Method for resource-aware distributed learning in a wireless communication system

Abstract

The disclosure describes a method of using pre-configured AI / ML (artificial intelligence / machine learning) based model alignment in a wireless mobile communication system including base stations such as gNB, TN, NTN and mobile stations such as UE. When AI / ML models are applied to a radio access network, signaling of model information exchange can not match, resulting in model operation failure. Therefore, model operation such as model training / inference / monitoring / update can be set up between the network and the UE by configuring model alignment and model identification assignment.

Description

Technical Field

[0001] This disclosure relates to AI / ML-based model alignment, wherein techniques are presented for pre-configuring and signaling specific information regarding the alignment of ML models applicable to radio access networks. Background Technology

[0002] Within 3GPP (3rd Generation Partnership Project), one of the selected research projects as part of the approved Release 18 package is AI / ML (Artificial Intelligence / Machine Learning), as described in the relevant document (RP-213599) submitted at 3GPP TSG (Technical Specification Group) RAN (Radio Access Networks) meeting #102. The formal title of the AI / ML research project is "Study on AI / ML for NR Air Interface." The goal of this research project is to establish a public AI / ML framework and areas where AI / ML-based technologies and use cases can yield benefits.

[0003] According to 3GPP, the main objective of this research project is to study an AI / ML framework for the air interface by considering performance, complexity, and potential specification impacts, leveraging target use cases. Specifically, AI / ML models, terminology, and descriptions used to identify common and specific characteristics of the framework are included as one of the key work scopes. Various aspects of the AI / ML framework are being considered for research, and one of the key projects concerns the lifecycle management of AI / ML models, which mandates multiple phases for model training, deployment, inference, monitoring, and updates. Furthermore, in 3GPP, a two-sided (AI / ML) model is defined as a paired AI / ML model on which joint inference is performed, where joint inference includes AI / ML inference performed jointly across the UE and the network. Additionally, for a one-sided (AI / ML) model, a UE-side (AI / ML) model is defined as an AI / ML model whose inference is performed entirely at the UE, and a network-side (AI / ML) model is defined as an AI / ML model whose inference is performed entirely at the network. Currently, the AI / ML specification work is in the work item discussion phase of version 19. Earlier, in 3GPP TR 37.817, version 17, entitled "Study on enhancement for Data Collection for NR and EN-DC," UE (User Equipment) mobility was also considered as one of the AI / ML use cases, and one scenario for model training / inference was that both functions reside within the RAN node. Subsequently, in version 18, a new work project, "Artificial Intelligence (AI) / Machine Learning (ML) for NG-RAN," was initiated to specify data collection enhancements and signaling support within the existing NG-RAN interface and architecture. For the aforementioned proactive standardization efforts, RAN-based AI / ML models are considered crucial for both the network and UE to meet any desired model operations (e.g., model training, inference, selection, handover, updates, monitoring, etc.). Model information can be signaled to pair the network-side model and the UE-side model for various lifecycle management (LCM) operations.

[0004] At the RAN#102 plenary meeting, a new WID for AI / ML for the air interface was approved for inclusion in version 19, which outlines the working objectives of a general AI / ML framework describing the collection of model training data on the UE side. For the FS_NR_AIML_Air study use case, the corresponding UE data collection methods are identified. The UE data collection mechanisms identified during the FS_NR_AIML_Air study (TR 38.843, Section 7.2.1.3.2) are analyzed, along with the impact and limitations of each method.

[0005] For UE data collection, the primary issue is whether UE data collection should be transparent. UE data collection consumes some computation cycles and memory within the UE. For example, UE data collection for AI / ML-based CSI compression and prediction may consume some CPU.

[0006] Currently, only 3 use cases are considered in Rel-19 WI: 1. CSI compression and prediction. 2. Beam management. 3. AIML-assisted localization.

[0007] WO 2022233511A2, entitled "EFFICIENT FEDERATED-LEARNING MODEL TRAINING IN WIRELESS COMMUNICATION SYSTEM," discloses measures for implementing / achieving efficient model training (including model collection and / or aggregation) for federated learning (including hierarchical federated learning) in a wireless communication system. Such measures exemplarily include: a federated learning training host configured for local model training determines how to perform local model training based on the availability of a cluster head and the computational and communication costs of the federated learning training task, and either performs local model training locally or delegates at least a portion of the federated learning training task to a cluster head. Additionally, such measures exemplarily include: the federated learning training host configured for local model training computes a similarity metric between the locally computed set of local model parameters and each set of local model parameters in the received set of local model parameters, and determines whether to act as a temporary cluster head for one or more federated learning training hosts.

[0008] IN 202341061007 A, entitled “DISTRIBUTED EDGE LEARNING WITH LOW POWER DEVICES IN UAVENABLEDIOT FOR 5G NETWORKS”, discloses an environment for edge learning in 5G networks.

