Data set sharing for AI / ML

Abstract

Apparatus and methods for data set sharing for AI / ML are provided. The first communication device comprises: at least one antenna; and a processor configured to: transmit a dataset notification including meta-information for a dataset of an artificial intelligence (AI) model to a second communication device; receiving, from the second communication device, a dataset request requesting to obtain the dataset; and transmitting a dataset report including the dataset to the second communication device.

Description

Technical Field

[0001] The present application generally relates to wireless communication systems, including apparatus and methods for dataset sharing for AI / ML in wireless communication systems. Background Art

[0002] Wireless mobile communication technologies use various standards and protocols to transmit data between base stations and wireless devices. Wireless communication system standards and protocols may include, for example, the 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) (e.g., 4G), 3GPP New Radio (NR) (e.g., 5G), and the IEEE 802.11 standard for wireless local area networks (WLANs), commonly referred to within industry organizations as WLANs. ).

[0003] As envisioned by 3GPP, different wireless communication system standards and protocols may use various radio access networks (RANs) to facilitate communication between base stations of the RAN (which may also be sometimes referred to as RAN nodes, network nodes, or simply nodes) and wireless devices called user equipment (UEs). 3GPP RANs may include, for example, Global System for Mobile Communications (GSM), Enhanced Data Rates for GSM Evolution (EDGE) RAN (GERAN), Universal Terrestrial Radio Access Network (UTRAN), Evolved Universal Terrestrial Radio Access Network (E-UTRAN), and / or Next Generation Radio Access Network (NG-RAN).

[0004] Each RAN can use one or more radio access technologies (RATs) to perform communications between base stations and UEs. For example, GERAN implements GSM and / or EDGE RATs, UTRAN implements Universal Mobile Telecommunications System (UMTS) RATs or other 3GPP RATs, E-UTRAN implements LTE RATs (sometimes referred to as LTE), and NG-RAN implements NR RATs (sometimes referred to herein as 5G RATs, 5G NR RATs, or simply NR). In some deployments, E-UTRAN may also implement NR RATs. In some deployments, NG-RAN may also implement LTE RATs.

[0005] The base stations used by the RAN may correspond to the RAN. An example of an E-UTRAN base station is an Evolved Universal Terrestrial Radio Access Network (E-UTRAN) Node B (often also denoted as an evolved Node B, enhanced Node B, eNodeB, or eNB). An example of an NG-RAN base station is a Next Generation Node B (sometimes also referred to as a gNodeB or gNB).

[0006] The RAN provides communication services with external entities through its connection to the Core Network (CN). For example, E-UTRAN may utilize the Evolved Packet Core (EPC), while NG-RAN may utilize the 5G Core Network (5GC).

[0007] Artificial intelligence (AI) is the simulation of human intelligence processes by machines, typically computer systems. Machine learning (ML), a subset of AI, creates algorithms and statistical models to perform specific tasks without explicit instructions, relying instead on patterns and reasoning. ML algorithms build mathematical models based on sample data, called training data, to make predictions or decisions without being specifically programmed for the task. When compared to today's fragile, manually designed systems, learned signal processing algorithms can support next-generation wireless systems with significantly reduced power consumption and improved density, throughput, and accuracy. Summary of the Invention

[0008] The present disclosure provides an apparatus and method for sharing datasets for AI / ML between different entities in a wireless communication system.

[0009] Embodiments disclosed herein include a first communication device comprising: at least one antenna; and a processor configured to: send a dataset notification including metadata of a dataset for an artificial intelligence (AI) model to a second communication device; receive a dataset request from the second communication device requesting the dataset; and send a dataset report including the dataset to the second communication device.

[0010] Embodiments disclosed herein include a second communication device comprising: at least one antenna; and a processor configured to: receive a dataset notification including metadata of a dataset for an artificial intelligence (AI) model from a first communication device; send a dataset request to the first communication device requesting the dataset; and receive a dataset report including the dataset from the first communication device.

[0011] Embodiments disclosed herein include a method performed by a first communication device, the method comprising: sending a dataset notification including metadata of a dataset for an artificial intelligence (AI) model to a second communication device; receiving a dataset request from the second communication device requesting to obtain the dataset; and sending a dataset report including the dataset to the second communication device.

[0012] Embodiments disclosed herein include a method performed by a second communication device, the method comprising: receiving a dataset notification including metadata of a dataset for an artificial intelligence (AI) model from a first communication device; sending a dataset request to the first communication device requesting to obtain the dataset; and receiving a dataset report including the dataset from the first communication device.

[0013] The embodiments disclosed herein include a non-transitory computer-readable storage medium having instructions stored thereon that, when executed by a processor of a first communication device, cause the processor to: send a dataset notification including metadata of a dataset for an artificial intelligence (AI) model to a second communication device; receive a dataset request from the second communication device requesting the dataset; and send a dataset report including the dataset to the second communication device.

[0014] The embodiments disclosed herein include a non-transitory computer-readable storage medium having instructions stored thereon that, when executed by a processor of a second communication device, cause the processor to: receive a dataset notification including metadata of a dataset for an artificial intelligence (AI) model from a first communication device; send a dataset request to the first communication device requesting the dataset; and receive a dataset report including the dataset from the first communication device. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] To easily identify the discussion of any particular element or action, the most significant digit(s) in a reference number refers to the drawing number that first introduces that element.

[0016] Figure 1 An example architecture of a wireless communication system according to the embodiments disclosed herein is illustrated.

[0017] Figure 2 A system for performing signaling between a wireless device and a network device according to embodiments disclosed herein is illustrated.

[0018] Figure 3 An example functional framework for AI / ML in a wireless communication system according to the embodiments disclosed herein is illustrated.

[0019] Figure 4 An example communication procedure between a UE and a gNB for preparing for data set sharing according to the embodiments disclosed herein is illustrated.

[0020] Figure 5 Another example communication process between a UE, a gNB, and a unified data management (UDM) for preparing data set sharing according to the embodiments disclosed herein is illustrated.

[0021] Figure 6 An example communication procedure for data set sharing from a UE to a gNB according to the embodiments disclosed herein is illustrated.

[0022] Figure 7 An example communication process for data set sharing from a gNB to a UE according to the embodiments disclosed herein is illustrated.

[0023] Figure 8 An example method performed by a first communications device according to embodiments disclosed herein is illustrated.

[0024] Figure 9 An example method performed by a second communications device according to embodiments disclosed herein is illustrated. DETAILED DESCRIPTION

[0025] Various embodiments are described with respect to a UE. However, reference to a UE is provided for illustrative purposes only. The example embodiments may be used with any wireless device that can establish a connection to a network and is configured with hardware, software, and / or firmware for exchanging information and data with the network. Therefore, a UE as described herein is intended to represent any suitable network-connected device.

