comprising: forwarding the dataset having an identifier of the AI ​​model to the Operation, Administration and Maintenance (OAM) function.

[0159] Example 26 includes a method for operating a network node, the method comprising: receiving an artificial intelligence (AI) model from an operation, management, and maintenance (OAM) node; generating a data measurement configuration (DMAC) of a container having a Radio Resource Control (RRC) identifier and a data collection (DC) configuration associated with the AI ​​model. DataMeasureConfig ) message; send the aforementioned message to User Equipment (UE) DataMeasureConfig The message includes: receiving a dataset collected from the UE based on the DC configuration; and performing online training of the AI ​​model based on the dataset.

[0160] Example 27 includes the method according to Example 26 or some other embodiment herein, wherein the AI ​​model is a first AI model, and the method further includes: generating a first model transfer having a Radio Resource Control (RRC) identifier and a container having the first AI model or a second AI model. ModelTransfer ) message; and send the message to the UE ModelTransfer Message; and receiving from the UE a second AI model having the RRC identifier and training corresponding to the first AI model or the second AI model. ModelTransfer information.

[0161] Example 28 includes the method according to Example 26 or some other embodiment herein, wherein the network node is a base station or location management function.

[0162] Example 29 includes a method to be implemented by a user equipment, the method comprising: receiving model lifecycle management configuration from a network node (…). ModelLCMConfig ) message, the model lifecycle management configuration ( ModelLCMConfig The message includes a radio resource identifier associated with an artificial intelligence (AI) model and a lifecycle management (LCM) configuration; and performs LCM operations based on the LCM configuration.

[0163] Example 30 includes the method of Example 29 or some other Example herein, wherein the LCM configuration is configured for configuring an LCM operation, the LCM operation comprising: model inference, monitoring, activation, deactivation, fallback, or handover, and the method further comprises: performing the LCM operation based on the LCM configuration.

[0164] Example 31 includes the method of Example 30 or some other Example herein, the method further comprising: sending a report to the network node based on the performance of the LCM operation.

[0165] Example 32 includes the method of Example 29 or some other Example herein, the method further comprising: receiving an indication from the network node to perform a model change, wherein the model change is a handover, activation, deactivation, or fallback; and performing the model change.

[0166] Example 33 includes a method to be implemented by a network node, the method comprising: generating a radio resource control (RRC) message to include a container, the container having an artificial intelligence (AI) model identifier (ID), a use case function ID corresponding to a use case function of an AI model, a data collection (DC) configuration of the AI model, a dataset associated with the AI model, an AI model, or a lifecycle management (LCM) configuration associated with the AI model; and sending the RRC message via a signaling radio bearer (SRB).

[0167] Example 34 includes the method of Example 33 or some other Example herein, wherein the SRB comprises a priority configured to be higher than an SRB4 priority and lower than an SRB0 / 1 priority.

[0168] Example 35 includes the method of Example 34 or some other Example herein, wherein the priority is configured by a base station using RRC signaling.

[0169] Example 36 includes the method of Example 33 or some other Example herein, wherein the SRB is SRB4, and the method further comprises: sending the RRC message in a first radio link layer (RLC) or logical channel (LCH) of SRB4, the first radio link layer (RLC) or logical channel (LCH) of the SRB4 being associated with a first priority that is higher than a second priority of a second RLC or LCH of the SRB4.

[0170] Example 37 includes the method of Example 33 or some other Example herein, wherein the RRC message includes an application layer segment identifier inside the container or outside the container.

[0171] Example 38 includes the method of example 33 or some other example herein, further comprising applying data compression to data within the container.

[0172] Another embodiment can include an apparatus comprising means for performing one or more elements of a method described in or related to any of examples 1-38, or any other method or process described herein.

[0173] Another embodiment can include one or more non-transitory computer-readable media comprising instructions to cause an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform one or more elements of a method described in or related to any of examples 1-38, or any other method or process described herein.

[0174] Another embodiment can include a device comprising logic, modules, or circuitry to perform one or more elements of a method described in or related to any of examples 1-38, or any other method or process described herein.

[0175] Another embodiment can include a method, technique, or process as described in or related to any of examples 1-38, or portions or aspects thereof.

