Machine learning model configuration in wireless networks

Abstract

Certain aspects of the present disclosure provide techniques and apparatus for determining neural network functions (NNFs) and configuring and using corresponding machine learning (ML) models to perform one or more ML-based wireless communication management procedures. An exemplary method performed by a user equipment includes transmitting, to a base station (BS), UE capability information indicating at least one radio capability of the UE and at least one machine learning (ML) capability of the UE, and receiving, from the BS based on the UE capability information, ML configuration information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

Description

[0001] Cross-references to related applications

[0002] This application claims priority to U.S. Application No. 17 / 348,522, filed June 15, 2021, which has been assigned to the assignee of this application and is expressly incorporated herein by reference in its entirety, as if fully set forth below and for the purposes described herein. Background Technology

[0003] Various aspects of this disclosure relate to wireless communications, and more specifically to techniques for configuring machine learning models in wireless communication networks.

[0004] Wireless communication systems are widely deployed to provide a variety of telecommunications services, such as telephone, video, data, messaging, broadcasting, or other similar services. These wireless communication systems can employ multiple access technologies that enable communication with multiple users by sharing available system resources (e.g., bandwidth, transmit power, or other resources). Multiple access technologies can rely on any of code division, time division, frequency division, orthogonal frequency division, single-carrier frequency division, or time-division synchronous code division, to name just a few. These and other multiple access technologies have been adopted in various telecommunications standards to provide a common protocol enabling different wireless devices to communicate at the city, national, regional, and even global levels.

[0005] Despite the tremendous technological advancements in wireless communication systems over the years, challenges remain. For example, complex and dynamic environments can still attenuate or block signals between wireless transmitters and receivers, disrupting various established wireless channel measurement and reporting mechanisms used to manage and optimize the use of limited wireless channel resources. Therefore, there is a need for further improvements to wireless communication systems to overcome these challenges. Summary of the Invention

[0006] Some aspects can be implemented in a method for wireless communication performed by at least one of a base station (BS) or other network entities. The method typically includes: receiving from a user equipment (UE) UE capability information indicating at least one radio capability and at least one machine learning (ML) capability of the UE; and transmitting to the UE ML configuration information based on the UE capability information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0007] Certain aspects can be implemented in a device for wireless communication. The device typically includes: a memory containing computer-executable instructions; and one or more processors configured to execute the computer-executable instructions and enable the device to: receive from a user equipment (UE) UE capability information indicating at least one radio capability and at least one machine learning (ML) capability of the UE; and, based on the UE capability information, transmit to the UE ML configuration information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0008] Some aspects can be implemented in a device for wireless communication. The device may include: means for receiving UE capability information from a user equipment (UE) indicating at least one radio capability and at least one machine learning (ML) capability of the UE; and means for transmitting ML configuration information to the UE based on the UE capability information, indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0009] Certain aspects may be implemented in a non-transitory computer-readable medium. The non-transitory computer-readable medium may include computer-executable instructions that, when executed by one or more processors of the device, cause the device to: receive UE capability information from a user equipment (UE) indicating at least one radio capability and at least one machine learning (ML) capability of the UE; and, based on the UE capability information, transmit to the UE ML configuration information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0010] Certain aspects may be implemented in a computer program product implemented on a computer-readable storage medium. The computer-readable storage medium may include code for performing wireless communication by a base station (BS) or other network entity, including code for: receiving UE capability information from a user equipment (UE) indicating at least one radio capability and at least one machine learning (ML) capability of the UE; and transmitting ML configuration information to the UE based on the UE capability information, indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0011] Certain aspects can be implemented in a method for wireless communication performed by a user equipment (UE). This method typically includes: transmitting UE capability information to a base station (BS) or other network entity, indicating at least one radio capability and at least one machine learning (ML) capability of the UE; and receiving, based on the UE capability information, ML configuration information from the BS or other network entity, indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0012] Certain aspects can be implemented in a device for wireless communication. The device typically includes: a memory containing computer-executable instructions; and one or more processors configured to execute the computer-executable instructions and cause the device to: transmit capability information to a base station (BS) or other network entity indicating at least one radio capability and at least one machine learning (ML) capability of the device; and receive, based on the capability information, ML configuration information from the BS or other network entity indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0013] Some aspects can be implemented in a device for wireless communication. The device may include: means for transmitting capability information to a base station (BS) or other network entity indicating at least one radio capability of the device and at least one machine learning (ML) capability of the device; and means for receiving, based on the capability information, ML configuration information from the BS or other network entity indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0014] Certain aspects may be implemented in a non-transitory computer-readable medium. This non-transitory computer-readable medium may include computer-executable instructions that, when executed by one or more processors of the device, cause the device to: transmit capability information to a base station (BS) or other network entity indicating at least one radio capability and at least one machine learning (ML) capability of the device; and receive, based on the capability information, ML configuration information from the BS or other network entity indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0015] Certain aspects may be implemented in a computer program product implemented on a computer-readable storage medium. The computer-readable storage medium may include code for: transmitting capability information to a base station (BS) or other network entity indicating at least one radio capability of a user equipment (UE) and at least one machine learning (ML) capability of the UE; and receiving, based on the UE capability information, ML configuration information from the BS or other network entity indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0016] Certain aspects can be implemented in a method for wireless communication performed by a first network entity. This method typically includes: receiving a first signal from a second network entity, the first signal including a first indication of at least one machine learning (ML) capability of a user equipment (UE) and a second indication corresponding to at least one neural network function (NNF) of an ML-based wireless communication management process associated with the UE; selecting at least one ML model for performing at least a portion of the ML-based wireless communication management process, at least in part based on the at least one ML capability of the UE; and transmitting a second signal to the second network entity, the second signal including ML configuration information indicating the at least one NNF and the at least one ML model.

[0017] Certain aspects can be implemented in a device for wireless communication. The device typically includes: a memory containing computer-executable instructions; and one or more processors configured to execute the computer-executable instructions and enable the device to: receive a first signal from a network entity, the first signal including a first indication of at least one machine learning (ML) capability of a user equipment (UE) and a second indication corresponding to at least one neural network function (NNF) of an ML-based wireless communication management process associated with the UE; select at least one ML model for performing at least a portion of the ML-based wireless communication management process, at least in part based on the at least one ML capability of the UE; and transmit a second signal to the network entity, the second signal including ML configuration information indicating the at least one NNF and the at least one ML model.

[0018] Certain aspects can be implemented in an apparatus for wireless communication. The apparatus may include: means for receiving a first signal from a network entity, the first signal including a first indication of at least one machine learning (ML) capability of a user equipment (UE) and a second indication corresponding to at least one neural network function (NNF) of an ML-based wireless communication management process associated with the UE; means for selecting at least one ML model for performing at least a portion of the ML-based wireless communication management process, at least partially based on the at least one ML capability of the UE; and means for transmitting a second signal to the network entity, the second signal including ML configuration information indicating the at least one NNF and the at least one ML model.

[0019] Certain aspects can be implemented in a non-transitory computer-readable medium for wireless communication. The non-transitory computer-readable medium may include computer-executable instructions that, when executed by one or more processors of the device, cause the device to: receive a first signal from a network entity, the first signal including a first indication of at least one machine learning (ML) capability of a user equipment (UE) and a second indication corresponding to at least one neural network function (NNF) of an ML-based wireless communication management process associated with the UE; select at least one ML model for performing at least a portion of the ML-based wireless communication management process, at least in part based on the at least one ML capability of the UE; and transmit a second signal to the network entity, the second signal including ML configuration information indicating the at least one NNF and the at least one ML model.

[0020] Certain aspects may be implemented in a computer program product implemented on a computer-readable medium. The computer-readable medium may include: code for receiving a first signal from a network entity, the first signal including a first indication of at least one machine learning (ML) capability of a user equipment (UE) and a second indication corresponding to at least one neural network function (NNF) of an ML-based wireless communication management process associated with the UE; code for selecting at least one ML model for performing at least a portion of the ML-based wireless communication management process, at least in part based on the at least one ML capability of the UE; and code for transmitting a second signal to the network entity, the second signal including ML configuration information indicating the at least one NNF and the at least one ML model.

[0021] Other aspects provide: an apparatus capable of operating to, being configured to, or otherwise adapted to perform the foregoing methods and those methods described elsewhere herein; a non-transitory computer-readable medium comprising instructions which, when executed by one or more processors of the apparatus, cause the apparatus to perform the foregoing methods and those methods described elsewhere herein; a computer program product implemented on a computer-readable storage medium including code for performing the foregoing methods and those methods described elsewhere herein; and an apparatus comprising means for performing the foregoing methods and those methods described elsewhere herein. As an example, an apparatus may include a processing system, a device having a processing system, or a processing system cooperating via one or more networks.

[0022] For illustrative purposes, the following description and figures illustrate certain features. Attached Figure Description

[0023] The accompanying drawings depict certain features of the aspects described herein and should not be considered as limiting the scope of this disclosure.

[0024] Figure 1 This is a block diagram that conceptually illustrates an exemplary wireless communication network.

[0025] Figure 2 It is a block diagram that conceptually illustrates various aspects of an exemplary base station and user equipment.

[0026] Figures 3A to 3D Various exemplary aspects of data structures used in wireless communication networks are described.

[0027] Figure 4 An exemplary network environment is shown in which nodes employ predictive machine learning models.

[0028] Figure 5 This is a call flowchart illustrating exemplary operations between a base station and user equipment for configuring and using machine learning models to perform one or more machine learning-based wireless communication management processes.

[0029] Figure 6 This is a flowchart illustrating an exemplary operation of wireless communication performed by a base station.

[0030] Figure 7 This is a flowchart illustrating an exemplary operation of wireless communication performed by a user equipment.

[0031] Figure 8 This is a flowchart illustrating an exemplary operation of wireless communication by a network entity.

[0032] Figures 9 to 10 Various aspects of an exemplary communication device are described.

[0033] Specific implementation method

[0034] This disclosure provides apparatus, methods, processing systems, and computer-readable media for performing neural network functions (NNFs) by configuring and using corresponding machine learning (ML) models to execute one or more ML-based wireless communication management processes.

[0035] For example, wireless communication devices can use machine learning to generate predictive ML models, which can then be used to form inferences about input data. In some cases, one or more ML models can be used to perform NNFs for executing one or more ML-based wireless communication management procedures. In some cases, it can be advantageous for the base station of a wireless communication network to dynamically configure the UE with at least one NNF and / or one or more corresponding ML models. This dynamic configuration can provide flexibility within the wireless communication network. For example, such dynamic configuration can allow the UE to avoid storing all possible ML models for a particular NNF, thus saving storage space on the UE. Instead, the UE can download a specific ML model (e.g., model structure and / or model parameters) individually when instructed to use a particular ML model. In addition, dynamic configuration can provide the base station with the flexibility to selectively choose which / which NNF(s) and / or(s) corresponding models to use to execute one or more ML-based wireless communication management procedures at any given time and for a specific scenario. Furthermore, dynamic configuration can allow the base station to dynamically update the ML models used for NNFs.

[0036] However, there is currently no way for a base station to configure a UE to use a specific NNF and / or ML model (e.g., to perform certain ML-based wireless communication management procedures). Therefore, aspects of this disclosure provide techniques to facilitate the configuration and use of NNFs and corresponding ML models within a wireless communication network.

[0037] Introduction to Wireless Communication Networks

[0038] Figure 1 An example of a wireless communication network 100 in which the aspects described herein can be implemented is depicted.

[0039] Typically, wireless communication network 100 includes base station (BS) 102, user equipment (UE) 104, and one or more core networks (such as evolved packet core (EPC) 160 and fifth generation (5G) core (5GC) network 190), which interoperate to provide wireless communication services.

[0040] Base station 102 can provide user equipment 104 with access to EPC 160 and / or 5GC 190, and can perform one or more of the following functions: transmission of user data, radio channel encryption and decryption, integrity protection, header compression, mobility control functions (e.g., handover, dual connectivity), inter-cell interference coordination, connection establishment and release, load balancing, distribution of non-access stratum (NAS) messages, NAS node selection, synchronization, radio access network (RAN) sharing, multimedia broadcast multicast service (MBMS), subscriber and equipment tracking, RAN information management (RIM), paging, location, and delivery of warning messages, and other functions. In various contexts, a base station may include and / or be referred to as a NodeB, evolved NodeB (eNB), next-generation NodeB (gNB), next-generation eNB (ng-eNB) (e.g., an eNB that has been enhanced to provide connectivity to both EPC 160 and 5GC 190), access point, transceiver base station, radio base station, radio transceiver, transceiver function, or transmit / receive point.

[0041] Base station 102 communicates wirelessly with UE 104 via communication link 120. Each base station 102 can provide communication coverage for various geographical coverage areas 110 that may overlap in some cases. For example, a small cell 102' (e.g., a low-power base station) may have a coverage area 110' that overlaps with a portion of the coverage area 110 of one or more macro cells (e.g., high-power base stations).

[0042] The communication link 120 between base station 102 and UE 104 may include uplink (UL) (also referred to as "reverse link") transmission from user equipment 104 to base station 102 and / or downlink (DL) (also referred to as "forward link") transmission from base station 102 to user equipment 104. In various aspects, the communication link 120 may utilize multiple-input multiple-output (MIMO) antenna technologies, including spatial multiplexing, beamforming, and / or transmit diversity.