[0026] Various embodiments are described in terms of gNBs. However, reference to gNBs is provided for illustrative purposes only. The example embodiments may be used with any network device in a network and configured with hardware, software, and / or firmware to implement any functionality of that network. Therefore, gNBs as described herein are intended to represent any suitable network device.

[0027] Figure 1 An example architecture of a wireless communication system 100 according to the embodiments disclosed herein is illustrated. The following description is provided for an example wireless communication system 100 operating in conjunction with the LTE system standard and / or the 5G or NR system standard provided by the 3GPP technical specifications.

[0028] like Figure 1 As shown, wireless communication system 100 includes UE 102 and UE 104 (although any number of UEs may be used). In this example, UE 102 and UE 104 are illustrated as smartphones (e.g., handheld touchscreen mobile computing devices capable of connecting to one or more cellular networks), but may include any mobile or non-mobile computing device configured for wireless communication.

[0029] UE 102 and UE 104 may be configured to be communicatively coupled to RAN 106. In an embodiment, RAN 106 may be NG-RAN, E-UTRAN, etc. UE 102 and UE 104 utilize connections (or channels) (shown as connection 108 and connection 110, respectively) with RAN 106, where each connection (or channel) includes a physical communication interface. RAN 106 may include one or more base stations, such as base station 112 and base station 114, that implement connection 108 and connection 110.

[0030] In this example, connections 108 and 110 are the air interfaces used to achieve this communicative coupling and may conform to the RAT used by RAN 106, such as, for example, LTE and / or NR.

[0031] In some embodiments, UE 102 and UE 104 may also directly exchange communication data via side link interface 116. UE 104 is shown as being configured to access an access point (shown as AP 118) via connection 120. For example, connection 120 may include a local wireless connection, such as a connection compliant with any IEEE 802.11 protocol, wherein AP 118 may include In this example, AP 118 may not be connected to another network (eg, the Internet) through CN 124.

[0032] In an embodiment, UE 102 and UE 104 may be configured to communicate with each other or with base station 112 and / or base station 114 over a multi-carrier communication channel using orthogonal frequency division multiplexing (OFDM) communication signals according to various communication techniques, such as, but not limited to, an orthogonal frequency division multiple access (OFDMA) communication technique (e.g., for downlink communication) or a single carrier frequency division multiple access (SC-FDMA) communication technique (e.g., for uplink and ProSe or sidelink communication), although the scope of the embodiment is not limited in this respect. An OFDM signal may include multiple orthogonal subcarriers.

[0033] In some embodiments, all or part of base station 112 or base station 114 may be implemented as one or more software entities running on a server computer as part of a virtual network. Additionally, or in other embodiments, base station 112 or base station 114 may be configured to communicate with each other via interface 122. In embodiments where wireless communication system 100 is an LTE system (e.g., when CN 124 is an EPC), interface 122 may be an X2 interface. This X2 interface may be defined between two or more base stations (e.g., two or more eNBs, etc.) connected to an EPC and / or between two eNBs connected to an EPC. In embodiments where wireless communication system 100 is an NR system (e.g., when CN 124 is a 5GC), interface 122 may be an Xn interface. This Xn interface is defined between two or more base stations (e.g., two or more gNBs, etc.) connected to a 5GC, between base station 112 (e.g., a gNB) and an eNB connected to a 5GC, and / or between two eNBs connected to a 5GC (e.g., CN 124).

[0034] The RAN 106 is shown as being communicatively coupled to the CN 124. The CN 124 may include one or more network elements 126 configured to provide various data and telecommunication services to customers / subscribers (e.g., UE 102 and users of UE 104) connected to the CN 124 via the RAN 106. The components of the CN 124 may be implemented in one physical device or separate physical devices that include components for reading and executing instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium).

[0035] In an embodiment, CN 124 may be an EPC, and RAN 106 may be connected to CN 124 via an S1 interface 128. In an embodiment, S1 interface 128 may be divided into two parts: an S1 user plane (S1-U) interface, which carries traffic data between base station 112 or base station 114 and a serving gateway (S-GW); and an S1-MME interface, which is a signaling interface between base station 112 or base station 114 and a mobility management entity (MME).

[0036] In an embodiment, CN 124 may be a 5GC, and RAN 106 may be connected to CN 124 via an NG interface 128. In an embodiment, NG interface 128 may be divided into two parts: an NG user plane (NG-U) interface, which carries traffic data between base station 112 or base station 114 and a user plane function (UPF); and an S1 control plane (NG-C) interface, which is a signaling interface between base station 112 or base station 114 and an access and mobility management function (AMF).

[0037] Generally speaking, application server 130 may be an element that provides applications (e.g., packet-switched data services) that utilize Internet Protocol (IP) bearer resources with CN 124. Application server 130 may also be configured to support one or more communication services (e.g., VoIP sessions, group communication sessions, etc.) for UE 102 and UE 104 via CN 124. Application server 130 may communicate with CN 124 via IP communication interface 132.

[0038] Figure 2 A system 200 is illustrated for performing signaling 234 between a wireless device 202 and a network device 218 according to embodiments disclosed herein. The system 200 can be part of a wireless communication system as described herein. The wireless device 202 can be, for example, a UE of the wireless communication system. The network device 218 can be, for example, a base station (e.g., an eNB or gNB) of the wireless communication system.

[0039] The wireless device 202 may include one or more processors 204. The processor 204 may execute instructions to perform various operations for the wireless device 202, as described herein. The processor 204 may include one or more baseband processors implemented using, for example, a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a controller, a field programmable gate array (FPGA) device, another hardware device, a firmware device, or any combination thereof, configured to perform the operations described herein.

[0040] The wireless device 202 may include a memory 206. The memory 206 may be a non-transitory computer-readable storage medium that stores instructions 208 (these instructions may include, for example, instructions to be executed by the processor 204). The instructions 208 may also be referred to as program code or a computer program. The memory 206 may also store data used by the processor 204 and results computed by the processor.

[0041] The wireless device 202 may include one or more transceivers 210, which may include radio frequency (RF) transmitter and / or receiver circuitry that facilitates signaling (e.g., signaling 234) between the wireless device 202 and other devices (e.g., network device 218) according to a corresponding RAT using an antenna 212 of the wireless device 202.