[0176] Another embodiment can include an apparatus comprising: one or more processors; and one or more computer-readable media comprising instructions to cause the one or more processors to perform the method, technique, or process of any of examples 1-38, or portions thereof, when executed by the one or more processors.

[0177] Another embodiment can include a signal as described in or related to any of examples 1-38, or portions or aspects thereof.

[0178] Another embodiment can include a datagram, information element, packet, frame, segment, PDU, or message as described in or related to any of examples 1-38, or portions or aspects thereof, or otherwise described in the present disclosure.

[0179] Another embodiment can include a signal encoded with data as described in or related to any of examples 1-38, or portions or aspects thereof, or otherwise described in the present disclosure.

[0180] Another embodiment can include a signal encoded with a datagram, IE, packet, frame, segment, PDU, or message as described in or related to any of examples 1-38, or portions or aspects thereof, or otherwise described in the present disclosure.

[0181] Another embodiment can include an electromagnetic signal carrying computer-readable instructions, wherein execution of the computer-readable instructions by one or more processors will cause the one or more processors to perform the method, techniques, or process as described in or related to any of embodiments 1-38, or portions thereof.

[0182] Another embodiment can include a computer program comprising instructions, wherein execution of the program by a processing element will cause the processing element to perform a method, techniques, or process, as described in or related to any of embodiments 1-38, or portions thereof.

[0183] Another embodiment can include a signal in a wireless network as shown and described herein.

[0184] Another embodiment can include a method of communicating in a wireless network as shown and described herein.

[0185] Another embodiment can include a system for providing wireless communication as shown and described herein.

[0186] Another embodiment can include an apparatus for providing wireless communication as shown and described herein.

[0187] Any of the embodiments described above can be combined with any other embodiment (or combination of embodiments), unless explicitly stated otherwise. The foregoing description of one or more implementations provides exemplification and description to enable a thorough understanding of the disclosure. However, various implementations can be practiced without resorting to the details shown. Aspects of the above described implementations can be modified or combined in various ways without departing from the scope of the disclosure.

[0188] While the above implementations have been described with some detail, once armed with the above disclosure, many variations and modifications will become apparent to those skilled in the art. It is intended that the claims be construed to include all such variations and modifications.

Claims

1. One or more computer-readable media having instructions that, when executed by one or more processors, cause a user equipment (UE) to: receive, from a base station, a capability query; generate, based on the capability query, a capability response to indicate an artificial intelligence (AI) model supported by the UE for a use case function; and send, to the base station, the capability response.

2. The one or more computer-readable media of claim 1, wherein the capability response is to indicate one or more AI models supported by the UE per use case function; or to indicate one or more use case functions supported by the UE per AI model.

3. The one or more computer-readable media of claim 1, wherein the instructions, when executed, further cause the UE to: generate a radio resource configuration (RRC) message that includes a container with an identifier associated with the AI model or the use case function; and send, to the base station, the RRC message, wherein the identifier is to be forwarded to an operations, administration, and maintenance (OAM) node for the OAM node to train the AI model for the use case function.

4. The one or more computer-readable media of claim 3, wherein the identifier is a first identifier associated with the AI model, and the RRC message further includes a second identifier associated with the use case function.

5. The one or more computer-readable media of claim 3, wherein the instructions, when executed, further cause the UE to: determine, based on a configuration from a manufacturer of the UE or based on signaling from a public land mobile network (PLMN), that the identifier is associated with the AI model, wherein the identifier is a globally unique identifier.

6. The one or more computer-readable media of claim 3, wherein the AI model is a first AI model, a second AI model is associated with the use case function, the identifier is a first identifier associated with the first AI model, and the RRC message further includes a second identifier associated with the second AI model.

7. The one or more computer-readable media of claim 3, wherein the identifier is an identifier of the AI model, the RRC message is a first RRC message, the container is a first container, and the instructions, when executed, further cause the UE to: receive, from the base station, a second RRC message that includes a second container with the identifier of the AI model to confirm that the OAM node supports the AI model.

8. The one or more computer-readable media of claim 1, wherein the instructions, when executed, further cause the UE to: send, to the base station, assistance information to provide an indication of a reduced AI model capability of the UE. ​ ​ ​ ​ ​ 9. The one or more computer-readable media of claim 8, wherein the indication corresponds to whether the UE is experiencing an overheat, memory shortage, compute resource shortage, or low battery condition.