[0043] Examples of UE 104 include cellular phones, smartphones, Session Initiation Protocol (SIP) phones, laptops, personal digital assistants (PDAs), satellite radios, Global Positioning System (GPS) devices, multimedia devices, video devices, digital audio players, cameras, game consoles, tablets, smart devices, wearable devices, vehicles, electricity meters, air pumps, large or small kitchen appliances, healthcare devices, implants, sensors / actuators, displays, or other wireless devices. Some UE 104s may be Internet of Things (IoT) devices (e.g., parking meters, air pumps, toasters, vehicles, heart monitors, or other IoT devices), always-on (AON) devices, or edge processing devices. UE 104 may also be more commonly referred to as a station, mobile station, subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handheld device, user agent, mobile client, or client.

[0044] Compared to lower-frequency communication, communication using higher frequency bands may have higher path loss and shorter distances. Therefore, some base stations (e.g., Figure 1 The base station 180 can utilize beamforming 182 with the UE 104 to improve path loss and range. For example, the base station 180 and the UE 104 can each include multiple antennas, such as antenna elements, antenna panels and / or antenna arrays, to facilitate beamforming.

[0045] In some cases, base station 180 may transmit beamformed signals to UE 104 in one or more transmit directions 182'. UE 104 may receive beamformed signals from base station 180 in one or more receive directions 182'. UE 104 may also transmit beamformed signals to base station 180 in one or more transmit directions 182'. Base station 180 may also receive beamformed signals from UE 104 in one or more receive directions 182'. Base station 180 and UE 104 may then perform beam training to determine the optimal receive and transmit directions for each of base station 180 and UE 104. It is worth noting that the transmit and receive directions of base station 180 may be the same or different. Similarly, the transmit and receive directions of UE 104 may be the same or different.

[0046] The wireless communication network 100 (and more specifically, base station 102) includes an ML model configuration component 199, which can be configured to perform... Figures 5 to 6 and Figure 8The document describes one or more operations, as well as other operations for determining the NNF and configuring and using the corresponding ML model to perform one or more ML-based wireless communication management procedures. The wireless communication network 100 (and more specifically, UE 104) also includes an ML model configuration component 198, which can be configured to perform... Figure 5 and Figure 7 The operations shown herein, as well as other operations described herein for determining NNFs and configuring and using corresponding ML models to perform one or more ML-based wireless communication management procedures.

[0047] Figure 2 Various aspects of an exemplary base station (BS) 102 and user equipment (UE) 104 are depicted.

[0048] Typically, base station 102 includes various processors (e.g., 220, 230, 238, and 240), antennas 234a-234t (collectively referred to as 234), transceivers 232a-232t (collectively referred to as 232) including modulators and demodulators (modems), and other aspects that enable the wireless transmission of data (e.g., from data source 212) and the wireless reception of data (e.g., to data sink 239). For example, base station 102 can transmit and receive data between itself and user equipment 104.

[0049] Base station 102 includes a controller / processor 240 that can be configured to implement various functions related to wireless communication. In the depicted example, controller / processor 240 includes components that can represent... Figure 1 The ML model configuration component 199 and the ML model configuration component 241. It is worth noting that although depicted as one aspect of the controller / processor 240, the ML model configuration component 241 may be additionally or alternatively implemented in various other aspects of the base station 102 in other specific implementations.

[0050] Typically, user equipment 104 includes various processors (e.g., 258, 264, 266, and 280), antennas 252a-252r (collectively referred to as 252), transceivers 254a-254r (collectively referred to as 254) including modulators and demodulators, and other aspects that enable wireless transmission of data (e.g., from data source 262) and wireless reception of data (e.g., to data sink 260).

[0051] User equipment 104 includes a controller / processor 280 that can be configured to implement various functions related to wireless communication. In the depicted example, the controller / processor 280 includes components that can represent... Figure 1The ML model configuration component 198 and the ML model configuration component 281. It is worth noting that although depicted as one aspect of the controller / processor 280, the ML model configuration component 281 may be additionally or alternatively implemented in various other aspects of the user equipment 104 in other specific implementations.

[0052] Figures 3A to 3D Describes the use of wireless communication networks (such as Figure 1 The data structure of the wireless communication network 100) in various aspects. Specifically, Figure 3A Figure 300 illustrates an example of the first subframe within a 5G (e.g., 5G NR) frame structure. Figure 3B Figure 330 illustrates an example of a DL channel within a 5G subframe. Figure 3C Figure 350 shows an example of a second subframe within a 5G frame structure, and Figure 3D Figure 380 shows an example of a UL channel within a 5G subframe.

[0053] Information about this disclosure will be provided later in this publication. Figure 1 , Figure 2 and Figures 3A to 3D Further discussion is needed.

[0054] Introduction to Machine Learning

[0055] Machine learning (ML) is generally the process of generating predictive ML models (e.g., artificial neural networks, trees, or other structures) that represent a generalized fit to a training set of data that is known in the prior knowledge. Applying the trained model to new data produces inferences, which can be used to gain insights about the new data.

[0056] Figure 4 An exemplary network environment 400 is shown in which node 420 employs a predictive ML model 424. For example... Figure 4 As shown, network environment 400 includes node 420, training system 430, and training storage 415 communicatively connected via network 405. Node 420 can be a UE (e.g., UE 104 in wireless communication network 100) or BS (e.g., BS 102 in wireless communication network 100). Network 405 can include wireless networks, such as wireless communication network 100, which can be a 5G NR network, a WiFi network, a Long Term Evolution (LTE) network, and / or another type of network. Although training system 430, node 420, and training storage 415 are... Figure 4While shown as separate components, training system 430, node 420, and training storage 415 can be implemented on any number of computing systems (whether as one or more standalone systems or in a distributed environment). Furthermore, although training system 430, node 420, and training storage 415 are shown communicating via network 405, they can also communicate via hardwired connections (or multiple hardwired connections) between training system 430, node 420, and / or training storage 415.

[0057] Training system 430 typically includes a prediction model training manager 432 that uses training data to generate a prediction ML model 424 for predicting output data based on input data. The prediction ML model 424 may be generated at least in part based on information in a training store 415. The training store 415 may include training data acquired before and / or after the deployment of node 420. Node 420 may be trained in a simulated communication environment (e.g., in field testing or drive testing) before deployment. After deployment, the training store 415 may be updated to include feedback associated with the duration of the packet buffer used by node 420. The training store may also be updated using information from other BSs and / or other UEs, for example, based on the experience learned by those BSs and / or UEs.

[0058] The predictive model training manager 432 can use information in the training store 415 to determine the predictive ML model 424 (e.g., an algorithm). The predictive model training manager 432 can use various different types of machine learning algorithms to form the predictive ML model 424. The training system 430 can reside on or at node 420, a BS in network 405, or a different entity that determines the predictive ML model 424. If located on a different entity, the predictive ML model 424 is provided to node 420. The training store 415 can be a storage device, such as a memory. The training store 415 can reside on or at node 420, the training system 430, or another entity in network 405. For example, the training store 415 can be in cloud storage. The training store 415 can receive training information from node 420, entities in network 405 (e.g., the BS and / or UE of network 405), the cloud, or other sources.

[0059] The predictive model training manager 432 can use any suitable machine learning algorithm to form the predictive ML model 424. In some non-limiting examples, the machine learning algorithm is a supervised learning algorithm, a deep learning algorithm, an artificial neural network algorithm, or other types of machine learning algorithm.

[0060] In some examples, deep convolutional networks (DCNs) are used to perform machine learning (e.g., as used by the prediction model training manager 432 in training system 430). A DCN is a network of convolutional networks configured with additional pooling and normalization layers. DCNs have achieved state-of-the-art performance on many tasks. DCNs can be trained using supervised learning, where both the input and output targets are known for many paradigms and are used to modify the network weights using gradient descent. A DCN can be a feedforward network. Furthermore, as mentioned above, the connections from neurons in the first layer of a DCN to the group of neurons in the next higher layer are shared across neurons in the first layer. The feedforward and shared connections of a DCN can be used for fast processing. For example, the computational cost of a DCN may be much smaller than that of a similarly sized neural network containing recurrent or feedback connections.

[0061] In some examples, neural networks are used to perform machine learning (e.g., by the predictive model training manager 432 in training system 430). Neural networks can be designed to have multiple connection patterns. In feedforward networks, information is passed from lower layers to higher layers, where each neuron in a given layer communicates with neurons in higher layers. Hierarchical representations can be built into successive layers of a feedforward network. Neural networks can also have recurrent or feedback (also known as top-down) connections. In recurrent connections, the output from a neuron in a given layer can be passed to another neuron in the same layer. Recurrent architectures can help identify patterns across more than one block of input data that is sequentially passed to the neural network. Connections from neurons in a given layer to neurons in lower layers are called feedback (or top-down) connections. Networks with many feedback connections can be helpful when the recognition of higher-level concepts can aid in discerning specific lower-level features of the input.

[0062] Artificial neural networks, which can be composed of interconnected groups of artificial neurons (e.g., neuron models), are computing devices or representations of methods performed by computing devices. These neural networks can be used in a variety of applications and / or devices, such as Internet Protocol (IP) cameras, Internet of Things (IoT) devices, autonomous vehicles, and / or service robots. Individual nodes in an artificial neural network can mimic biological neurons by taking input data and performing simple operations on that data. The results of the simple operations performed on the input data are selectively passed to other neurons. Weights are associated with each vector and node in the network, and these values ​​constrain how the input data relates to the output data. For example, the input data of each node can be multiplied by its corresponding weight value, and the products can be summed. The sum of the products can be adjusted by optional biases, and an activation function can be applied to the results to produce the node's output signal or "output activation." The weights can initially be determined by an iterative stream of training data through the network (e.g., weights are established during a training phase, in which the network learns how to identify a particular category based on typical input data features of that category).

[0063] Machine learning can be implemented using different types of artificial neural networks (e.g., those used by the prediction model training manager 432 in training system 430), such as recurrent neural networks (RNNs), multilayer perceptron (MLP) neural networks, convolutional neural networks (CNNs), etc. RNNs work on the principle of storing the output of a layer and feeding that output back to the input to help predict the outcome of that layer. In an MLP neural network, data can be fed into the input layer, and one or more hidden layers provide an abstraction level to the data. The output layer can then be used to make predictions based on this abstract data. MLPs are particularly suitable for classification prediction problems where the input is assigned a category or label.

[0064] A Convolutional Neural Network (CNN) is a type of feedforward artificial neural network. A CNN can comprise a collection of artificial neurons, each possessing a receptive field (e.g., a localized region of the input space) that collectively construct an input space. CNNs have numerous applications. In particular, CNNs are widely used in pattern recognition and classification. In a hierarchical neural network architecture, the output of a neuron in the first layer becomes the input of a neuron in the second layer, the output of the neuron in the second layer becomes the input of a neuron in the third layer, and so on. CNNs can be trained to recognize hierarchical structures of features. The computation in a CNN architecture can be distributed across a set of processing nodes, which can be configured in one or more chains of operations. These multi-layered architectures can be trained layer by layer and fine-tuned using backpropagation.

[0065] Various aspects related to machine learning model configuration

[0066] As mentioned above, machine learning can be used to generate predictive ML models that can be used to form inferences about input data. In other words, ML models can be used to predict output data based on data input into the ML model. An ML model can consist of a model structure and a set of parameters. The model structure can include, for example, a graphical structure with vectors or matrices, and the set of parameters can include a set of weights with specific values ​​for the vectors or matrices. In some cases, the set of parameters can be location- or configuration-specific.

[0067] In some cases, ML models can be used to implement wireless communication networks (such as...) Figure 1 The neural network function (NNF) in the wireless communication network 100 shown herein can be configured to execute one or more ML-based wireless communication management procedures within the wireless communication network. In some cases, different NNFs may correspond to different ML-based wireless communication management procedures. For example, these ML-based wireless communication management procedures may include cell reselection procedures, idle or inactive mode measurement procedures, radio resource management (RRM) measurement procedures, etc. Each different NNF can be identified by an NNF identifier (ID) corresponding to a specific ML-based wireless communication management procedure.

[0068] In some cases, different ML models can be used to perform ML-based wireless communication management procedures. In other words, multiple ML models can be associated with a specific NNF and used to perform a specific ML-based wireless communication management procedure. In some cases, the specific ML model to be used to perform the NNF can be indicated via an ML model ID.

[0069] As described above, NNFs can be used within a wireless communication network to perform certain ML-based wireless communication management procedures. In some cases, it can be advantageous for the base station of the wireless communication network to manage these NNFs and for the corresponding ML models to be configured and used at the user equipment within the wireless communication network. However, currently there is no way for the base station to configure the UE to use one or more specific ML models to perform NNFs to execute certain ML-based wireless communication management procedures. Therefore, aspects of this disclosure provide techniques to facilitate the configuration and use of ML models within a wireless communication network to perform certain NNFs.

[0070] An exemplary call flow for determining the NNF and configuring and using the corresponding ML model.