[0042] The wireless device 202 may include one or more antennas 212 (e.g., one, two, four, or more). For embodiments with multiple antennas 212, the wireless device 202 may leverage the spatial diversity of such multiple antennas 212 to transmit and / or receive multiple different data streams on the same time-frequency resources. This behavior may be referred to as, for example, multiple-input, multiple-output (MIMO) behavior (referring to the multiple antennas used at each of the transmitting and receiving devices to implement this aspect). MIMO transmission by the wireless device 202 may be implemented based on precoding (or digital beamforming) applied at the wireless device 202, which multiplexes the data streams across the antennas 212 based on known or assumed channel characteristics, such that each data stream is received at an appropriate signal strength relative to the other streams and at a desired location in the spatial domain (e.g., the location of the receiver associated with that data stream). Certain embodiments may utilize single-user MIMO (SU-MIMO) methods (where data streams are all directed to a single receiver) and / or multi-user MIMO (MU-MIMO) methods (where individual data streams may be directed to separate (different) receivers at different locations in the spatial domain).

[0043] In certain embodiments with multiple antennas, the wireless device 202 may implement analog beamforming techniques whereby the phases of the signals transmitted by the antennas 212 are relatively adjusted so that the (joint) transmissions of the antennas 212 can be directed (this is sometimes referred to as beam steering).

[0044] The wireless device 202 may include one or more interfaces 214. The interfaces 214 may be used to provide input to or output from the wireless device 202. For example, the wireless device 202 (UE) may include interfaces 214, such as a microphone, a speaker, a touch screen, and buttons, to allow a user of the UE to provide input to and / or output to the UE. Other interfaces of such a UE may consist of transmitters, receivers, and other circuits (e.g., in addition to the transceiver 210 / antenna 212 already described) that allow the UE to communicate with other devices, and may communicate according to known protocols (e.g., and etc.) to perform the operation.

[0045] The network device 218 may include one or more processors 220. The processor 220 may execute instructions to perform various operations of the network device 218, as described herein. The processor 204 may include one or more baseband processors implemented using, for example, a CPU, a DSP, an ASIC, a controller, an FPGA device, another hardware device, a firmware device, or any combination thereof configured to perform the operations described herein.

[0046] The network device 218 may include a memory 222. The memory 222 may be a non-transitory computer-readable storage medium that stores instructions 224 (which may include, for example, instructions to be executed by the processor 220). The instructions 224 may also be referred to as program code or a computer program. The memory 222 may also store data used by the processor 220 and results calculated by the processor.

[0047] The network device 218 may include one or more transceivers 226, which may include RF transmitter and / or receiver circuitry that facilitates signaling (e.g., signaling 234) between the network device 218 and other devices (e.g., wireless device 202) using an antenna 228 of the network device 218 in accordance with a corresponding RAT.

[0048] The network device 218 may include one or more antennas 228 (e.g., one, two, four, or more). In embodiments with multiple antennas 228, the network device 218 may perform MIMO, digital beamforming, analog beamforming, beamsteering, etc. as described.

[0049] The network device 218 may include one or more interfaces 230. The interfaces 230 may be used to provide input to or output from the network device 218. For example, a network device 218 that is a base station may include an interface 230 consisting of a transmitter, a receiver, and other circuits (e.g., in addition to the transceiver 226 / antenna 228 already described). These interfaces enable the base station to communicate with other equipment in the core network and / or enable the base station to communicate with external networks, computers, databases, etc., to achieve the purpose of operating, managing, and maintaining the base station or other equipment operably connected to the base station.

[0050] AI / ML can be applied to wireless communication systems. Use cases include channel state information (CSI) feedback enhancement (e.g., overhead reduction, improved accuracy and prediction), beam management (BM) (e.g., beam prediction in the time and spatial domains for overhead and delay reduction, and beam selection accuracy improvement), and positioning accuracy enhancement for different scenarios (including, for example, those with severe NLOS conditions).

[0051] Figure 3 An example functional framework for AI / ML in a wireless communication system according to the embodiments disclosed herein is illustrated.

[0052] Data collection 302 is a function that provides input data to the model training and model inference functions. AI / ML algorithm-specific data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) is not performed in the data collection function. Examples of input data may include measurements from the UE or various network entities, feedback from actors, and output from AI / ML models. Training data is the data required as input to the AI / ML model training function. Inference data is the data required as input to the AI / ML model inference function.

[0053] Model training 304 is a function that performs AI / ML model training, validation, and testing, which can generate model performance metrics as part of the model testing process. If necessary, the model training function is also responsible for data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) based on the training data delivered by the data collection function. Model deployment / update is the initial deployment of the trained, validated, and tested AI / ML model to the model inference function or the delivery of an updated model to the model inference function.

[0054] Model reasoning 306 is a function that provides AI / ML model reasoning output (e.g., prediction or decision). When applicable, the model reasoning function can provide model performance feedback to the model training function. If necessary, the model reasoning function is also responsible for data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) based on the reasoning data delivered by the data collection function. The output is the reasoning output of the AI / ML model generated by the model reasoning function. The details of the reasoning output are use case specific. When available, model performance feedback can be used to monitor the performance of the AI / ML model.

[0055] Actors 308 receive output from the model inference function and trigger or execute corresponding actions. Actors can trigger actions for other entities or for themselves. Feedback is the information that may be needed to export training data, inference data, or monitor the performance of the AI / ML model and its impact on the network by updating KPIs and performance counters (for example, through the model monitoring function).

[0056] The datasets required for model training, model monitoring, or model inference at an entity can be shared with other entities. The present disclosure provides an apparatus and method for sharing datasets for AI / ML between different entities in a wireless communication system.

[0057] An example communication process is described below to illustrate various aspects of the present disclosure. The example communication process is described with respect to AI. However, reference to AI is provided for illustrative purposes only and may be replaced with ML.

[0058] Figure 4 This section illustrates an example communication process between a UE and a gNB for preparing for dataset sharing according to the embodiments disclosed herein. The dataset sharing preparation includes AI capability exchange, dataset sharing consent, and model registration, and can be performed before data sharing between different entities.

[0059] First, the UE and gNB exchange AI capabilities. At step 402, when the gNB needs to know the UE's AI capabilities, it sends a capability query to the UE. At step S404, in response to the received capability query, the UE sends capability information indicating the UE's AI capabilities to the gNB.

[0060] A data set sharing consent exchange is then performed between the UE and the gNB. At step 406, the gNB sends a data set sharing query to the UE, querying whether the UE agrees to data set sharing. At step 408, in response to the received data set sharing query, the UE sends a data set sharing response, indicating that the UE agrees to data set sharing.

[0061] Finally, model registration is performed between the US and the gNB. At step S410, the UE sends a model registration request to the gNB to register the list of AI models available at the UE. The model registration request may include the model ID of the available AI model. In some implementations, the model ID may be generated by the UE and unique with respect to the PLMN. For example, PLMN information may be part of the model ID. In some implementations, the model ID may be assigned by the network. The model registration request may also include metadata of the available AI model. The AI ​​model metadata describes the AI ​​model and may include one or more of the following: training status, use case, functionality / purpose (e.g., input / output of the AI ​​model), requirements / performance (e.g., latency benchmark of the ML model, memory requirements of the ML model, accuracy of the ML model), compression status of the AI ​​model, or inference / operation conditions (e.g., urban, indoor, dense macro, SINR). The model registration request may also include the model status of the AI ​​model at the UE, such as supported but not downloaded, well-trained, poorly trained, etc.