10. The one or more computer-readable media of claim 8, wherein the instructions, when executed, further cause the UE to: transmit the assistance information in a radio resource control (RRC) message or in uplink control information.

11. A method to be implemented by an operations, administration, and maintenance (OAM) node, the method comprising: identifying an artificial intelligence (AI) model supported by a user equipment (UE); downloading the AI model from a server; and transmitting, to a unified data management (UDM) function of a core network, an identifier of a data collection configuration associated with the AI model to determine whether the UE authorizes sharing of a data set based on the data collection configuration.

12. The method of claim 11, the method further comprising: confirming that the AI model is supported by the OAM node; and communicating, to the UE, a confirmed model identifier (ID) associated with the AI model.

13. The method of claim 11, the method further comprising: a response from the UDM to indicate that the UE does not authorize sharing of the data set based on the data collection configuration.

14. The method of claim 11, the method further comprising: generating a container to include one or more identifiers associated with one or more data collection configurations, respectively, wherein the one or more identifiers include the identifier and the one or more data collection configurations include the data collection configuration, wherein the transmitting the identifier includes transmitting the container.

15. The method of claim 11, the method further comprising: receiving, from the UE via a base station, a data set corresponding to the data collection configuration, wherein the data set is transmitted to the base station in a container of a radio resource control (RRC) message; and training the AI model based on the data set to generate a trained AI model.

16. The method of claim 15, wherein the AI model is associated with channel state information (CSI) compression or beam management, and the method further comprising: communicating, to the UE via a base station, the trained AI model.

17. The method of claim 15, wherein the AI model is associated with UE positioning, and the method further comprising: communicating, to the UE via a location management function (LMF), the trained AI model.

18. An apparatus having circuitry for causing a unified data management (UDM) function to: receive, from an operations, administration, and maintenance (OAM) node, an identifier of a data collection configuration associated with an artificial intelligence (AI) model; determine whether a user equipment (UE) authorizes sharing of a data set associated with the data collection configuration; and sending a message to the OAM or to an access and mobility management function (AMF) based on the determining whether the UE authorizes sharing of the data set.

19. The apparatus of claim 18, wherein the determining whether the UE authorizes sharing of the data set comprises determining that the UE authorizes sharing of the data set, and the circuitry further causes the UDM function to: send the message to an AMF to activate the data collection configuration.

20. A method to be implemented in a base station, the method comprising: receiving an artificial intelligence (AI) model configuration information element (IE) having an identifier of a data collection configuration associated with an AI model supported by a user equipment (UE); and sending a radio resource control (RRC) message having the AI model configuration IE to the UE to activate the data collection configuration at the UE.

21. The method of claim 20, wherein the RRC message further includes an RRC identifier associated with a globally unique identifier of the AI model.

22. The method of claim 20, further comprising: receiving the AI model configuration IE over an N2 interface, wherein the RRC message is a data measurement configuration message.

23. A network node, the network node comprising interface circuitry; and a processing circuit coupled with the interface circuit, the processing circuit to: receive an artificial intelligence (AI) model from an operations, administration, and maintenance (OAM) node via the interface circuit; generating a data measurement configuration (D DataMeasureConfig ) message with a radio resource control (RRC) identifier and a container with a data collection (DC) configuration associated with the AI model; The interface circuit transmits the data to the user equipment (UE). DataMeasureConfig information; receive a data set collected based on the DC configuration from the UE via the interface circuit; and perform online training of the AI model based on the data set.

24. A method to be implemented by a user equipment (UE), the method comprising: receive a model lifecycle management configuration ( ModelLCMConfig ) message from a network node, the model lifecycle management configuration ( ModelLCMConfig ) message including a radio resource identifier associated with an artificial intelligence (AI) model and a lifecycle management (LCM) configuration; and perform a lifecycle management (LCM) operation based on the LCM configuration.

25. A method to be implemented by a network node, the method comprising: generating a radio resource control (RRC) message to include a container having an artificial intelligence (AI) model identifier (ID), a use case function ID corresponding to a use case function of an AI model, a data collection (DC) configuration of an AI model, a data set associated with an AI model, an AI model, or a lifecycle management (LCM) configuration associated with an AI model; and sending the RRC message via a signaling radio bearer (SRB).

Images