[0071] Figure 5 This is a flowchart illustrating an exemplary operation 500 between BS 502 and UE 504 for configuring and using the NNF and its corresponding ML model to perform one or more ML-based wireless communication management procedures. In some cases, BS 502 may be... Figure 1The example shown is BS102 in the wireless communication network 100. Additionally, UE 504 can be... Figure 1 The example of UE 104 is shown below. Furthermore, in some cases, a Universal Mobile Telecommunications System (UMTS) air interface (Uu interface) can be established to facilitate communication between BS 502 and UE 504; however, in other cases, different types of interfaces may be used.

[0072] As shown in the figure, BS 502 may include multiple logical entities for performing one or more operations in operation 500, such as a centralized unit control plane (CU-CP) 506 for managing the Radio Resource Control (RRC) layer and Packet Data Convergence Protocol (PDCP) layer of BS 502, a centralized unit ML plane (CU-MLP) 508 for managing machine learning functions, and a distributed unit (DU) 510 for managing portions of the Radio Link Control (RLC) layer, Media Access Control (MAC) layer, and Physical (PHY) layer of BS 502. Although CU-MLP 508 is in Figure 5 It is shown as a separate part of BS 502 from CU-CP 506, but CU-MLP 508 may alternatively be implemented as part of CU-CP or as part of a separate network entity from BS 502.

[0073] Operation 500 begins at 520, where UE 504 transmits UE capability information to BS 502, indicating at least one radio capability and at least one ML capability of the UE. In some cases, UE 504 may transmit the UE capability information in an RRC connection establishment message during the Radio Resource Control (RRC) establishment procedure. In some cases, the UE capability information may be received by BS 502's CU-CP 506.

[0074] In some cases, the UE's radio capabilities can indicate the UE's ability to perform one or more radio communication management procedures, which may be ML-based. For example, the UE's radio capabilities can indicate at least one of the following: the ability to perform a (ML-based) cell reselection procedure, the ability to perform a (ML-based) idle or inactive mode measurement procedure, the ability to perform a (ML-based) radio resource management (RRM) measurement procedure, the ability to perform a (ML-based) radio link monitoring (RLM) procedure, the ability to perform a (ML-based) channel state information (CSI) measurement procedure, the ability to perform a (ML-based) precoding matrix indicator (PMI), rank indicator (RI), and channel quality indicator (CQI) feedback procedure, the ability to perform a (ML-based) radio link failure (RLF) and beam failure recovery (BFR) procedure, or the ability to perform a (ML-based) RRM relaxation procedure.

[0075] In some cases, at least one ML capability of a UE may indicate one or more capabilities supported by the UE for performing ML. For example, the ML capability of a UE may indicate at least one of the following: ML training capability, ML inference capability, processing capability, one or more supported ML model formats, one or more supported ML libraries, or indication of one or more locally cached ML models.

[0076] At 525, BS 502's CU-CP 506 determines whether to use ML functions to perform one or more wireless communication management procedures. For example, in some cases, BS 502's CU-CP 506 can select the ML-based wireless communication management procedure to be used at the UE and determine the at least one NNF used to perform at least a portion of the selected ML-based wireless communication management procedure. In other words, at 525, BS 502 can determine which NNFs the UE supports (if any) and can generate a requested list of NNFs.

[0077] Subsequently, at 530, CU-CP 506 sends a context establishment request message to CU-MLP 508 to establish a context for UE 504 to use the at least one NNF. The context establishment request message includes an indication of the at least one NNF determined by CU-CP 506 (e.g., a list of requested NNFs) and a request to CU-MLP 508 for support of using the at least one NNF. Although Figure 5 The use of a context establishment request message for requesting CU-MLP 508 support using at least one of the NNFs is illustrated, but it should be understood that other or additional signaling procedures may be used.

[0078] The context establishment request message may include at least one ML capability of the UE. The CU-MLP 508 can then select at least one ML model for use in at least one NNF to perform at least a portion of an ML-based wireless communication management process. In some cases, the CU-MLP 508 may select the at least one ML model based at least in part on the at least one ML capability of the UE. Additionally, in some cases, the CU-MLP 508 may select the at least one ML model based on at least one of the cell ID, gNB ID, or UE context information. In some cases, the UE context information may indicate information such as UE type, data radio bearer (DRB) configuration, and / or antenna switching (AS) configuration.

[0079] Subsequently, at 545, CU-MLP 508 sends a context establishment response message to CU-CP 506. The context establishment response message can provide an indication that a UE context has been successfully established for at least one NNF indicated in the context establishment request message (e.g., a list of accepted NNFs). Additionally, the context establishment response message can provide an indication of at least one ML model selected for at least one NNF to perform at least a portion of the ML-based wireless communication management process. In some cases, the indication of at least one ML model may include an ML model ID.

[0080] Subsequently, as shown at 550, BS 502 transmits ML configuration information to UE 504 based on the UE capability information received at 520 (e.g., via CU-CP 506). In some cases, BS 502 may transmit the ML configuration information in an RRC reconfiguration message. The ML configuration information may include at least one NNF (e.g., a list of accepted NNFs) and an indication of at least one ML model corresponding to the at least one NNF. In some cases, the at least one NNF is indicated by an NNF ID, and the at least one ML model is indicated by an ML model ID. As described above, the at least one ML model may be associated with a model structure and one or more sets of parameters (e.g., weights, biases, and / or activation functions). In some cases, the ML model ID may indicate the model structure associated with the at least one ML model, while the one or more sets of parameters may be indicated by a parameter set ID in the ML configuration information.

[0081] In some cases, ML models and ML model structures can be associated with multiple sets of parameters (e.g., one or more sets of parameters may include multiple parameter groups). In this case, each set of parameters can be effectively used with a model structure specific to a particular geographic region (such as a cell) or a particular configuration. In other words, different sets of parameters can be used with a model structure depending on, for example, the specific geographic region or configuration of UE 504. This allows UE 504 to use a model structure to perform ML-based wireless communication management procedures while adaptively changing the set of parameters used with that model structure according to the specific geographic region or configuration of UE 504.

[0082] Subsequently, as shown at 555, in response to receiving an RRC reconfiguration message including ML configuration information, UE 504 sends an RRC reconfiguration complete message to BS 502, which can indicate that UE 504 has successfully received the ML configuration information.

[0083] As shown at 560, if at least one ML model corresponding to at least one NNF is not already locally stored at UE 504, then UE 504 may then download the at least one ML model. For example, in some cases, if the ML model is not locally stored by UE 504, then UE 504 may receive from BS 502 at least one of the ML model structure corresponding to the at least one ML model or at least one set of parameters (e.g., weights, biases, and / or activation functions) associated with the ML model structure.

[0084] Subsequently, as shown at 565, once UE 504 has received the ML structure and / or one or more parameters associated with at least one ML model, UE 504 transmits a message to BS 502 indicating that at least one ML model is ready for use by UE 504. In some cases, this message may be received by CU-CP 506 and subsequently forwarded to CU-MLP 508, as shown at 570.

[0085] Subsequently, at some future point in time, BS 502 may determine that UE 504 should use at least one NNF and at least one corresponding ML model to perform at least a portion of the ML-based wireless communication management process. In this case, BS 502 transmits a signal to UE 504 to activate the use of at least one ML model to perform at least a portion of the ML-based wireless communication management process. In some cases, the signal activating the use of at least one ML model may be transmitted by BS 502 (e.g., via CU-CP 506 or CU-MLP 508) in the Media Access Control-Control Element (MAC-CE) or in RRC signaling.

[0086] Subsequently, UE 504 can perform ML-based wireless communication management procedures using at least one ML model based on the activation signal. In some cases, performing ML-based wireless communication management procedures may include: inputting one or more input variables into at least one ML model; and obtaining an output from at least one ML model based on the one or more input variables.

[0087] As described above, the at least one ML model can be associated with a model structure and one or more sets of parameters. When performing an ML-based wireless communication management procedure, UE 504 can determine a specific set of parameters to be used in conjunction with the ML structure to process one or more input variables. This determination can be based on, for example, a specific geographic area (such as a cell) or configuration of UE 504. For example, in some cases, when UE 504 is in a first cell, UE 504 can select a first set of parameters to use with the model structure to perform the ML-based wireless communication management procedure. In other cases, when UE 504 is in a second cell, UE 504 can select a second set of parameters to use with the model structure to perform the ML-based wireless communication management procedure.

[0088] For example, an ML-based wireless communication management process may include cell reselection and idle / inactive mode measurement procedures. In this case, to perform the cell reselection and idle / inactive mode measurement procedures, UE 504 may input one or more input variables into a model structure of an ML model specifically trained for cell reselection and idle / inactive mode measurements. As mentioned above, UE 504 may also select a set of parameters (e.g., weights, biases, and / or activation functions) to be used with the model structure. In some cases, the input variables of the ML model may include, for example, serving cell measurements (e.g., Reference Signal Received Power (RSRP) measurements associated with BS 502), neighboring cell measurements, and the service specified by UE 504. The at least one ML model may take into account the serving cell measurements, neighboring cell measurements, and the service specified by UE 504, and provide an output indicating the target cell to be reselected and / or the target cell to be used for idle / inactive mode measurements.

[0089] Additional considerations regarding the use of NNF and corresponding ML models in both base stations and user equipment.

[0090] In some cases, NNFs and corresponding ML models can be used in both BS 502 and UE 504. For example, as described above, BS 502 can configure UE 504 with a specific NNF and a corresponding first ML model to perform at least a portion of one or more ML-based wireless communication management procedures. However, BS 502 can also configure itself with a second ML model to perform at least a portion of the ML-based wireless communication management procedures. In some cases, the first ML model and the second ML model can include matching ML models, whereby the output of one ML model is used as the input of the other ML model.

[0091] For example, return to Figure 5 After establishing the context associated with at least one NNF for UE 504 and selecting at least one ML model (e.g., a first ML model) corresponding to at least one NNF, CU-MLP 508 may additionally determine a second ML model for BS 502 to perform at least a portion of the ML-based wireless communication management process. In some cases, CU-MLP 508 may determine another ML model for BS 502 based on ML capability information associated with DU 510. As shown at 535, CU-MLP 508 may then send an ML model establishment request message to DU 510, requesting DU 510 to establish the second ML model for performing at least a portion of the ML-based wireless communication management process. Subsequently, once the second ML model has been established, DU 510 sends an ML model establishment response message to CU-MLP 508 at 540, indicating that the establishment of the second ML model has been completed.

[0092] As stated above, in some cases, the first ML model configured for use by UE 504 can be matched with the second ML model configured for use by BS 502. In other words, the output of the first ML model used by UE 504 can be used as the input of the second ML model used by BS 502.

[0093] For example, as described above, BS 502 can transmit a signal activating the use of a first ML model (e.g., at least one ML model, as described above) for performing an ML-based wireless communication management process. BS 502 can also activate the use of a second ML model at BS 502. Thereafter, UE 504 can perform the ML-based wireless communication management process by inputting one or more input variables into the first ML model and obtaining an output from the first ML model based on said one or more input variables. UE 504 can then transmit the output of the first ML model, which can be received by BS 502. BS can then input the output received from UE into the second ML model at BS 502 and obtain the output of the second ML model at BS.

[0094] An exemplary use case of this matching ML model configuration between the first and second ML models can be used to perform functions such as data compression for reporting Channel State Information (CSI) feedback. For example, in wireless communication networks employing a large number of antennas, such as in high-volume Multiple-Input Multiple-Output (MIMO) scenarios, the CSI feedback can be very large. In this case, when UE 504 transmits CSI feedback to BS 502, UE 504 can use the first ML model to compress the CSI feedback to save time and frequency resources on the radio channel. Once BS 502 receives the CSI feedback, BS 502 can use the second ML model to decompress the CSI feedback.

[0095] Exemplary methods for determining NNFs and configuring and using the corresponding ML models

[0096] Figure 6 This is a flowchart illustrating an exemplary operation 600 for wireless communication. Operation 600 can be performed, for example, by a BS (e.g., Figure 1 The operation 600 is executed by one or more processors (e.g., BS102 in the wireless communication network 100) and / or another network entity (e.g., a device implementing CU-CP 506 or CU-MLP508) to determine the NNF and configure and use the corresponding ML model to perform one or more ML-based wireless communication management procedures. Operation 600 can be implemented in one or more processors (e.g., Figure 2 Software components executed and running on the controller / processor 240. Furthermore, signal transmission and reception performed by the BS (and / or other network entities) in operation 600 can be, for example, by one or more transceivers and antennas (e.g., Figure 2 This is achieved via transceiver 232 and antenna 234. In some aspects, signal transmission and / or reception by the BS (and / or other network entities) can be achieved by obtaining and / or outputting signals via a bus interface of one or more processors (e.g., controller / processor 240, including ML model configuration component 241).

[0097] Operation 600 begins at box 610, wherein UE capability information indicating at least one radio capability and at least one machine learning (ML) capability of the UE is received from the user equipment (UE).

[0098] In box 620, the BS or other network entity transmits to the UE, based on the UE capability information, at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF, ML configuration information.

[0099] In some cases, in the ML configuration information, at least one NNF is indicated by an NNF identifier (ID), and at least one ML model is indicated by an ML model ID.

[0100] In some cases, at least one NNF includes at least a portion of an ML-based wireless communication management process, and the ML model is configured to perform at least a portion of the ML-based wireless communication management process.