[0062] At step S412, in response to the received model registration request, the gNB sends a model registration response to the UE. The model registration response may include the model ID of the supported AI model. If a sub-dataset ID is assigned to an AI model among the supported AI models, the model registration response may also include the sub-dataset ID. The sub-dataset ID may identify a sub-dataset of the dataset of the AI ​​model. If no sub-dataset ID is assigned, the dataset for the AI ​​model may be identified by the model ID of the AI ​​model. In other words, one model ID may identify one dataset.

[0063] The Model Registration Request and Model Registration Response exchanged between the UE and gNB regarding AI model registration can be radio resource control (RRC) messages, non-access stratum (NAS) messages, or user plane (UP) services. The UE may transmit a Model Registration Request when its list of available AI models changes. If the Model Registration Request is a UP service, the Model Registration Request may be part of the IP payload (i.e., part of an IP packet), or included in a new Layer 2 control PDU with a special QFI (QoS Flow Identifier), or included in a reserved DRB (Data Radio Bearer), or followed by a Medium Access Control (MAC) Control Element (CE). Depending on the type of Model Registration Request, the network termination entity for the Model Registration Request may vary. For example, if the Model Registration Request is an RRC message, the model information may be stored in the gNB; if the Model Registration Request is a NAS message, the model information may be stored in the core network; or if the Model Registration Request is a UP service, the model information may be stored in the OAM.

[0064] Figure 5 Another example communication process for preparing dataset sharing between a UE, a gNB, and a unified data management (UDM) according to the embodiments disclosed herein is illustrated. The dataset sharing preparation includes AI capability exchange, dataset sharing consent, and model registration, and can be performed before data sharing between different entities. Steps S502 and S504 for exchanging AI capabilities can be performed with Figure 4 Steps S402 and S404 are the same as those in step 540, and steps S510 and S512 for exchanging model registrations can be the same as those in step 540. Figure 4 However, steps S506 and S508 for exchanging data set sharing consent are the same as Figure 4 Steps S406 and S408 are different. Figure 5In this scenario, the gNB obtains the UE's consent for dataset sharing from the UDM rather than the UE. At step S506, the gNB sends a dataset sharing query to the UDM to inquire whether the UE agrees to dataset sharing. At step S508, the gNB receives a dataset sharing response from the UDM indicating that the UE agrees to dataset sharing.

[0065] Figure 6 The following illustrates an example communication process for sharing a dataset from a UE to a gNB according to the embodiments disclosed herein. Prior to dataset sharing, capability exchange and online / offline training / inference may be performed between the UE and the gNB. At step S602, the UE and the gNB perform capability exchange, enabling the gNB to learn the UE's AI capabilities. Examples of capability exchange include: Figure 4 Steps S402 and S404 and Figure 5 At step S604, online / offline training / inference may be performed by the UE and / or the gNB.

[0066] Dataset sharing can then be performed between the UE and gNB. Due to large payload overhead and varying data quality, it may not always be necessary to transmit data sets between the gNB and UE. If key data set information can be exchanged between the UE and gNB, the necessity of data set transmission can be determined based on this information, and unnecessary transmission can be avoided, thereby reducing communication overhead.

[0067] At step S608, the UE sends a dataset notification to the gNB indicating the presence of a dataset for an AI model available at the UE. The dataset notification may be an RRC message and may include a model ID for the AI ​​model. The dataset notification may also include metadata about the dataset. The metadata about the dataset describes the dataset and may include one or more of the following: data type (e.g., measurement results or ground truth labels), data usage (e.g., model training, model monitoring, or model inference), a use case indicating the results of applying the AI ​​model (e.g., CSI compression, beam management, positioning), sample size, payload size or required memory size before and / or after compression, last update time (e.g., a timestamp of the last update time), environmental information (e.g., SINR / RSRP / RSRQ values / histograms, UE location, serving / neighboring cell IDs), or a percentage of individual data samples. To enhance privacy, the use case may be indirectly indicated, for example, by using an identifier.

[0068] The dataset notification may also include dataset quality of the dataset. The dataset quality of the dataset may include one or more of the following: accuracy (e.g., inference accuracy, training accuracy, and ground truth label accuracy), variance, completeness, or latency (e.g., the delay between generating the original data and collecting the data by the UE).

[0069] In some implementations, a dataset notification can be sent upon request (e.g., in response to a dataset query sent from the gNB at step S606 to inquire whether there are any datasets for the AI ​​model available at the UE). The dataset query can be an RRC message and can include a model ID of the AI ​​model and can also include the data type of the dataset (e.g., measurement results or ground truth labels). If a sub-dataset ID is assigned, the dataset notification can also include the sub-dataset ID after the model ID.

[0070] In some implementations, step S606 is not necessary, and the data set notification may be sent periodically. The time interval for sending the data set notification may be configured by an RRC message (eg, RRCReconfiguration).

[0071] In some implementations, step S606 is not necessary, and the data set notification may be triggered by an event (e.g., a condition being met). The condition may be configured by an RRC message (e.g., RRCReconfiguration) and may include one or more thresholds for data set quality.

[0072] At step S610, the gNB determines that a data set is required. The gNB may determine the required data set based on metadata of the data set. Alternatively, the gNB may determine the required data set based on data set quality (e.g., if the data set quality exceeds a threshold). This allows the gNB to obtain a data set of higher quality and avoid transmitting a data set of lower quality, thereby reducing communication overhead.

[0073] At step S612, in response to the determination in step 610, the gNB sends a dataset request to the UE requesting the dataset. The dataset request may include the model ID of the AI ​​model and may also include the data type of the dataset (e.g., measurement results or ground truth labels). If a sub-dataset ID is assigned, the dataset request may also include the sub-dataset ID after the model ID. Similar to a dataset query, the dataset request may be sent on request, periodically, or triggered by an event.

[0074] The dataset request may be an RRC message or a UP service. If the dataset request is an RRC message, a new RRC message (e.g., DatasetRequest) may be introduced to support on-demand transmission of dataset requests. A prohibit timer may be configured to be associated with the dataset request to avoid sending the dataset request too frequently. If the dataset request is a UP service, the dataset request may be part of the IP payload (i.e., part of the IP packet), or in a new Layer 2 control PDU with a special QFI (QoS Flow Identifier) ​​or in a reserved DRB (Data Radio Bearer), or after a Medium Access Control (MAC) Control Element (CE).