[0009] US Patent Application No. 2022036123 A1, entitled "MACHINE LEARNING MODEL SCALING SYSTEM WITH ENERGY EFFICIENTNETWORK DATA TRANSFER FOR POWER AWARE HARDWARE," discloses a Machine Learning Model Exchange (MLMS) framework that selects and exchanges machine learning (ML) models in an energy- and communication-efficient manner, while adapting ML models to real-time changes in system constraints. The MLMS framework includes ML model search strategies that enable ML models to flexibly adapt to a wide variety of computing systems and / or environmental changes. Energy and communication efficiency are achieved through a similarity-based ML model selection process that selects replacement ML models from currently deployed ML models with the most overlapping pre-trained parameters, minimizing memory write operation overhead.

[0010] WO 2023209577 A1, entitled "ML MODEL SUPPORT AND MODEL ID HANDLING BY UE AND NETWORK," discloses instructing and configuring machine learning (ML) model support in a user equipment (UE) within a network. The UE may provide model support information to nodes, including ML type and / or version information of at least one model associated with a specific functionality. Network nodes use the support information and, optionally, previous model ID assignment information from the current node or other nodes to determine at least one model ID and signal this at least one model ID to the UE. The model ID refers to the model in subsequent model handling-related signaling between the UE and the node. The node assigns model IDs such that both the UE and the node are aware that a given model ID refers to a given model or model version (e.g., the mapping is unique for the UE). The model ID may also be unique within a functional domain (e.g., CSI reporting).

[0011] US2022004929 A1, entitled "On-Device Machine Learning Platform," discloses systems and methods for on-device machine learning. Specifically, this disclosure relates to an on-device machine learning platform and related technologies that enable on-device prediction, training, example collection, and / or other machine learning tasks or functionalities. The on-device machine learning platform may include a context provider that can securely inject contextual features into collected training examples and / or input data provided by clients for generating predictions / inference. Therefore, the on-device machine learning platform enables centralized collection of training examples, model training, and the use of machine learning models as services supplied to applications or other clients.

[0012] US Patent 2020388035 A1, entitled "Systems and Methods for Medical Acquisition Processing and Machine Learning for Anatomical Assessment," discloses systems and methods for determining anatomical structures directly from raw medical acquisitions using a machine learning system. One method includes: obtaining raw medical acquisition data from the transport and collection of energy and particles traveling through and originating from the body of one or more individuals; obtaining a parameterized model associated with the anatomical structures of each of the one or more individuals; determining one or more parameters for the parameterized model, wherein the parameters are associated with the raw medical acquisition data; training a machine learning system to predict one or more values ​​for each of the determined parameters for the parameterized model based on the raw medical acquisition data; acquiring medical acquisitions for a selected patient; and using the trained machine learning system to determine parameter values ​​for a patient-specific parameterized model for that patient.

[0013] US2020401939 A1, entitled "Systems and Methods for Preparing Data for Use by Machine Learning Algorithms," discloses historical data used to train machine learning algorithms. This data can have thousands of records and hundreds of fields, and inevitably includes erroneous data that affects the accuracy and usability of the primary model machine learning algorithm. To improve dataset integrity, the historical data is divided into a clean dataset without invalid data values ​​and an erroneous dataset with invalid data values. The clean dataset is used to generate: a second-level model machine learning algorithm, which is trained to generate replacement values ​​for individual invalid data values ​​in multiple complete data records; and a third-level model machine learning clustering algorithm, which is trained to generate replacement values ​​for multiple invalid data values ​​from multiple complete data records. Augmented training data is created by replacing the invalid data values ​​in the erroneous dataset with the replacement data values. This augmented training data is combined with the clean data to train a more accurate and useful primary model.

[0014] US Patent 2023022050 A1, entitled "AIML-BASED CONTINUOUS DELIVERY FOR NETWORKS," discloses an application in a distributed computing environment. Telemetry data corresponding to the application's deployment is collected. This telemetry data is received by a machine learning model, which is trained using test telemetry data to determine whether the deployment is successful or failed. Successful inference causes the deployment to continue, while failed inference causes the application to roll back.

[0015] Advanced distributed learning methods (such as split learning and federated learning) promise to enhance network operation and performance. However, UEs face constraints in terms of energy and computational resources and capabilities. For advanced distributed AI / ML, the efficient use of these limited resources and capabilities is crucial. Currently, the network cannot determine whether a UE supports distributed learning or the extent to which it does. Therefore, the network randomly selects UEs to participate in distributed learning, which can lead to non-convergence of AI / ML models.

[0016] The problem is addressed through a newly proposed UE capability, in which the UE indicates to the NW whether it supports distributed learning. If the UE supports distributed learning, it provides the NW with supplemental information regarding whether it wishes to participate in distributed learning.