[0101] In some cases, ML-based wireless communication management processes include at least one of the following: cell reselection process, idle or inactive mode measurement process, radio resource management (RRM) measurement process, radio link monitoring (RLM) process, channel state information (CSI) measurement process, precoding matrix indicator (PMI), rank indicator (RI) and channel quality indicator (CQI) feedback process, radio link failure (RLF) and beam failure recovery (BFR) process, or power saving process.

[0102] In some cases, operation 600 may also include the BS or other network entity selecting an ML-based wireless communication management procedure to be used at the UE. In some cases, operation 600 may also include the BS or other network entity determining at least one NNF corresponding to the selected ML-based wireless communication management procedure. In some cases, operation 600 may also include the BS or other network entity selecting at least one ML model for performing at least a portion of the ML-based wireless communication management procedure, based at least in part on at least one of the UE's radio capabilities or the UE's ML capabilities.

[0103] In some cases, Operation 600 may also include the BS or other network entity establishing a context for the UE based at least in part on at least one NNF, at least one ML model, and at least one ML capability of the UE.

[0104] In some cases, at least one ML model used to perform at least a portion of an ML-based wireless communication management process includes an ML model structure and a set of or more sets of parameters associated with the ML model structure. In some cases, the set of or more sets of parameters includes a set of weights, which may be indicated by a parameter set ID in the ML configuration information. In some cases, operation 600 may also include the BS or other network entity transmitting to the UE at least one of the ML model structure or the set of or more sets of parameters associated with the ML model structure.

[0105] In some cases, operation 600 may also include the BS or other network entity receiving from the UE a signal indicating that at least one ML model is ready for use after transmitting ML configuration information. In some cases, operation 600 may also include the BS or other network entity transmitting to the UE a signal activating the use of at least one ML model to perform at least a portion of an ML-based wireless communication management process. In some cases, transmitting the signal activating the use of at least one ML model includes transmitting the signal activating the use of at least one ML model in a Media Access Control (MAC-CE) element or in Radio Resource Control (RRC) signaling.

[0106] In some cases, operation 600 may also include activating the use of a second ML model at the BS or other network entity, receiving the UE output of the at least one ML model from the UE, inputting the UE output to the second ML model at the BS or other network entity, and obtaining the output of the second ML model at the BS or other network entity.

[0107] In some cases, receiving UE capability information includes receiving UE capability information in the Radio Resource Control (RRC) connection establishment message. Additionally, in some cases, transmitting ML configuration information includes transmitting ML configuration information in the Radio Resource Control (RRC) reconfiguration message.

[0108] In some cases, at least one radio capability of the UE includes at least one of the following: the ability to perform ML-based cell reselection procedures, the ability to perform ML-based idle or inactive mode measurement procedures, the ability to perform ML-based radio resource management (RRM) measurement procedures, the ability to perform ML-based radio link monitoring (RLM) procedures, the ability to perform ML-based channel state information (CSI) measurement procedures, the ability to perform ML-based precoding matrix indicator (PMI), rank indicator (RI), and channel quality indicator (CQI) feedback procedures, the ability to perform ML-based radio link failure (RLF) and beam failure recovery (BFR) procedures, or the ability to perform ML-based RRM relaxation procedures.

[0109] In some cases, at least one ML capability of a UE includes at least one of the following: ML training capability, ML inference capability, processing capability, one or more supported ML model formats, one or more supported ML libraries, or indication of one or more locally cached ML models.

[0110] Figure 7 This is a flowchart illustrating an exemplary operation 700 for wireless communication according to certain aspects of this disclosure. Operation 700 can be performed, for example, by a UE (e.g., Figure 1 The UE 104 in the wireless communication network 100 performs one or more ML-based wireless communication management procedures to determine the NNF and configure and use the corresponding ML model. Operation 700 may be complementary to operation 600 performed by the BS and / or other network entities. Operation 700 may be implemented in one or more processors (e.g., Figure 2 The software components executed and running on the controller / processor 280. Furthermore, the signal transmission and reception performed by the UE in operation 600 can be, for example, by one or more transceivers and antennas (e.g., Figure 2 This is achieved via transceiver 254 and antenna 252. In some aspects, signal transmission and / or reception by the UE can be achieved by obtaining and / or outputting signals via a bus interface of one or more processors (e.g., controller / processor 280, including ML model configuration component 281).

[0111] Operation 700 begins at box 710, wherein UE capability information indicating at least one radio capability and at least one machine learning (ML) capability of the UE is transmitted to the base station (BS) and / or other network entities.

[0112] In box 720, the UE receives ML configuration information from the BS or other network entity based on UE capability information, indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0113] In some cases, the at least one NNF is indicated by an NNF identifier (ID) in the ML configuration information, and the at least one ML model is indicated by an ML model ID. Additionally, in some cases, the at least one NNF includes at least a portion of an ML-based wireless communication management process, and the ML model is configured to perform at least a portion of the ML-based wireless communication management process.

[0114] In some cases, ML-based wireless communication management processes include at least one of the following: cell reselection process, idle or inactive mode measurement process, radio resource management (RRM) measurement process, radio link monitoring (RLM) process, channel state information (CSI) measurement process, precoding matrix indicator (PMI), rank indicator (RI) and channel quality indicator (CQI) feedback process, radio link failure (RLF) and beam failure recovery (BFR) process, or power saving process.

[0115] In some cases, at least one ML model used to perform ML-based wireless communication management procedures includes an ML model structure and a set of or more sets of parameters associated with the ML model structure. In some cases, the set of or more sets of parameters includes a set of weights, which may be indicated by a parameter set ID in the ML configuration information. In some cases, operation 700 also includes the UE receiving an indication of at least one (or more) of the ML model structure or the set of or more sets of parameters associated with the ML model structure from the BS or other network entity.

[0116] In some cases, operation 700 also includes the UE transmitting a signal to the BS or other network entity indicating that at least one ML model is ready for use after receiving an indication of the ML model structure or at least one of a set or more sets of parameters.

[0117] In some cases, operation 700 also includes the UE receiving from the BS or other network entity a signal activating the use of at least one ML model to perform at least a portion of an ML-based wireless communication management process. In some cases, receiving the signal activating the use of at least one ML model includes receiving the signal activating the use of at least one ML model in a Media Access Control Element (MAC-CE) or in Radio Resource Control (RRC) signaling.

[0118] In some cases, Operation 700 also includes the UE using at least one ML model to perform an ML-based wireless communication management procedure based on activation signaling. In some cases, performing an ML-based wireless communication management procedure includes: inputting one or more input variables into at least one ML model; obtaining an output from at least one ML model based on one or more input variables; and transmitting the output of at least one ML model to a BS or other network entity.

[0119] In some cases, transmitting UE capability information includes transmitting UE capability information during the Radio Resource Control (RRC) connection establishment process. In some cases, receiving ML configuration information includes receiving ML configuration information in a Radio Resource Control (RRC) reconfiguration message.

[0120] In some cases, at least one radio capability of the UE includes at least one of the following: the ability to perform ML-based cell reselection procedures, the ability to perform ML-based idle or inactive mode measurement procedures, the ability to perform ML-based radio resource management (RRM) measurement procedures, the ability to perform ML-based radio link monitoring (RLM) procedures, the ability to perform ML-based channel state information (CSI) measurement procedures, the ability to perform ML-based precoding matrix indicator (PMI), rank indicator (RI), and channel quality indicator (CQI) feedback procedures, the ability to perform ML-based radio link failure (RLF) and beam failure recovery (BFR) procedures, or the ability to perform ML-based RRM relaxation procedures.

[0121] In some cases, at least one ML capability of a UE includes at least one of the following: ML training capability, ML inference capability, processing capability, one or more supported ML model formats, one or more supported ML libraries, or indication of one or more locally cached ML models.

[0122] Figure 8 This is a flowchart illustrating an exemplary operation 800 for wireless communication. Operation 800 may be performed, for example, by a first network entity (e.g., a device implementing CU-MLP 508) to determine the Non-Non-Functional (NNF) and to configure and use a corresponding ML model to perform one or more ML-based wireless communication management procedures. Operation 800 may be implemented as a software component that executes and runs on one or more processors. Furthermore, the signal transmission and reception performed by the network entity in operation 800 may be implemented, for example, by one or more transceivers and antennas. In some aspects, the signal transmission and / or reception performed by the network entity may be implemented by obtaining and / or outputting signals via a bus interface of one or more processors.

[0123] Operation 800 begins in block 810, wherein a first signal is received from a second network entity, the first signal including a first indication of at least one machine learning (ML) capability of a user equipment (UE) and a second indication of at least one neural network function (NNF) corresponding to an ML-based wireless communication management process associated with the UE.

[0124] In block 820, the first network entity selects at least one ML model for performing at least a portion of the ML-based wireless communication management process, based at least in part on the at least one ML capability of the UE.

[0125] In box 830, the first network entity transmits a second signal to the second network entity, the second signal including ML configuration information indicating at least one NNF and at least one ML model.

[0126] In some cases, in the ML configuration information: at least one NNF is indicated by an NNF identifier (ID), and at least one ML model is indicated by an ML model ID.

[0127] In some cases, the at least one ML model includes an ML model structure and one or more sets of parameters associated with the ML model structure.

[0128] In some cases, operation 800 also includes determining the ML model structure and one or more sets of parameters associated with the ML model structure.

[0129] In some cases, one or more sets of parameters include at least one set of weights, and one or more parameters are indicated by a parameter set ID in the ML configuration information.

[0130] In some cases, operation 800 also includes transmitting a third signal to a third network entity, the third signal including the at least one NNF and an indication of a second ML model corresponding to the at least one NNF for performing at least another part of an ML-based wireless communication management process.

[0131] In some cases, the first network entity includes a Centralized Unit Machine Learning Plane (CU-MLP) entity, the second network entity includes a Centralized Unit Control Plane (CU-CP) entity, and the third network entity includes a Distributed Unit (DU) entity. In some cases, the CU-CP entity is part of a base station, and the CU-MLP entity is at least part of an entity separate from the base station.

[0132] In some cases, operation 800 also includes receiving a third signal forwarded from the UE by a second network entity, the third signal indicating that at least one ML model is ready for use.

[0133] In some cases, Operation 800 also includes establishing a context for the UE against at least one NNF. In this case, the first signal includes a context establishment request message, and the second signal includes a context establishment response message.

[0134] In some cases, at least one ML capability of a UE includes at least one of the following: ML training capability, ML inference capability, processing capability, one or more supported ML model formats, one or more supported ML libraries, or indication of one or more locally cached ML models.

[0135] In some cases, operation 800 also includes transmitting a third signal to the UE to activate the use of at least one ML model to perform at least a portion of an ML-based wireless communication management process. In some cases, transmitting the third signal to activate the use of at least one ML model includes transmitting the third signal to activate the use of at least one ML model in a Media Access Control Element (MAC-CE) or in Radio Resource Control (RRC) signaling.

[0136] Exemplary wireless communication device

[0137] Figure 9 The description includes operations that can be operated to, configured to, or adapted to perform the techniques disclosed herein (such as regarding...). Figures 5 to 6 and Figure 8 An exemplary communication device 900 comprising various components (depicted and described in the text). In some examples, the communication device 900 may be, for example, about... Figure 1 and Figure 2 The described base station 102, or other network entities such as CU-CP 506 or CU-MLP 508 (e.g., when implemented as a standalone device).

[0138] The communication device 900 includes a processing system 902 coupled to a transceiver 908 (e.g., a transmitter and / or receiver). The transceiver 908 is configured to transmit (or send) and receive signals from the communication device 900, such as the various signals described herein, via an antenna 910. The processing system 902 may be configured to perform processing functions for the communication device 900, including processing signals received by and / or to be transmitted by the communication device 900.

[0139] Processing system 902 includes one or more processors 920 coupled to computer-readable medium / memory 930 via bus 906. In some aspects, the computer-readable medium / memory 930 is configured to store instructions (e.g., computer-executable code) that, when executed by the one or more processors 920, cause the one or more processors 920 to perform. Figures 5 to 6 and Figure 8 The operations shown herein, or other operations used to perform the various techniques discussed herein for determining NNFs and configuring and using the corresponding ML models to perform one or more ML-based wireless communication management procedures.

[0140] In the illustrated example, the computer-readable medium / memory 930 stores code 931 for receiving, code 932 for transmitting, code 933 for selecting, code 934 for determining, code 935 for establishing, code 936 for activating, code 837 for input, and code 838 for obtaining.

[0141] In the depicted example, the one or more processors 920 include circuitry configured to implement code stored in a computer-readable medium / memory 930, including circuitry 921 for receiving, circuitry 922 for transmitting, circuitry 923 for selecting, circuitry 924 for determining, circuitry 925 for establishing, circuitry 926 for activating, circuitry 827 for input, and circuitry 828 for obtaining.

[0142] The various components of the communication device 900 can provide for performing the methods described herein (including those described herein). Figures 5 to 6 and Figure 8 The apparatus described in the method.

[0143] In some examples, the means for transmitting or sending (or the means for outputting for transmission) may include in Figure 2 The base station 102 shown includes a transceiver 232 and / or (multiple) antennas 234 and / or Figure 9 The transceiver 908 and antenna 910 of the communication device 900.