[0075] At step S614, the UE sends a dataset report including the dataset to the gNB. The dataset report may also include a model ID, metadata, and dataset quality. In some implementations, the dataset report is an RRC message. The address for downloading the dataset from the server (e.g., a server URL) is included in the RRC message, or the dataset is included in a transparent container in the RRC message. The model ID, metadata, and dataset quality (if necessary) are included in an IE (Information Element) of the RRC message. Thus, the dataset can be transmitted as a server URL or in an existing open format. In some implementations, the dataset report is a non-access stratum (NAS) message. In some implementations, the dataset report is user plane (UP) traffic. For example, UP traffic can be part of an IP packet or in a Layer 2 control PDU with a reserved 5G QoS Indication (5QI) or QoS Flow Identifier (QFI). For another example, on a reserved data radio bearer (DRB), UP traffic can be established via specific control signaling with a reserved 5QI or QFI, or the UP traffic can be behind a medium access control (MAC) control element (CE). For another example, the control signaling used to establish a PDU session to include a data set can use a reserved 5QI / QFI. Depending on the different requirements for data set transmission, different 5QI / QFIs can be assigned to the established PDU session. This provides greater flexibility if the UE needs to transmit more than one data set.

[0076] Figure 7 The following illustrates an example communication process for sharing a data set from a gNB to a UE according to the embodiments disclosed herein. Prior to data set sharing, capability exchange and online / offline training / inference may be performed between the UE and the gNB. At step S702, the UE and the gNB perform capability exchange, enabling the gNB to learn the UE's AI capabilities. Examples of capability exchange include: Figure 4 Steps S402 and S404 and Figure 5 At step S704, online / offline training / inference may be performed by the UE and / or the gNB.

[0077] Dataset sharing can then be performed between the UE and gNB. Due to large payload overhead and varying data quality, it may not always be necessary to transmit data sets between the gNB and UE. If key data set information can be exchanged between the UE and gNB, the necessity of data set transmission can be determined based on this information, and unnecessary transmission can be avoided, thereby reducing communication overhead.

[0078] At step S708, the gNB sends a dataset notification to the UE, indicating the presence of a dataset for an AI model available at the gNB. The dataset notification may be an RRC message and may include a model ID for the AI ​​model. The dataset notification may also include metadata about the dataset. The metadata about the dataset describes the dataset and may include one or more of the following: data type (e.g., measurement results or ground truth labels), data usage (e.g., model training, model monitoring, or model inference), a use case indicating the results of applying the AI ​​model (e.g., CSI compression, beam management, positioning), sample size, payload size or required memory size before and / or after compression, last update time (e.g., a timestamp of the last update time), environmental information (e.g., SINR / RSRP / RSRQ values / histograms, gNB location, serving / neighboring cell IDs), or a percentage of individual data samples. To enhance privacy, the use case may be indirectly indicated, for example, by using an identifier.

[0079] The dataset notification may also include dataset quality of the dataset. The dataset quality of the dataset may include one or more of the following: accuracy (e.g., inference accuracy, training accuracy, and ground truth label accuracy), variance, completeness, or latency (e.g., the delay between generating the original data and collecting the data by the UE).

[0080] In some implementations, a dataset notification can be sent upon request (e.g., in response to a dataset query sent from the UE at step S706 to inquire whether there are any datasets for the AI ​​model available at the gNB). The dataset query can be an RRC message and can include a model ID of the AI ​​model and can also include the data type of the dataset (e.g., measurement results or ground truth labels). If a sub-dataset ID is assigned, the dataset notification can also include the sub-dataset ID after the model ID. The dataset notification can be an RRC message (e.g., UE assistance information or DatasetQualityRequest).

[0081] In some implementations, step S706 is not necessary and the dataset notification may be sent periodically. However, whether / when to transmit the dataset notification may depend on the gNB implementation.

[0082] At step S710, the UE determines that a data set is required. The UE may determine the required data set based on metadata of the data set. Alternatively, the UE may determine the required data set based on the quality of the data set (e.g., if the data set quality is greater than a threshold). In this way, the UE can obtain a data set of higher quality and avoid transmitting a data set of lower quality, thereby reducing communication overhead.

[0083] At step S712, in response to the determination in step 710, the UE sends a dataset request to the gNB requesting the dataset. The dataset request may include the model ID of the AI ​​model and may also include the data type of the dataset (e.g., measurement results or ground truth labels). If a sub-dataset ID is assigned, the dataset request may also include the sub-dataset ID after the model ID. Similar to a dataset query, the dataset request may be sent on a per-request basis, periodically, or triggered by an event.

[0084] The dataset request may be an RRC message or a UP service. If the dataset request is an RRC message, a new RRC message (e.g., DatasetRequest) may be introduced to support on-demand transmission of dataset requests. A prohibit timer may be configured to be associated with the dataset request to avoid sending the dataset request too frequently. If the dataset request is a UP service, the dataset request may be part of the IP payload (i.e., part of the IP packet), or in a new Layer 2 control PDU with a special QFI (QoS Flow Identifier) ​​or in a reserved DRB (Data Radio Bearer), or after a Medium Access Control (MAC) Control Element (CE).

[0085] At step S714, the gNB sends a dataset report including the dataset to the UE. The dataset report may also include a model ID, metadata, and dataset quality. In some implementations, the dataset report is an RRC message. The address for downloading the dataset from the server (e.g., a server URL) is included in the RRC message, or the dataset is included in a transparent container in the RRC message. The model ID, metadata, and dataset quality (if necessary) may be included in an IE (Information Element) of the RRC message. Thus, the dataset may be transmitted as a server URL or in an existing open format. In some implementations, the dataset report is a non-access stratum (NAS) message. In some implementations, the dataset report is user plane (UP) traffic. For example, UP traffic may be part of an IP packet or in a Layer 2 control PDU with a reserved 5G QoS Indication (5QI) or QoS Flow Identifier (QFI). For another example, UP traffic can be established on a reserved data radio bearer (DRB) via specific control signaling with a reserved 5QI or QFI, or it can be behind a medium access control (MAC) control element (CE). For another example, control signaling used to establish a PDU session to include a data set can use a reserved 5QI / QFI. Different 5QI / QFIs can be assigned to established PDU sessions based on the different requirements for data set transmission. This provides greater flexibility if the gNB needs to transmit more than one data set. However, whether / when data set reports are delivered to the UE via RRC messages or UP traffic may be gNB implementation-specific.