[0144] In some examples, the means for receiving (or the means for obtaining) may include in Figure 2 The base station shown includes a transceiver 232 and / or (multiple) antennas 234 and / or Figure 9 The transceiver 908 and antenna 910 of the communication device 900.

[0145] In some examples, the means for selection, the means for determination, the means for establishment, the means for activation, the means for input, and the means for acquisition can include various processing system components, such as: Figure 9 One or more of the 920 processors, or Figure 2 The base station 102 depicted includes various aspects such as a receiver processor 238, a transmitter processor 220, a TX MIMO processor 230, and / or a controller / processor 240 (including an ML model configuration component 241).

[0146] It is worth noting that, Figure 9 This is just an example, and many other examples and configurations of the communication device 900 are possible.

[0147] Figure 10 The description includes operations that can be operated to, configured to, or adapted to perform the techniques disclosed herein (such as regarding...). Figure 5 and Figure 7 An exemplary communication device 1000 comprising various components (described and depicted in the operation). In some examples, the communication device 1000 may be, for example, about... Figure 1 and Figure 2User equipment 104 as described.

[0148] Communication device 1000 includes a processing system 1002 coupled to a transceiver 1008 (e.g., a transmitter and / or receiver). Transceiver 1008 is configured to transmit (or send) and receive signals from communication device 1000, such as the various signals described herein, via antenna 1010. Processing system 1002 may be configured to perform processing functions for communication device 1000, including processing signals received by and / or to be transmitted by communication device 1000.

[0149] Processing system 1002 includes one or more processors 1020 coupled to computer-readable medium / memory 1030 via bus 1006. In some aspects, computer-readable medium / memory 1030 is configured to store instructions (e.g., computer-executable code) that, when executed by the one or more processors 1020, cause the one or more processors 1020 to perform. Figure 5 and Figure 7 The operations shown herein, or other operations used to perform the various techniques discussed herein for determining NNFs and configuring and using the corresponding ML models to perform one or more ML-based wireless communication management procedures.

[0150] In the depicted example, computer-readable medium / memory 1030 stores code 1031 for transmission, code 1032 for reception, and code 1033 for execution.

[0151] In the depicted example, the one or more processors 1020 include circuitry configured to implement code stored in a computer-readable medium / memory 1030, including circuitry 1021 for transmitting, circuitry 1022 for receiving, and circuitry 1023 for executing.

[0152] The various components of the communication device 1000 can provide for performing the methods described herein (including those described herein). Figure 5 and Figure 7 The apparatus described in the method.

[0153] In some examples, the means for transmitting or sending (or the means for outputting for transmission) may include in Figure 2 The user equipment 104 shown includes a transceiver 254 and / or (multiple) antennas 252 and / or Figure 10 The transceiver 1008 and antenna 1010 of the communication device 1000.

[0154] In some examples, the means for receiving (or the means for obtaining) may include in Figure 2The user equipment 104 shown includes a transceiver 254 and / or (multiple) antennas 252 and / or Figure 10 The transceiver 1008 and antenna 1010 of the communication device 1000.

[0155] In some examples, the means of execution may include various processing system components, such as: Figure 10 One or more processors 1020, or Figure 2 The user equipment 104 depicted includes various aspects such as a receiver processor 258, a transmitter processor 264, a TX MIMO processor 266, and / or a controller / processor 280 (including an ML model configuration component 281).

[0156] It is worth noting that, Figure 10 This is just an example, and many other examples and configurations of the communication device 1000 are possible.

[0157] Exemplary terms

[0158] Specific implementation examples are described in the following numbered clauses:

[0159] Clause 1: A method for wireless communication by a user equipment (UE), comprising: transmitting UE capability information to a base station (BS) or other network entity indicating at least one radio capability of the UE and at least one machine learning (ML) capability of the UE; and receiving, based on the UE capability information, ML configuration information from the BS or other network entity indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0160] Clause 2: The method according to Clause 1, wherein in the ML configuration information: the at least one NNF is indicated by an NNF identifier (ID), and the at least one ML model is indicated by an ML model ID.

[0161] Clause 3: The method according to any one of Clauses 1 to 2, wherein: the at least one NNF includes at least a portion of an ML-based wireless communication management process, and the ML model is configured to perform the at least a portion of the ML-based wireless communication management process.

[0162] Clause 4: The method described in Clause 3, wherein the ML-based wireless communication management process includes at least one of the following: cell reselection process, idle or inactive mode measurement process, radio resource management (RRM) measurement process, radio link monitoring (RLM) process, channel state information (CSI) measurement process, precoding matrix indicator (PMI), rank indicator (RI) and channel quality indicator (CQI) feedback process, radio link failure (RLF) and beam failure recovery (BFR) process, or power saving process.

[0163] Clause 5: The method according to any one of Clauses 3 to 4, wherein the at least one ML model for performing the ML-based wireless communication management process includes an ML model structure and one or more parameters associated with the ML model structure.

[0164] Clause 6: According to the method described in Clause 5, the one or more parameters include a set of weights, and the one or more parameters are indicated by a parameter set ID in the ML configuration information.

[0165] Clause 7: The method according to any one of Clauses 5 to 6 further includes receiving, from the BS or other network entity, at least one of the ML model structure or the one or more parameters associated with the ML model structure.

[0166] Clause 8: The method according to Clause 7 further comprises: after receiving the ML model structure or at least one of the one or more parameters, transmitting a signal to the BS or other network entity indicating that the at least one ML model is ready for use.

[0167] Clause 9: The method according to any one of Clauses 3 to 8 further includes receiving from the BS or other network entity a signal activating the use of the at least one ML model to perform at least a portion of the ML-based wireless communication management process.

[0168] Clause 10: The method according to Clause 9 further includes using the at least one ML model based on the activation signal to perform the ML-based wireless communication management process, wherein performing the ML-based wireless communication management process includes: inputting one or more input variables into the at least one ML model; obtaining an output from the at least one ML model based on the one or more input variables; and transmitting the output of the at least one ML model to the BS or other network entity.

[0169] Clause 11: The method according to any one of Clauses 9 to 10, wherein receiving the signal for activating the at least one ML model includes receiving the signal for activating the at least one ML model in a Media Access Control Control Element (MAC-CE) or in Radio Resource Control (RRC) signaling.

[0170] Clause 12: The method according to any one of Clauses 1 to 11, wherein transmitting the UE capability information includes transmitting the UE capability information during the Radio Resource Control (RRC) connection establishment process.

[0171] Clause 13: The method according to any one of Clauses 1 to 12, wherein receiving the ML configuration information includes receiving the ML configuration information in a Radio Resource Control (RRC) reconfiguration message.

[0172] Clause 14: The method according to any one of Clauses 1 to 13, wherein the at least one radio capability of the UE includes at least one of the following: the ability to perform a cell reselection procedure based on ML, the ability to perform an idle or inactive mode measurement procedure based on ML, the ability to perform a radio resource management (RRM) measurement procedure based on ML, the ability to perform a radio link monitoring (RLM) procedure based on ML, the ability to perform a channel state information (CSI) measurement procedure based on ML, the ability to perform a precoding matrix indicator (PMI), rank indicator (RI), and channel quality indicator (CQI) feedback procedure based on ML, the ability to perform a radio link failure (RLF) and beam failure recovery (BFR) procedure based on ML, or the ability to perform an RRM relaxation procedure based on ML.

[0173] Clause 15: The method according to any one of Clauses 1 to 14, wherein the at least one ML capability of the UE includes at least one of the following: ML training capability, ML inference capability, processing capability, one or more supported ML model formats, one or more supported ML libraries, or indication of one or more locally cached ML models.

[0174] Clause 16: A method for wireless communication by a base station (BS) or other network entity, comprising: receiving from a user equipment (UE) UE capability information indicating at least one radio capability of the UE and at least one machine learning (ML) capability of the UE; and transmitting to the UE ML configuration information based on the UE capability information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF.

[0175] Clause 17: The method according to Clause 16, wherein in the ML configuration information: the at least one NNF is indicated by an NNF identifier (ID), and the at least one ML model is indicated by an ML model ID.

[0176] Clause 18: The method according to any one of Clauses 16 to 17, wherein: the at least one NNF includes at least a portion of an ML-based wireless communication management process, and the ML model is configured to perform the at least a portion of the ML-based wireless communication management process.

[0177] Clause 19: The method described in Clause 18, wherein the ML-based wireless communication management process includes at least one of the following: cell reselection process, idle or inactive mode measurement process, radio resource management (RRM) measurement process, radio link monitoring (RLM) process, channel state information (CSI) measurement process, precoding matrix indicator (PMI), rank indicator (RI) and channel quality indicator (CQI) feedback process, radio link failure (RLF) and beam failure recovery (BFR) process, or power saving process.

[0178] Clause 20: The method according to any one of Clauses 18 to 19 further includes: selecting the ML-based wireless communication management procedure to be used at the UE; determining the at least one NNF corresponding to the selected ML-based wireless communication management procedure; and selecting the at least one ML model for performing the at least part of the ML-based wireless communication management procedure based at least in part on the at least one of the at least one radio capability of the UE or the at least one ML capability of the UE.

[0179] Clause 21: The method according to Clause 20 further includes establishing a context for the UE based at least in part on the at least one NNF, the at least one ML model, and the at least one ML capability of the UE.

[0180] Clause 22: The method according to any one of Clauses 18 to 21, wherein the at least one ML model for performing at least a portion of the ML-based wireless communication management process includes an ML model structure and one or more parameters associated with the ML model structure.

[0181] Clause 23: The method according to Clause 22, wherein the one or more parameters comprise a set of weights, and the one or more parameters are indicated by a parameter set ID in the ML configuration information.

[0182] Clause 24: The method according to Clause 22 further includes transmitting to the UE at least one of the ML model structure or one or more parameters associated with the ML model structure.

[0183] Clause 25: The method according to Clause 24 further includes receiving from the UE a signal indicating that the at least one ML model is ready for use after transmitting the ML configuration information.

[0184] Clause 26: The method according to any one of Clauses 18 to 25 further includes transmitting to the UE a signal activating the use of the at least one ML model to perform at least a portion of the ML-based wireless communication management process.

[0185] Clause 27: The method according to Clause 26 further includes: activating the use of a second ML model at the BS or other network entity; receiving UE output of the at least one ML model from the UE; inputting the UE output to the second ML model at the BS or other network entity; and obtaining the output of the second ML model at the BS or other network entity.

[0186] Clause 28: The method according to any one of Clauses 26 to 27, wherein transmitting the signal used to activate the at least one ML model comprises transmitting the signal used to activate the at least one ML model in a Media Access Control Control Element (MAC-CE) or in Radio Resource Control (RRC) signaling.

[0187] Clause 29: The method according to any one of Clauses 16 to 28, wherein receiving the UE capability information includes receiving the UE capability information in a Radio Resource Control (RRC) Connection Establishment message.

[0188] Clause 30: The method according to any one of Clauses 16 to 29, wherein transmitting the ML configuration information includes transmitting the ML configuration information in a Radio Resource Control (RRC) reconfiguration message.

[0189] Clause 31: The method according to any one of Clauses 16 to 30, wherein the at least one radio capability of the UE includes at least one of the following: the ability to perform a cell reselection procedure based on ML, the ability to perform an idle or inactive mode measurement procedure based on ML, the ability to perform a radio resource management (RRM) measurement procedure based on ML, the ability to perform a radio link monitoring (RLM) procedure based on ML, the ability to perform a channel state information (CSI) measurement procedure based on ML, the ability to perform a precoding matrix indicator (PMI), rank indicator (RI), and channel quality indicator (CQI) feedback procedure based on ML, the ability to perform a radio link failure (RLF) and beam failure recovery (BFR) procedure based on ML, or the ability to perform an RRM relaxation procedure based on ML.

[0190] Clause 32: The method according to any one of Clauses 16 to 31, wherein the at least one ML capability of the UE includes at least one of the following: ML training capability, ML inference capability, processing capability, one or more supported ML model formats, one or more supported ML libraries, or indication of one or more locally cached ML models.

[0191] Clause 33: A method for wireless communication by a first network entity, comprising: receiving a first signal from a second network entity, the first signal including a first indication of at least one machine learning (ML) capability of a user equipment (UE) and a second indication corresponding to at least one neural network function (NNF) of an ML-based wireless communication management process associated with the UE; selecting at least one ML model for performing at least a portion of the ML-based wireless communication management process based at least in part on the at least one ML capability of the UE; and transmitting a second signal to the second network entity, the second signal including ML configuration information indicating the at least one NNF and the at least one ML model.

[0192] Clause 34: The method according to Clause 33, wherein in the ML configuration information: the at least one NNF is indicated by an NNF identifier (ID), and the at least one ML model is indicated by an ML model ID.

[0193] Clause 35: The method according to any one of Clauses 33 to 34, wherein the at least one ML model includes an ML model structure and a set or more sets of parameters associated with the ML model structure.

[0194] Clause 36: The method according to Clause 35 further includes determining the ML model structure and the set or more sets of parameters associated with the ML model structure.

[0195] Clause 37: The method according to any one of Clauses 35 to 36, wherein: the set or more sets of parameters includes at least one set of weights, and the one or more parameters are indicated by a parameter set ID in the ML configuration information.

[0196] Clause 38: The method according to any one of Clauses 33 to 37 further includes transmitting a third signal to a third network entity, the third signal including the at least one NNF and an indication of a second ML model corresponding to the at least one NNF for performing at least another part of the ML-based wireless communication management process.

[0197] Clause 39: The method according to Clause 38, wherein: the first network entity includes a Centralized Unit Machine Learning Plane (CU-MLP) entity, the second network entity includes a Centralized Unit Control Plane (CU-CP) entity, and the third network entity includes a Distributed Unit (DU) entity.

[0198] Clause 40: The method according to Clause 39, wherein the CU-CP entity is part of the base station and the CU-MLP entity is at least part of an entity separate from the base station.

[0199] Clause 41: The method according to any one of Clauses 33 to 39 further includes receiving a third signal forwarded from the UE by the second network entity, the third signal indicating that the at least one ML model is ready for use.

[0200] Clause 42: The method according to any one of Clauses 33 to 41 further includes establishing a context for the UE for the at least one NNF.

[0201] Clause 43: The method according to Clause 42, wherein the first signal includes a context establishment request message and the second signal includes a context establishment response message.

[0202] Clause 44: The method according to any one of Clauses 33 to 43, wherein the at least one ML capability of the UE includes at least one of the following: ML training capability, ML inference capability, processing capability, one or more supported ML model formats, one or more supported ML libraries, or indication of one or more locally cached ML models.

[0203] Clause 45: The method according to any one of Clauses 33 to 44 further includes transmitting to the UE a third signal activating the use of the at least one ML model to perform at least a portion of the ML-based wireless communication management process.

[0204] Clause 46: The method according to Clause 45, wherein transmitting the third signal used to activate the at least one ML model includes transmitting the third signal used to activate the at least one ML model in a Media Access Control Control Element (MAC-CE) or in Radio Resource Control (RRC) signaling.

[0205] Clause 47: An apparatus comprising: a memory including executable instructions; and one or more processors configured to execute the executable instructions and cause the apparatus to perform a method according to any one of Clauses 1 to 46.

[0206] Clause 48: An apparatus comprising means for performing the method according to any one of Clauses 1 to 46.

[0207] Clause 49: A non-transitory computer-readable medium comprising: executable instructions that, when executed by one or more processors of the device, cause the device to perform the method according to any one of Clauses 1 to 46.

[0208] Clause 50: A computer program product implemented on a computer-readable storage medium, comprising code for performing a method according to any one of Clauses 1 to 46.

[0209] Additional wireless communication network considerations

[0210] The techniques and methods described herein can be used in a variety of wireless communication networks (or wireless wide area networks (WWANs)) and radio access technologies (RATs). While aspects may be described herein using terms commonly associated with 3G, 4G, and / or 5G (e.g., 5G New Radio (NR)) wireless technologies, aspects of this disclosure are equally applicable to other communication systems and standards not explicitly mentioned herein.

[0211] 5G wireless communication networks can support a variety of advanced wireless communication services, such as enhanced mobile broadband (eMBB), millimeter wave (mmWave), machine-type communication (MTC), and / or ultra-reliable low-latency communication (URLLC) for mission-critical applications. These and other services may include latency and reliability specifications.

[0212] return Figure 1 Various aspects of this disclosure can be implemented within an exemplary wireless communication network 100.

[0213] In 3GPP, the term "cell" can refer to the coverage area of ​​a NodeB and / or the narrowband subsystem serving that coverage area, depending on the context in which the term is used. In NR systems, the term "cell" is used interchangeably with BS, next-generation NodeB (gNB or gNodeB), access point (AP), distributed unit (DU), carrier, or transmit / receive point. A BS can provide communication coverage for macrocells, picocells, femtocells, and / or other cell types.

[0214] Macro cells typically cover a relatively large geographic area (e.g., a radius of several kilometers) and allow unrestricted access by UEs with service subscriptions. Pico cells cover a relatively small geographic area (e.g., a stadium) and allow unrestricted access by UEs with service subscriptions. Femto cells cover a relatively small geographic area (e.g., a home) and allow restricted access by UEs associated with that femto cell (e.g., UEs in a closed subscriber group (CSG) and UEs belonging to users in the home). A BS used for a macro cell may be referred to as a macro BS. A BS used for a pico cell may be referred to as a pico BS. A BS used for a femto cell may be referred to as a femto BS, a home BS, or a home NodeB.

[0215] Base station 102 configured for 4G LTE (collectively referred to as Evolved Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access Network (E-UTRAN)) can interface with EPC 160 via a first backhaul link 132 (e.g., S1 interface). Base station 102 configured for 5G (e.g., 5G NR or Next Generation RAN (NG-RAN)) can interface with 5GC 190 via a second backhaul link 184. Base station 102 can communicate with each other directly or indirectly (e.g., via EPC 160 or 5GC 190) on a third backhaul link 134 (e.g., X2 interface). The third backhaul link 134 can typically be wired or wireless.

[0216] Small cell 102' can operate in licensed and / or unlicensed spectrum. When operating in unlicensed spectrum, small cell 102' can employ NR and use the same 5 GHz unlicensed spectrum as Wi-Fi AP 150. Small cell 102' employing NR in unlicensed spectrum can improve coverage of the access network and / or increase the capacity of the access network.

[0217] Some base stations (such as gNBs) can operate in conventional sub-6 GHz spectrum, millimeter wave (mmWave) frequencies, and / or near-mmWave frequencies to communicate with UE 104. When a gNB operates in mmWave or near-mmWave frequencies, it can be referred to as an mmWave base station.

[0218] The communication link 120 between base station 102 and, for example, UE 104, can be transmitted via one or more carriers. For example, for each carrier allocated in a carrier aggregation of up to Y x MHz (x component carriers) for transmission in each direction, base station 102 and UE 104 can use a spectrum with a bandwidth of up to Y MHz (e.g., 5 MHz, 10 MHz, 15 MHz, 20 MHz, 100 MHz, 400 MHz, and other MHz). Carriers may or may not be adjacent to each other. Carrier allocation can be asymmetric with respect to DL and UL (e.g., more or fewer carriers may be allocated to DL compared to UL). Component carriers may include primary component carriers and one or more secondary component carriers. The primary component carrier may be referred to as the primary cell (PCell) and the secondary component carrier may be referred to as the secondary cell (SCell).

[0219] The wireless communication network 100 also includes a Wi-Fi access point (AP) 150 that communicates with a Wi-Fi station (STA) 152 via a communication link 154 in, for example, unlicensed spectrum of 2.4 GHz and / or 5 GHz. When communicating in unlicensed spectrum, STA 152 and AP 150 can perform a free channel assessment (CCA) to determine whether a channel is available before communication.

[0220] Some UEs 104 may communicate with each other using a device-to-device (D2D) communication link 158. The D2D communication link 158 may use DL / UL WWAN spectrum. The D2D communication link 158 may use one or more sidelink channels, such as the Physical Sidelink Broadcast Channel (PSBCH), Physical Sidelink Discovery Channel (PSDCH), Physical Sidelink Shared Channel (PSSCH), and Physical Sidelink Control Channel (PSCCH). D2D communication may be conducted via a variety of wireless D2D communication systems, such as, for example, FlashLinQ, WiMedia, Bluetooth, ZigBee, Wi-Fi based on the IEEE 802.11 standard, 4G (e.g., LTE), or 5G (e.g., NR), to name just a few options.

[0221] EPC 160 may include a Mobility Management Entity (MME) 162, other MMEs 164, a Serving Gateway 166, a Multimedia Broadcast Multicast Service (MBMS) Gateway 168, a Broadcast Multicast Service Center (BM-SC) 170, and a Packet Data Network (PDN) Gateway 172. MME 162 may communicate with the Home Subscriber Server (HSS) 174. MME 162 is the control node that handles signaling between UE 104 and EPC 160. Generally, MME 162 provides bearer and connection management.

[0222] Generally, user Internet Protocol (IP) packets are transmitted through Serving Gateway 166, which is itself connected to PDN Gateway 172. PDN Gateway 172 provides UE IP address allocation and other functions. PDN Gateway 172 and BM-SC 170 are connected to IP service 176, which may include, for example, the Internet, intranet, IP Multimedia Subsystem (IMS), PS streaming service, and / or other IP services.

[0223] The BM-SC 170 provides functionality for MBMS user service provisioning and delivery. The BM-SC 170 can serve as an entry point for MBMS transmissions by content providers, authorize and initiate MBMS bearer services within a Public Land Mobile Network (PLMN), and schedule MBMS transmissions. The MBMS gateway 168 can distribute MBMS services to base stations 102 belonging to Multicast-Broadcast Single Frequency Network (MBSFN) areas belonging to Broadcast Specific Services, and is responsible for session management (start / stop) and collecting eMBMS-related billing information.

[0224] 5GC 190 may include Access and Mobility Management Function (AMF) 192, other AMFs 193, Session Management Function (SMF) 194, and User Plane Function (UPF) 195. AMF 192 may communicate with Unified Data Management Function (UDM) 196.

[0225] The AMF 192 is typically the control node that handles signaling between UE 104 and 5GC 190. Generally, the AMF 192 provides QoS flow and session management.

[0226] All user Internet Protocol (IP) packets are transmitted via UPF 195, which connects to IP service 197 and provides the UE with IP address allocation and other functions for 5GC 190. IP service 197 may include, for example, the Internet, intranet, IP Multimedia Subsystem (IMS), PS streaming service, and / or other IP services.

[0227] return Figure 2 The document describes BS102 and UE 104 (e.g., which can be used to implement various aspects of this disclosure). Figure 1 Various exemplary components of the wireless communication network 100.

[0228] At BS102, the transmit processor 220 can receive data from the data source 212 and control information from the controller / processor 240. This control information may be for the Physical Broadcast Channel (PBCH), Physical Control Format Indicator Channel (PCFICH), Physical Hybrid ARQ Indicator Channel (PHICH), Physical Downlink Control Channel (PDCCH), Group Common PDCCH (GC PDCCH), and others. In some examples, this data may be for the Physical Downlink Shared Channel (PDSCH).

[0229] The Media Access Control (MAC)-Control Element (MAC-CE) is a MAC layer communication structure that can be used for the exchange of control commands between wireless nodes. The MAC-CE can be carried in a shared channel, such as the Physical Downlink Shared Channel (PDSCH), Physical Uplink Shared Channel (PUSCH), or Physical Sidelink Shared Channel (PSSCH).

[0230] Processor 220 can process (e.g., encoding and symbol mapping) data and control information to obtain data symbols and control symbols, respectively. Transmit processor 220 can also generate reference symbols (such as those for primary synchronization signal (PSS), secondary synchronization signal (SSS), PBCH demodulation reference signal (DMRS), and channel state information reference signal (CSI-RS)).

[0231] The transmit (TX) multiple-input multiple-output (MIMO) processor 230 can perform spatial processing (e.g., precoding) on ​​data symbols, control symbols, and / or reference symbols, where applicable, and can provide the output symbol stream to the modulators (MODs) in transceivers 232a-232t. Each modulator in transceivers 232a-232t can process its own output symbol stream (e.g., for OFDM) to obtain an output sample stream. Each modulator can further process (e.g., convert to analog, amplify, filter, and up-convert) the output sample stream to obtain a downlink signal. The downlink signal from the modulators in transceivers 232a-232t can be transmitted via antennas 234a-234t respectively.

[0232] At UE 104, antennas 252a-252r can receive downlink signals from BS 102 and can provide the received signals to demodulators (DEMODs) in transceivers 254a-254r respectively. Each demodulator in transceivers 254a-254r can adjust (e.g., filter, amplify, down-convert, and digitize) the corresponding received signal to obtain an input sample. Each demodulator can further process the input sample (e.g., for OFDM) to obtain the received symbols.

[0233] MIMO detector 256 acquires received symbols from all demodulators in transceivers 254a-254r, performs MIMO detection on the received symbols where applicable, and provides the detected symbols. Receiver processor 258 can process (e.g., demodulate, deinterleave, and decode) the detected symbols, provide decoded data from UE 104 to data sink 260, and provide decoded control information to controller / processor 280.

[0234] On the uplink, at UE 104, transmit processor 264 can receive and process data from data source 262 (e.g., for the Physical Uplink Shared Channel (PUSCH)) and control information from controller / processor 280 (e.g., for the Physical Uplink Control Channel (PUCCH)). Transmit processor 264 can also generate reference symbols for reference signals (e.g., sounding reference signals (SRS)). Symbols from transmit processor 264 may be pre-encoded by TX MIMO processor 266, where applicable, further processed by modulators in transceivers 254a-254r (e.g., for SC-FDM), and transmitted to BS 102.

[0235] At BS102, uplink signals from UE 104 can be received by antennas 234a-t, processed by demodulators in transceivers 232a-232t, detected by MIMO detector 236 where applicable, and further processed by receiver processor 238 to obtain decoded data and control information transmitted by UE 104. Receiver processor 238 can provide the decoded data to data sink 239 and the decoded control information to controller / processor 240.

[0236] Memory 242 and 282 can store data and program code for BS102 and UE 104, respectively.

[0237] Scheduler 244 can schedule UE to perform data transmission on the downlink and / or uplink.