[0086] In this disclosure, RRC messages, NAS messages, and UP services can be selected based on the different latency, robustness, and payload size requirements of the actual system. Compared to NAS and UP, RRC messages can result in lower signaling latency because they terminate in the gNB. RRC messages can be incrementally configured and are more robust than UP services. However, RRC messages cannot deliver data sets with large payload sizes and require RRC messages to support segmentation, which places higher demands on RRC messages. NAS messages can be more robust than UP services and support the delivery of data sets with large payload sizes. However, compared to RRC messages, NAS messages may incur higher latency. UP services support the delivery of data sets with large payload sizes. However, compared to RRC messages, UP services incur higher latency and are less robust than RRC and NAS.

[0087] Figure 8An example method performed by a first communication device according to embodiments disclosed herein is illustrated. At step S802, the first communication device sends a dataset notification including metadata of a dataset for an artificial intelligence (AI) model to a second communication device. At step S804, the first communication device receives a dataset request from the second communication device requesting the dataset. At step S806, the first communication device sends a dataset report including the dataset to the second communication device.

[0088] The above reference Figure 6 Any of the steps described in the UE or referenced above Figure 7 Any of the steps described in the gNB in ​​the embodiment of the present invention may be performed by the first communication device and will not be repeated here for the sake of brevity. Figure 8 All of the steps shown in are necessary, and the first communication device may only perform one or more of these steps.

[0089] Figure 9 An example method performed by a second communication device according to embodiments disclosed herein is illustrated. At step S902, the second communication device receives a dataset notification from a first communication device, including metadata of a dataset for an artificial intelligence (AI) model. At step S904, the second communication device sends a dataset request to the first communication device requesting the dataset. At step S906, the second communication device receives a dataset report from the first communication device including the dataset.

[0090] The above reference Figure 6 Any of the steps described in the gNB or referenced above Figure 7 Any of the steps described in the UE in the embodiment may be performed by the second communication device and will not be repeated here for the sake of brevity. Figure 9 All of the steps shown in are necessary, and the second communication device may only perform one or more of these steps.

[0091] Embodiments contemplated herein include a first communication device comprising: at least one antenna; and a processor configured to: send a dataset notification comprising metadata of a dataset for an artificial intelligence (AI) model to a second communication device; receive a dataset request from the second communication device requesting the dataset; and send a dataset report comprising the dataset to the second communication device.

[0092] In some embodiments of the present disclosure, the data set notification is sent periodically or triggered by an event.

[0093] In some embodiments of the present disclosure, the processor is further configured to: receive a dataset query from the second communication device to inquire whether there is any dataset for the AI ​​model, and wherein the dataset notification is sent in response to the dataset query.

[0094] In some embodiments of the present disclosure, the dataset query further includes the data type of the dataset.

[0095] In some embodiments of the present disclosure, the metadata of the data set includes one or more of the following: data type, data usage, use case, sample size, payload size or required memory size, latest update time, environmental information, or percentage of a single data sample.

[0096] In some embodiments of the present disclosure, the dataset notification further includes a dataset quality of the dataset.

[0097] In some embodiments of the present disclosure, the dataset quality includes one or more of the following: accuracy, variance, completeness, or latency.

[0098] In some embodiments of the present disclosure, the data set report is a radio resource control (RRC) message, and the data set is included in a transparent container of the RRC message.

[0099] In some embodiments of the present disclosure, the data set report is a radio resource control (RRC) message, and an address for downloading the data set from a server is included in the RRC message.

[0100] In some embodiments of the present disclosure, the data set report is a non-access stratum (NAS) message.

[0101] In some embodiments of the present disclosure, the data set report is user plane (UP) traffic.

[0102] In some embodiments of the present disclosure, the UP service is part of an IP packet, or in a layer 2 control PDU with a reserved 5G QoS indication (5QI) or QoS flow identifier (QFI), or on a reserved data radio bearer (DRB); the UP service is established via specific control signaling with a reserved 5QI or QFI; or the UP service is after a medium access control (MAC) control element (CE).

[0103] In some embodiments of the present disclosure, the processor is further configured to: receive a dataset sharing query from the second communication device to inquire whether the first communication device agrees to dataset sharing; and send a dataset sharing response to the second communication device indicating that the first communication device agrees to dataset sharing.

[0104] In some embodiments of the present disclosure, the processor is further configured to: send a model registration request including model IDs of available AI models and model descriptions of these available AI models to the second communication device; and receive a model registration response including model IDs of supported AI models from the second communication device.

[0105] In some embodiments of the present disclosure, the model registration response also includes a sub-dataset ID assigned to at least one of the supported AI models.

[0106] In some embodiments of the present disclosure, each of the model registration request and the model registration response is a radio resource control (RRC) message, a non-access stratum (NAS) message, or a user plane (UP) traffic.

[0107] Embodiments contemplated herein include an apparatus comprising means for performing one or more elements of a method comprising: sending a dataset notification comprising meta-information of a dataset for an artificial intelligence (AI) model to a second communication device; receiving a dataset request from the second communication device requesting the dataset; and sending a dataset report comprising the dataset to the second communication device. The apparatus may be, for example, Figure 6 The UE device or Figure 7 The gNB device in the.

[0108] Embodiments contemplated herein include one or more non-transitory computer-readable media including instructions that cause an electronic device, when executed by one or more processors of the electronic device, to perform one or more elements of a method comprising: sending a dataset notification including metadata of a dataset for an artificial intelligence (AI) model to a second communication device; receiving a dataset request from the second communication device requesting the dataset; and sending a dataset report including the dataset to the second communication device. The non-transitory computer-readable medium may be, for example, Figure 6 UE memory or Figure 7 The memory of the gNB in

[0109] Embodiments contemplated herein include an apparatus comprising logic components, modules, or circuits for performing one or more elements of a method comprising: sending a dataset notification comprising metadata of a dataset for an artificial intelligence (AI) model to a second communication device; receiving a dataset request from the second communication device requesting the dataset; and sending a dataset report comprising the dataset to the second communication device. The apparatus may be, for example, Figure 6 The UE device or Figure 7 The gNB device in the.

[0110] Embodiments contemplated herein include an apparatus comprising one or more processors and one or more computer-readable media comprising instructions that, when executed by the one or more processors, cause the one or more processors to perform one or more elements of a method comprising: sending a dataset notification comprising meta-information of a dataset for an artificial intelligence (AI) model to a second communication device; receiving a dataset request from the second communication device requesting the dataset; and sending a dataset report comprising the dataset to the second communication device. The apparatus may be, for example, Figure 6 The UE device or Figure 7 The gNB device in the.

[0111] Embodiments contemplated herein include a signal as described in or relating to one or more elements of a method comprising the following operations: sending a dataset notification comprising metadata of a dataset for an artificial intelligence (AI) model to a second communication device; receiving a dataset request from the second communication device to obtain the dataset; and sending a dataset report comprising the dataset to the second communication device.

[0112] Embodiments contemplated herein include a computer program or computer program product comprising instructions, wherein execution of the program by a processor causes the processor to perform one or more elements of a method comprising: sending a dataset notification comprising meta-information of a dataset for an artificial intelligence (AI) model to a second communication device; receiving a dataset request from the second communication device requesting the dataset; and sending a dataset report comprising the dataset to the second communication device. The processor may be Figure 6 The UE's processor, or Figure 7 These instructions may be located, for example, at the processor of the gNB. Figure 6 UE or Figure 7 in the processor of the gNB and / or in the memory of the UE or the gNB.

[0113] Embodiments contemplated herein include a second communication device comprising: at least one antenna; and a processor configured to: receive a dataset notification comprising metadata of a dataset for an artificial intelligence (AI) model from a first communication device; send a dataset request to the first communication device requesting the dataset; and receive a dataset report comprising the dataset from the first communication device.

[0114] In some embodiments of the present disclosure, the data set notification is sent periodically.

[0115] In some embodiments of the present disclosure, the processor is further configured to: send a dataset query to the first communication device to inquire whether there is any dataset for the AI ​​model.

[0116] In some embodiments of the present disclosure, the dataset query further includes the data type of the dataset.

[0117] In some embodiments of the present disclosure, the metadata of the data set includes one or more of the following: data type, data usage, use case, sample size, payload size or required memory size, latest update time, environmental information, or percentage of a single data sample.

[0118] In some embodiments of the present disclosure, the dataset notification further includes a dataset quality of the dataset.

[0119] In some embodiments of the present disclosure, the dataset quality includes one or more of the following: accuracy, variance, completeness, or latency.

[0120] In some embodiments of the present disclosure, the processor is further configured to: send a dataset sharing query to the first communication device to inquire whether the first communication device agrees to dataset sharing; and receive a dataset sharing response from the first communication device indicating that the first communication device agrees to dataset sharing.

[0121] In some embodiments of the present disclosure, the processor is further configured to: send a dataset sharing query to a unified data management (UDM) to inquire whether the first communication device agrees to dataset sharing; and receive a dataset sharing response from the UDM indicating that the first communication device agrees to dataset sharing.

[0122] In some embodiments of the present disclosure, the processor is further configured to: receive a model registration request including model IDs of available AI models and model descriptions of these available AI models from the first communication device; and send a model registration response including model IDs of supported AI models to the first communication device.

[0123] In some embodiments of the present disclosure, the model registration response also includes a sub-dataset ID assigned to at least one of the supported AI models.

[0124] Embodiments contemplated herein include an apparatus comprising means for performing one or more elements of a method comprising: receiving a dataset notification comprising meta-information of a dataset for an artificial intelligence (AI) model from a first communication device; sending a dataset request to the first communication device requesting the dataset; and receiving a dataset report comprising the dataset from the first communication device. The apparatus may be, for example, Figure 6 gNB in ​​or Figure 7 The UE device in the device.

[0125] Embodiments contemplated herein include one or more non-transitory computer-readable media including instructions that cause an electronic device, when executed by one or more processors of the electronic device, to perform one or more elements of a method comprising: receiving a dataset notification including metadata of a dataset for an artificial intelligence (AI) model from a first communication device; sending a dataset request to the first communication device requesting the dataset; and receiving a dataset report including the dataset from the first communication device. The non-transitory computer-readable medium may be, for example, Figure 6 gNB in ​​or Figure 7 The memory of the UE in .

[0126] Embodiments contemplated herein include an apparatus comprising logic components, modules, or circuits for performing one or more elements of a method comprising: receiving a dataset notification comprising metadata of a dataset for an artificial intelligence (AI) model from a first communication device; sending a dataset request to the first communication device requesting the dataset; and receiving a dataset report comprising the dataset from the first communication device. The apparatus may be, for example, Figure 6 gNB in ​​or Figure 7 The UE device in the device.

[0127] Embodiments contemplated herein include an apparatus comprising one or more processors and one or more computer-readable media comprising instructions that, when executed by the one or more processors, cause the one or more processors to perform one or more elements of a method comprising: receiving a dataset notification comprising meta-information of a dataset for an artificial intelligence (AI) model from a first communication device; sending a dataset request to the first communication device requesting the dataset; and receiving a dataset report comprising the dataset from the first communication device. The apparatus may be, for example, Figure 6 gNB in ​​or Figure 7 The UE device in the device.

[0128] Embodiments contemplated herein include a signal as described in or relating to one or more elements of a method comprising the following operations: receiving a dataset notification comprising metadata of a dataset for an artificial intelligence (AI) model from a first communication device; sending a dataset request to the first communication device requesting the dataset; and receiving a dataset report comprising the dataset from the first communication device.

[0129] Embodiments contemplated herein include a computer program or computer program product comprising instructions, wherein execution of the program by a processing element causes the processing element to perform one or more elements of a method comprising: receiving a dataset notification comprising meta-information of a dataset for an artificial intelligence (AI) model from a first communication device; sending a dataset request to the first communication device requesting the dataset; and receiving a dataset report comprising the dataset from the first communication device. The processor may be Figure 6 gNB in ​​or Figure 7 These instructions may be located, for example, at Figure 6 gNB in ​​or Figure 7 in a processor of the UE and / or in a memory of the gNB or the UE.

[0130] For one or more embodiments, at least one of the components described in one or more of the preceding figures may be configured to perform one or more operations, techniques, processes, and / or methods as described herein. For example, a baseband processor as described herein in conjunction with one or more of the preceding figures may be configured to operate according to one or more of the examples described herein. For another example, circuitry associated with a UE, base station, network element, etc., as described above in conjunction with one or more of the preceding figures, may be configured to operate according to one or more of the examples described herein.

[0131] Unless otherwise expressly stated, any of the above embodiments may be combined with any other embodiment (or combination of embodiments). The foregoing description of one or more specific implementations provides illustration and description, but is not intended to be exhaustive or to limit the scope of the embodiments to the precise forms disclosed. In view of the above teachings, modifications and variations are possible or can be obtained from the practice of the various embodiments.

[0132] Embodiments and implementations of the systems and methods described herein may include various operations that may be embodied in machine-executable instructions to be executed by a computer system. A computer system may include one or more general-purpose or special-purpose computers (or other electronic devices). A computer system may include hardware components that include specific logic for performing the operations; or may include a combination of hardware, software, and / or firmware.