[0238] 5G can utilize Orthogonal Frequency Division Multiplexing (OFDM) with a cyclic prefix (CP) on both the uplink and downlink. 5G can also support half-duplex operation using Time Division Duplex (TDD). OFDM and Single-Carrier Frequency Division Multiplexing (SC-FDM) divide the system bandwidth into multiple orthogonal subcarriers, often referred to as frequency modulation and frequency grids. Each subcarrier can be modulated with data. Modulation symbols can be transmitted in the frequency domain using OFDM and in the time domain using SC-FDM. The spacing between adjacent subcarriers can be fixed, and the total number of subcarriers can depend on the system bandwidth. In some examples, the minimum resource allocation (called a resource block (RB)) can be 12 consecutive subcarriers. The system bandwidth can also be divided into subbands. For example, a subband can cover multiple RBs. NR can support a base subcarrier spacing (SCS) of 15 kHz and can define other SCSs (e.g., 30 kHz, 60 kHz, 120 kHz, 240 kHz, and others) relative to the base SCS.

[0239] As mentioned above, Figures 3A to 3D Describes the use of wireless communication networks (such as Figure 1 Various exemplary aspects of the data structure of the wireless communication network 100.

[0240] In various aspects, the 5G NR frame structure can be Frequency Division Duplex (FDD), where subframes within a specific set of subcarriers (carrier system bandwidth) are dedicated to either DL (dedicated deep learning) or UL (ultimate learning). The 5G frame structure can also be Time Division Duplex (TDD), where subframes within a specific set of subcarriers (carrier system bandwidth) are dedicated to both DL and UL. Figure 3A and Figure 3C In the provided example, the 5G frame structure is assumed to be TDD, where subframe 4 is configured with slot format 28 (mostly DL) and subframe 3 is configured with slot format 34 (mostly UL), where D is DL, U is UL, and X is for flexible use between DL and UL. Although subframes 3 and 4 are shown with slot formats 34 and 28 respectively, any particular subframe can be configured with any of the various available slot formats 0-61. Slot formats 0 and 1 are DL and UL, respectively. Other slot formats 2-61 include a mixture of DL, UL, and flexible symbols. The UE is configured with a slot format via the received Slot Format Indicator (SFI) (dynamically configured via DL Control Information (DCI) or semi-statically / statically configured via Radio Resource Control (RRC) signaling). Note that the following description also applies to 5G frame structures for TDD.

[0241] Other wireless communication technologies may have different frame structures and / or different channels. A frame (10 ms) can be divided into 10 subframes (1 ms) of the same size. Each subframe may include one or more time slots. Subframes may also include micro-time slots, which may include 7, 4, or 2 symbols. In some examples, each time slot may include 7 or 14 symbols, depending on the time slot configuration.

[0242] For example, for slot configuration 0, each slot may include 14 symbols, while for slot configuration 1, each slot may include 7 symbols. Symbols on the DL can be Cyclic Prefix (CP) OFDM (CP-OFDM) symbols. Symbols on the UL can be CP-OFDM symbols (for high-throughput scenarios) or Discrete Fourier Transform (DFT) Extended OFDM (DFT-s-OFDM) symbols (also known as Single Carrier Frequency Division Multiple Access (SC-FDMA) symbols) (for power-constrained scenarios; limited to a single stream).

[0243] The number of time slots within a subframe is based on the time slot configuration and the digital scheme. For time slot configuration 0, different digital schemes (μ) 0 to 5 allow for 1, 2, 4, 8, 16, and 32 time slots per subframe, respectively. For time slot configuration 1, different digital schemes 0 to 2 allow for 2, 4, and 8 time slots per subframe, respectively. Therefore, for time slot configuration 0 and digital scheme μ, there are 14 symbols / time slots and 2μ time slots / subframe. The subcarrier spacing and symbol length / duration vary depending on the digital scheme. The subcarrier spacing can be equal to 2. μ ×15kHz, where μ is the digital scheme from 0 to 5. Thus, digital scheme μ=0 has a subcarrier spacing of 15kHz, and digital scheme μ=5 has a subcarrier spacing of 480kHz. The symbol length / duration is inversely related to the subcarrier spacing. Figures 3A to 3D Examples are provided for slot configuration 0 with 14 symbols per slot and digital scheme μ=2 with 4 slots per subframe. The slot duration is 0.25 ms, the subcarrier spacing is 60 kHz, and the symbol duration is approximately 16.67 μs.

[0244] A resource grid can be used to represent the frame structure. Each time slot consists of a resource block (RB) extending for 12 consecutive subcarriers (also known as a physical RB (PRB)). The resource grid is divided into multiple resource elements (REs). The number of bits carried by each RE depends on the modulation scheme.

[0245] like Figure 3A As shown, some REs carry information for the UE (e.g., Figure 1 and Figure 2The reference (pilot) signal (RS) for the UE (104) may include a demodulation RS (DM-RS) (denoted as Rx for a particular configuration, where 100x is the port number, but other DM-RS configurations are also possible) and a channel state information reference signal (CSI-RS) for channel estimation at the UE. The RS may also include a beam measurement RS (BRS), a beam refinement RS (BRRS), and a phase tracking RS (PT-RS).

[0246] Figure 3B Examples of various DL channels within a subframe of a frame are shown. The Physical Downlink Control Channel (PDCCH) carries the DCI within one or more Control Channel Elements (CCEs), each CCE comprising nine RE groups (REGs), each REG comprising four consecutive REs in an OFDM symbol.

[0247] The Primary Synchronization Signal (PSS) can be located within symbol 2 of a specific subframe of the frame. The PSS is generated by the UE (e.g., Figure 1 and Figure 2 104) is used to determine subframe / symbol timing and physical layer identifier.

[0248] The secondary synchronization signal (SSS) is located in symbol 4 of a specific subframe of the frame. The SSS is used by the UE to determine the physical layer cell identifier group number and radio frame timing.

[0249] Based on the Physical Layer Identifier and Physical Layer Cell Identifier Group Number, the UE can determine the Physical Cell Identifier (PCI). Based on the PCI, the UE can determine the location of the aforementioned DM-RS. The Physical Broadcast Channel (PBCH), carrying the Master Information Block (MIB), can be logically grouped with the PSS and SSS to form a Synchronization Signal (SS) / PBCH block. The MIB provides the number of RBs in the system bandwidth and the System Frame Number (SFN). The Physical Downlink Shared Channel (PDSCH) carries user data, broadcast system information not transmitted via the PBCH (e.g., System Information Block (SIB)), and paging messages.

[0250] like Figure 3CAs shown, some REs carry DM-RS for channel estimation at the base station (indicated as R for a specific configuration, but other DM-RS configurations are possible). The UE can transmit DM-RS for the Physical Uplink Control Channel (PUCCH) and DM-RS for the Physical Uplink Shared Channel (PUSCH). The PUSCH DM-RS can be transmitted in the first or second symbol of the PUSCH. PUCCH DM-RS can be transmitted in different configurations depending on whether a short or long PUCCH is transmitted and the specific PUCCH format used. The UE can transmit a Sounding Reference Signal (SRS). The SRS can be transmitted in the last symbol of a subframe. The SRS can have a comb structure, and the UE can transmit the SRS on one of the combs. The SRS can be used by the base station for channel quality estimation to enable frequency-dependent scheduling of the UL.

[0251] Figure 3D Examples of various UL channels within a subframe of a frame are shown. The PUCCH can be located as indicated in one configuration. The PUCCH carries uplink control information (UCI), such as scheduling requests, channel quality indicators (CQI), precoding matrix indicators (PMI), rank indicators (RI), and HARQ ACK / NACK feedback. The PUSCH carries data and can additionally be used to carry buffer status reports (BSR), power headroom reports (PHR), and / or UCI.

[0252] Additional considerations

[0253] The foregoing description provides examples of configuring and using ML models to perform one or more NNFs (such as ML-based wireless communication management processes in communication systems). The foregoing description is provided to enable any person skilled in the art to implement the various aspects described herein. The examples discussed herein do not limit the scope, applicability, or aspects set forth in the claims. Various modifications to these aspects will be apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects. For example, changes may be made to the function and arrangement of the elements discussed without departing from the scope of this disclosure. Various processes or components may be omitted, substituted, or added as appropriate in the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described for some examples may be combined in some other examples. For example, an apparatus or method of practice may be implemented using any number of aspects set forth herein. Additionally, the scope of this disclosure is intended to cover such apparatus or methods implemented using other structures, functions, or structures and functions that complement or replace the various aspects of this disclosure set forth herein. It should be understood that any aspect of the disclosure herein may be embodied by one or more elements of the claims.

[0254] The techniques described in this article can be used in various wireless communication technologies, such as 5G (e.g., 5G NR), 3GPP Long Term Evolution (LTE), LTE-A Advanced (LTE-A), Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single Carrier Frequency Division Multiple Access (SC-FDMA), Time Division Synchronous Code Division Multiple Access (TD-SCDMA), and other networks. The terms "network" and "system" are often used interchangeably. CDMA networks can implement radio technologies such as Universal Terrestrial Radio Access (UTRA), cdma2000, and others. UTRA includes Wideband CDMA (WCDMA) and other variants of CDMA. cdma2000 covers the IS-2000, IS-95, and IS-856 standards. TDMA networks can implement radio technologies such as Global System for Mobile Communications (GSM). OFDMA networks can implement radio technologies such as NR (e.g., 5G RA), evolved UTRA (E-UTRA), Ultra Mobile Broadband (UMB), IEEE 802.11 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802.20, Flash-OFDMA, etc. UTRA and E-UTRA are parts of the Universal Mobile Telecommunications System (UMTS). LTE and LTE-A are UMTS versions using E-UTRA. UTRA, E-UTRA, UMTS, LTE, LTE-A, and GSM are described in documents from an organization called the 3rd Generation Partnership Project (3GPP). CDMA2000 and UMB are described in documents from an organization called the 3rd Generation Partnership Project 2 (3GPP2). NR is an emerging wireless communication technology under development.

[0255] The various illustrative logic blocks, modules, and circuits described in connection with this disclosure can be implemented or executed using a general-purpose processor, DSP, ASIC, field-programmable gate array (FPGA) or other programmable logic device (PLD), discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in alternatives, the processor may be any commercially available processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors working in conjunction with a DSP core, a system-on-a-chip (SoC), or any other such configuration.

[0256] If implemented in hardware, an exemplary hardware configuration could include a processing system in a wireless node. This processing system can be implemented using a bus architecture. Depending on the specific application and overall design constraints of the processing system, the bus can include any number of interconnect buses and bridges. The bus can link various circuits together, including processors, machine-readable media, and bus interfaces. The bus interface can be used to connect network adapters, etc., to the processing system via the bus. The network adapter can be used to implement signal processing functions at the PHY layer. In user equipment (see...) Figure 1 In such cases, the user interface (e.g., keypad, display, mouse, joystick, touchscreen, biometric sensor, proximity sensor, light-emitting element, etc.) can also be connected to the bus. The bus can also link various other circuits, such as timing sources, peripheral devices, voltage regulators, power management circuits, etc., which are well known in the art and will not be described further. The processor can be implemented using one or more general-purpose processors and / or special-purpose processors. Examples include microprocessors, microcontrollers, DSP processors, and other circuits capable of executing software. Those skilled in the art will recognize that the aforementioned functions of the processing system depend on how best to achieve them depending on the specific application and the overall design constraints imposed on the system.

[0257] If implemented in software, the functions can be stored or transmitted as one or more instructions or code on or through a computer-readable medium. Software should be broadly interpreted as representing instructions, data, or any combination thereof, whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise. Computer-readable media includes both computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one location to another. The processor may be responsible for managing the bus and general-purpose processing, which includes executing software modules stored on the machine-readable storage medium. The computer-readable storage medium may be coupled to the processor, allowing the processor to read information from and write information to the storage medium. Alternatively, the storage medium may be integrated with the processor. For example, the machine-readable medium may include a transmission line, a carrier wave modulated by data, and / or a separate computer-readable storage medium containing instructions, all accessible to the processor via a bus interface. Alternatively or additionally, the machine-readable medium or any portion thereof may be integrated into the processor, for example, with a cache and / or a general-purpose register file. As an example, examples of machine-readable storage media may include RAM (random access memory), flash memory, ROM (read-only memory), PROM (programmable read-only memory), EPROM (erasable programmable read-only memory), EEPROM (electrically erasable programmable read-only memory), registers, disks, optical disks, hard disks, or any other suitable storage media, or any combination thereof. Machine-readable media may be implemented in computer program products.

[0258] Software modules may comprise a single instruction or a number of instructions, and may be distributed across several different code segments, across different programs, and across multiple storage media. Computer-readable media may include several software modules. A software module includes instructions that, when executed by a device such as a processor, cause the processing system to perform various functions. Software modules may include transfer modules and receive modules. Each software module may reside in a single storage device or be distributed across multiple storage devices. For example, when a triggering event occurs, a software module may be loaded from a hard disk drive into RAM. During the execution of a software module, the processor may load some of the instructions into a cache to improve access speed. One or more cache lines may then be loaded into a general-purpose register file for processor execution. When the functionality of a software module is referred to below, it will be understood that such functionality is implemented by the processor when executing the instructions from that software module.

[0259] As used herein, the term “exemplary” means “serving as an example, instance, or illustration.” Any aspect described herein as “exemplary” is not necessarily to be construed as preferred or superior to other aspects.