[0133] It should be appreciated that the systems described herein include descriptions of specific embodiments. These embodiments can be combined into a single system, partially combined into other systems, separated into multiple systems, or otherwise divided or combined. In addition, it is contemplated that parameters, attributes, aspects, etc. of one embodiment may be used in conjunction with another embodiment. For clarity, these parameters, attributes, aspects, etc. are described only in relation to one or more embodiments, and it should be appreciated that these parameters, attributes, aspects, etc. may be combined with or substituted for parameters, attributes, aspects, etc. of another embodiment unless expressly stated otherwise herein.

[0134] It is understood that the use of personally identifiable information should be subject to privacy policies and practices that are generally recognized to meet or exceed industry or government requirements for maintaining user privacy. Specifically, personally identifiable information data should be managed and processed to minimize the risk of unintentional or unauthorized access or use, and the nature of authorized use should be clearly stated to users.

[0135] Although the foregoing has been described in considerable detail for purposes of clarity, it will be apparent that certain changes and modifications may be made without departing from the principles of the invention. It should be noted that there are many alternative ways of implementing both the processes and the apparatus described herein. The embodiments of the present invention are therefore to be considered illustrative and not restrictive, and the description is not to be limited to the details given herein, but may be modified within the scope and equivalents of the appended claims.

Claims

1. A first communication device, comprising: at least one antenna; and A processor configured to: sending a data set notification including meta information of a data set for an artificial intelligence (AI) model to a second communication device; receiving a data set request from the second communication device to obtain the data set; as well as A data set report including the data set is sent to the second communications device. 2 . The first communication device according to claim 1 , wherein the data set notification is sent periodically or is triggered by an event.

3. The first communication device of claim 1 , wherein the processor is further configured to: receive a dataset query from the second communication device to inquire whether there are any datasets for the AI ​​model, and wherein the dataset notification is sent in response to the dataset query. The first communication device according to claim 3 , wherein the data set query further includes a data type of the data set.

5. The first communication device according to claim 1, wherein the meta-information of the data set includes one or more of the following: data type, data usage, use case, sample size, payload size or required memory size, latest update time, environmental information or percentage of a single data sample. The first communication device of claim 1 , wherein the data set notification further comprises a data set quality of the data set. 7 . The first communication device of claim 6 , wherein the data set quality comprises one or more of: accuracy, variance, completeness, or latency.

8. The first communication device of claim 1, wherein the data set report is a radio resource control (RRC) message, and the data set is included in a transparent container of the RRC message. 9 . The first communication device according to claim 1 , wherein the data set report is a radio resource control (RRC) message, and an address for downloading the data set from a server is included in the RRC message.

10. The first communications device of claim 1, wherein the data set report is a non-access stratum (NAS) message.

11. The first communication device of claim 1, wherein the data set report is user plane (UP) traffic.

12. The first communication device according to claim 11, wherein one of the following situations exists: The UP service is part of an IP packet, or in a Layer 2 control PDU with a reserved 5G QoS Indication (5QI) or QoS Flow Identifier (QFI), or on a reserved Data Radio Bearer (DRB); The UP service is established via specific control signaling with a reserved 5QI or QFI; or The UP service is behind a medium access control (MAC) control element (CE).

13. The first communication device of claim 1 , wherein the processor is further configured to: receiving, from the second communication device, a data set sharing query for querying whether the first communication device agrees to data set sharing; and A data set sharing response is sent to the second communication device indicating that the first communication device agrees to the data set sharing.

14. The first communication device of claim 1 , wherein the processor is further configured to: Sending a model registration request including a model ID of an available AI model and a model description of the available AI model to the second communication device; and A model registration response including a model ID of a supported AI model is received from the second communication device. 15 . The first communications device of claim 14 , wherein the model registration response further comprises a sub-dataset ID assigned to at least one of the supported AI models. 16 . The first communication device of claim 14 , wherein each of the model registration request and the model registration response is a radio resource control (RRC) message, a non-access stratum (NAS) message, or a user plane (UP) traffic.

17. A second communication device, comprising: at least one antenna; and A processor configured to: receiving, from a first communication device, a dataset notification including meta information of a dataset for an artificial intelligence (AI) model; Sending a data set request to the first communication device to obtain the data set; as well as A data set report including the data set is received from the first communications device. The second communication device according to claim 17 , wherein the data set notification is sent periodically.

19. The second communication device of claim 17, wherein the processor is further configured to send a dataset query to the first communication device to inquire whether there is any dataset for the AI ​​model.

20. The second communication device of claim 19, wherein the data set query further includes a data type of the data set.

21. The second communication device of claim 17, wherein the meta-information of the data set comprises one or more of the following: data type, data usage, use case, sample size, payload size or required memory size, latest update time, environmental information, or percentage of a single data sample.

22. The second communications device of claim 17, wherein the data set notification further comprises a data set quality of the data set.

23. The second communications device of claim 22, wherein the data set quality comprises one or more of: accuracy, variance, completeness, or latency.

24. The second communication device of claim 17, wherein the processor is further configured to: sending a data set sharing query to the first communication device to inquire whether the first communication device agrees to share the data set; and A data set sharing response is received from the first communication device indicating that the first communication device agrees to data set sharing.

25. The second communication device of claim 17, wherein the processor is further configured to: sending a data set sharing query to a unified data management (UDM) to inquire whether the first communication device agrees to share the data set; and A data set sharing response is received from the UDM indicating that the first communication device agrees to data set sharing.

26. The second communication device of claim 17, wherein the processor is further configured to: receiving, from the first communication device, a model registration request including a model ID of an available AI model and a model description of the available AI model; and A model registration response including a model ID of a supported AI model is sent to the first communication device. 27 . The first communications device of claim 26 , wherein the model registration response further comprises a sub-dataset ID assigned to at least one of the supported AI models.

28. A method performed by a first communication device, the method comprising: sending a data set notification including meta information of a data set for an artificial intelligence (AI) model to a second communication device; receiving a data set request from the second communication device to obtain the data set; as well as A data set report including the data set is sent to the second communications device.

29. A method performed by a second communication device, the method comprising: receiving, from a first communication device, a dataset notification including meta information of a dataset for an artificial intelligence (AI) model; Sending a data set request to the first communication device to obtain the data set; as well as A data set report including the data set is received from the first communications device.

30. A non-transitory computer-readable storage medium having instructions stored thereon, the instructions, when executed by a processor of a first communication device, causing the processor to: sending a data set notification including meta information of a data set for an artificial intelligence (AI) model to a second communication device; receiving a data set request from the second communication device requesting to obtain the data set; and A data set report including the data set is sent to the second communications device.

31. A non-transitory computer-readable storage medium having instructions stored thereon, the instructions, when executed by a processor of a second communication device, causing the processor to: receiving, from a first communication device, a dataset notification including meta information of a dataset for an artificial intelligence (AI) model; Sending a data set request to the first communication device to obtain the data set; as well as A data set report including the data set is received from the first communications device.

Images