[0260] As used herein, the phrase “at least one of” in a list of entries refers to any combination of those entries (including a single member). For example, “at least one of a, b, or c” is intended to cover a, b, c, ab, ac, bc, and abc, as well as any combination with multiple identical elements (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or any other ordering of a, b, and c).

[0261] As used herein, the term "determine" encompasses a wide variety of actions. For example, "determine" can include calculation, computation, processing, derivation, research, searching (e.g., looking in a table, database, or other data structure), ascertaining, and similar actions. Furthermore, "determine" can include receiving (e.g., receiving information), accessing (e.g., accessing data in memory), etc. Additionally, "determine" can include parsing, selecting, choosing, building, and so on.

[0262] The methods disclosed herein include one or more steps or actions for implementing the methods. The steps and / or actions of the methods may be interchanged without departing from the scope of the claims. That is, unless a specific order of steps or actions is specified, the order and / or use of specific steps and / or actions may be modified without departing from the scope of the claims. Furthermore, the various operations of the methods described above can be performed by any suitable means capable of performing the corresponding functions. Such means may include various hardware and / or (multiple) software components and / or (multiple) modules, including but not limited to circuits, application-specific integrated circuits (ASICs), or processors. Typically, where operations shown in the figures are present, those operations may have corresponding paired means plus functional components with similar numbering.

[0263] The following claims are not intended to be limited to the aspects shown herein, but should be granted the full scope consistent with the language of the claims. Within the claims, unless expressly stated otherwise, references to elements in the singular form are not intended to mean “one and only one,” but rather “one or more.” Unless otherwise specifically stated, the term “some” refers to one or more. No element of any claim should be interpreted in accordance with 35 U.S.SC §112(f) unless the element is expressly referred to by the phrase “means for…” or, in the case of a method claim, by the phrase “steps for…”. All structural and functional equivalents of the elements of the aspects described throughout this disclosure that are now or hereafter known to a person skilled in the art are expressly incorporated herein and are intended to be covered by the claims. Furthermore, nothing disclosed herein is intended to be offered to the public, whether or not such disclosure is expressly referred to in the claims.

Claims

1. An apparatus for wireless communication by a user equipment (UE), comprising: Includes one or more memories containing executable instructions; and One or more processors are configured to execute the executable instructions and enable the UE: Transmit UE capability information to network entities, indicating at least one radio capability of the UE and at least one machine learning (ML) capability of the UE; Based on the UE capability information, the network entity receives ML configuration information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF, wherein the at least one NNF includes at least one ML-based wireless communication management procedure, and the at least one ML model is configured to perform the at least one ML-based wireless communication management procedure, wherein: the at least one ML-based wireless communication management procedure includes a channel state information (CSI) measurement procedure; and the at least one ML model includes at least one ML model for performing CSI feedback compression for the CSI measurement procedure; as well as The UE receives an activation signal from the network entity to activate the use of the at least one ML model to perform the ML-based wireless communication management process.

2. The apparatus according to claim 1, wherein in the ML configuration information: The at least one NNF is indicated by the NNF identifier ID, and The at least one ML model is indicated by an ML model ID.

3. The apparatus of claim 1, wherein the ML-based wireless communication management process further comprises at least one of the following: The process of re-election of neighborhoods, Measurement process in idle or inactive mode. Radio Resource Management (RRM) measurement process, Radio link monitoring (RLM) process, Feedback process for precoding matrix indicator PMI, rank indicator RI, and channel quality indicator CQI. Radio link failure (RLF) and beam failure recovery (BFR) procedures, or Power saving process.

4. The apparatus of claim 1, wherein the at least one ML model for performing the ML-based wireless communication management process comprises an ML model structure and one or more sets of parameters associated with the ML model structure.

5. The apparatus according to claim 4, wherein: The set or more sets of parameters include at least one set of weights, and The set or multiple sets of parameters are indicated by parameter set IDs in the ML configuration information.

6. The apparatus according to claim 4, wherein, The one or more processors are further configured to enable the UE to receive from the network entity the ML model structure or at least one of the one or more sets of parameters associated with the ML model structure.

7. The apparatus according to claim 6, wherein, The one or more processors are further configured to cause the UE to transmit a signal to the network entity indicating that the at least one ML model is ready for use after receiving the ML model structure or at least one of the set or more sets of parameters.

8. The apparatus according to claim 1, wherein, The one or more processors are further configured to cause the UE to perform the ML-based wireless communication management process using the at least one ML model based on the activation signal, wherein, in order to perform the ML-based wireless communication management process, the one or more processors are further configured to cause the UE to: Input one or more input variables into the at least one ML model; as well as Obtain the output based on the one or more input variables from the at least one ML model.

9. The apparatus according to claim 1, wherein, The one or more processors are also configured to enable the UE to receive the activation signal in the Media Access Control (MAC-CE) control element or in the Radio Resource Control (RRC) signaling.

10. The apparatus according to claim 1, wherein, The one or more processors are also configured to cause the UE to transmit UE capability information during the Radio Resource Control (RRC) connection establishment process.

11. The apparatus according to claim 1, wherein, The one or more processors are further configured to enable the UE to receive the ML configuration information, including receiving the ML configuration information in a Radio Resource Control (RRC) reconfiguration message.

12. The apparatus of claim 1, wherein the at least one radio capability of the UE comprises at least one of the following: The ability to perform ML-based cell reselection processes. The ability to perform ML-based measurement procedures in idle or inactive modes. The ability to perform ML-based Radio Resource Management (RRM) measurement procedures. The ability to perform ML-based radio link monitoring (RLM) procedures. The ability to perform ML-based Channel State Information (CSI) measurement procedures. The ability to perform ML-based feedback processes for precoding matrix indicator (PMI), rank indicator (RI), and channel quality indicator (CQI). The ability to perform ML-based radio link failure (RLF) and beam failure recovery (BFR) procedures, or The ability to perform ML-based RRM relaxation processes.

13. The apparatus of claim 1, wherein the at least one ML capability of the UE comprises at least one of the following: ML training capability ML inference capabilities Processing capacity One or more supported ML model formats, One or more supported ML libraries, or Instructions for one or more local cached ML models.

14. An apparatus for wireless communication by a network entity, comprising: Includes one or more memories containing executable instructions; and One or more processors are configured to execute the executable instructions and enable the network entity: Receive UE capability information from the user equipment (UE) indicating at least one radio capability and at least one machine learning (ML) capability of the UE; Based on the UE capability information, the system transmits to the UE at least one neural network function (NNF) and ML configuration information corresponding to at least one ML model for the at least one NNF. The at least one NNF includes at least one ML-based wireless communication management procedure, and the at least one ML model is configured to execute the at least one ML-based wireless communication management procedure, wherein: the at least one ML-based wireless communication management procedure includes a channel state information (CSI) measurement procedure; and the at least one ML model includes at least one ML model for performing CSI feedback compression for the CSI measurement procedure; and The network entity transmits an activation signal to the UE to activate the use of at least one ML model to perform the ML-based wireless communication management process.

15. The apparatus of claim 14, wherein in the ML configuration information: The at least one NNF is indicated by the NNF identifier ID, and The at least one ML model is indicated by an ML model ID.

16. The apparatus of claim 14, wherein the ML-based wireless communication management process further comprises at least one of the following: The process of re-election of neighborhoods, Measurement process in idle or inactive mode. Radio Resource Management (RRM) measurement process, Radio link monitoring (RLM) process, Feedback process for precoding matrix indicator PMI, rank indicator RI, and channel quality indicator CQI. Radio link failure (RLF) and beam failure recovery (BFR) procedures, or Power saving process.

17. The apparatus according to claim 14, wherein, The one or more processors are also configured to cause the network entity to: Select the ML-based wireless communication management procedure to be used at the UE; Determine at least one NNF corresponding to the selected ML-based wireless communication management procedure; as well as The at least one ML model for performing the ML-based wireless communication management process is selected based at least in part on at least one of the at least one radio capability of the UE or at least one ML capability of the UE.

18. The apparatus according to claim 17, wherein, The one or more processors are also configured to enable the network entity to establish a context for the UE based at least in part on the at least one NNF, the at least one ML model, and the at least one ML capability of the UE.

19. The apparatus according to claim 14, wherein: The at least one ML model used to perform the ML-based wireless communication management process includes an ML model structure and one or more sets of parameters associated with the ML model structure. The set or more sets of parameters include a set of weights, and The set or multiple sets of parameters are indicated by parameter set IDs in the ML configuration information.

20. The apparatus according to claim 19, wherein, The one or more processors are also configured to cause the network entity to: Transmit the ML model structure or at least one of the set or more sets of parameters associated with the ML model structure to the UE; as well as After transmitting the ML configuration information, the UE receives a signal indicating that at least one ML model is ready for use.

21. The apparatus according to claim 14, wherein, The one or more processors are further configured to cause the network entity to transmit an activation signal to the UE in a Media Access Control (MAC-CE) control element or in Radio Resource Control (RRC) signaling to activate the use of the at least one ML model to perform the ML-based wireless communication management process.

22. The apparatus according to claim 14, wherein, The one or more processors are also configured to cause the network entity to: Activate the use of the second ML model at the network entity; Receive UE output of at least one ML model from the UE; The UE output is input into the second ML model at the network entity; as well as Obtain the output of the second ML model at the network entity.

23. The apparatus according to claim 14, wherein, The one or more processors are also configured to cause the network entity to: The UE capability information is received in the Radio Resource Control (RRC) connection establishment message, and The one or more processors are also configured to cause the network entity to transmit the ML configuration information in a Radio Resource Control (RRC) reconfiguration message.

24. The apparatus of claim 14, wherein the at least one radio capability of the UE comprises at least one of the following: The ability to perform ML-based cell reselection processes. The ability to perform ML-based measurement procedures in idle or inactive modes. The ability to perform ML-based Radio Resource Management (RRM) measurement procedures. The ability to perform ML-based radio link monitoring (RLM) procedures. The ability to perform ML-based Channel State Information (CSI) measurement procedures. The ability to perform ML-based feedback processes for precoding matrix indicator (PMI), rank indicator (RI), and channel quality indicator (CQI). The ability to perform ML-based radio link failure (RLF) and beam failure recovery (BFR) procedures, or The ability to perform ML-based RRM relaxation processes.

25. The apparatus of claim 14, wherein the at least one ML capability of the UE comprises at least one of the following: ML training capability ML inference capabilities Processing capacity One or more supported ML model formats, One or more supported ML libraries, or Instructions for one or more local cached ML models.

26. A method for wireless communication by a user equipment (UE), comprising: Transmit UE capability information to network entities, indicating at least one radio capability of the UE and at least one machine learning (ML) capability of the UE; Based on the UE capability information, the network entity receives ML configuration information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF, wherein the at least one NNF includes at least one ML-based wireless communication management procedure, and the at least one ML model is configured to perform the at least one ML-based wireless communication management procedure, wherein: the at least one ML-based wireless communication management procedure includes a channel state information (CSI) measurement procedure; and the at least one ML model includes at least an ML model for performing CSI feedback compression for the CSI measurement procedure; as well as The UE receives an activation signal from the network entity to activate the use of the at least one ML model to perform the ML-based wireless communication management process.

27. A non-transitory computer-readable medium for wireless communication, comprising executable instructions that, when executed by one or more processors of a device, cause the device to: Transmit capability information to network entities indicating at least one radio capability of the device and at least one machine learning (ML) capability of the device; Based on the capability information, the network entity receives ML configuration information indicating at least one neural network function (NNF) and at least one ML model corresponding to the at least one NNF, wherein the at least one NNF includes at least one ML-based wireless communication management procedure, and the at least one ML model is configured to execute the at least one ML-based wireless communication management procedure, wherein: The at least one ML-based wireless communication management process includes a Channel State Information (CSI) measurement process; and the at least one ML model includes at least an ML model for performing CSI feedback compression for the CSI measurement process; and The device receives an activation signal from the network entity to activate the use of the at least one ML model to perform the ML-based wireless communication management process.

28. A method for wireless communication by a network entity, comprising: Receive UE capability information from the user equipment (UE) indicating at least one radio capability and at least one machine learning (ML) capability of the UE; Based on the UE capability information, the system transmits to the UE at least one neural network function (NNF) and ML configuration information corresponding to at least one ML model for the at least one NNF. The at least one NNF includes at least one ML-based wireless communication management procedure, and the at least one ML model is configured to execute the at least one ML-based wireless communication management procedure, wherein: the at least one ML-based wireless communication management procedure includes a channel state information (CSI) measurement procedure; and the at least one ML model includes at least one ML model for performing CSI feedback compression for the CSI measurement procedure; and The network entity transmits an activation signal to the UE to activate the use of at least one ML model to perform the ML-based wireless communication management process.

29. A non-transitory computer-readable medium for wireless communication, comprising executable instructions that, when executed by one or more processors of a device, cause the device to perform the method according to claim 28.

30. An apparatus for wireless communication by a user equipment (UE), comprising means for performing the steps of the method according to claim 26.

31. An apparatus for wireless communication by a network entity, comprising means for performing the steps of the method according to claim 28.

32. A program product comprising instructions that, when executed by one or more processors of a device, cause the device to perform the method according to claim 26 or 28.

Images