[0017] Network awareness enables UEs to participate in distributed learning and their capabilities, and distributed learning can be configured accordingly, thus reducing signaling overhead and resulting in UE energy savings.

[0018] The existing technologies demonstrate contributions relevant to key areas of the chosen classification. The listed existing technologies encompass various approaches to AI / ML-based applications for federated learning. However, the proposed concept features a differentiated mechanism that considers UE constraints.

[0019] This disclosure addresses the cited problems through proposed embodiments and, as a first aspect, describes a method for resource-aware distributed learning in a wireless communication system, wherein the UE processes the following steps: a) The UE checks whether it has received a UE capability request. b) If the check in step a is correct, the UE indicates distributed learning capability. c) The UE checks whether it has received the capability report configuration. d) If the check in step c is correct, the UE checks whether the reporting criteria are met. e) If the check in step d is correct, the UE reports an update to the distributed learning capability.

[0020] In some embodiments of the method according to the first aspect, the method is characterized in that the distributed learning is federated and / or split-based, or any method that includes the method of training a global model in smaller, geographically dispersed parts.

[0021] In some embodiments of the method according to the first aspect, the method is characterized in that the distributed learning is federated learning and / or split learning, and uses L1 / L2 / L3 UEAssistanceInformation messages.

[0022] In some embodiments of the method according to the first aspect, the method is characterized in that the UE determines its indication to change based on an implementation and sends updates periodically or aperiodically, wherein the implementation uses a dataset threshold, dataset availability, energy level and / or application layer configuration.

[0023] In some embodiments of the method according to the first aspect, the method is characterized in that the dataset threshold is configured by the gNB via a UE-specific message or a system information message.

[0024] This disclosure addresses the cited problems through proposed embodiments and, as a second aspect, describes resource-aware distributed learning for wireless communication systems, wherein the gNB processes the following steps: a. gNB indicates support for distributed learning b. The gNB inquires with the UE about support for distributed learning. c. gNB provides configuration for distributed learning and UE capability reporting. d. gNB checks whether it has received a report on the distributed learning UE capability. e. If step d is correct, the gNB adjusts the distributed learning configuration and / or the configuration used for distributed learning UE capability reporting.

[0025] In some embodiments of the method according to the second aspect, the method is characterized in that the gNB configures the UE to, for example, periodically report updates based on UE-specific messages or system information messages.

[0026] In some embodiments of the method according to the second aspect, the method is characterized in that the periodic reporting is triggered based on the periodicity configured by the gNB.

[0027] In some embodiments of the method according to the second aspect, the method is characterized in that the non-periodic reporting is triggered based on the UE implementation.

[0028] In some embodiments of the method according to the second aspect, the method is characterized in that the UE implements a method to check whether the UE's energy state is low or whether the UE's buffer is full.

[0029] In some embodiments of the method according to the second aspect, the method is characterized in that the UE implements a check to see if the UE is unwilling to send any UL data and / or the UE checks whether it is based on UE mobility.

[0030] According to a third aspect, this disclosure relates to an apparatus for resource-aware distributed learning in a wireless communication system, the apparatus including a wireless transceiver and a processor coupled to a memory storing computer program instructions therein, the instructions being configured to perform the steps according to the first aspect.

[0031] According to the fourth aspect, this disclosure relates to an apparatus for resource-aware distributed learning in a wireless communication system, the apparatus including a wireless transceiver and a processor coupled to a memory storing computer program instructions therein, the instructions being configured to perform the steps according to the second aspect.

[0032] According to the fifth aspect, this disclosure addresses the cited problems through proposed embodiments and is described through a user equipment including the device according to the third aspect.

[0033] According to the sixth aspect, this disclosure addresses the cited problems through proposed embodiments and is described by means of a gNB including the device according to the third aspect.

[0034] According to the seventh aspect, this disclosure relates to wireless communication for resource-aware distributed learning, wherein the wireless communication system includes at least one user equipment according to the fifth aspect and at least one gNB according to the sixth aspect, wherein the user equipment and the gNB each include a processor coupled to a memory storing computer program instructions therein, and perform the steps of the first aspect for the UE and the second aspect for the gNB.

[0035] According to the eighth aspect, this disclosure relates to a computer program product comprising instructions that, when executed by at least one processor, configure the at least one processor to perform the method according to the first aspect, and / or configure the at least one processor to perform a method for exchanging data according to any embodiment of the embodiments of this disclosure. The computer program product may use any programming language and may be in the form of source code, object code, or any intermediate form between source code and object code, such as a partially compiled form, or any other desired form.

[0036] According to a ninth aspect, this disclosure relates to a computer-readable storage medium including instructions that, when executed by at least one processor, configure the at least one processor to perform a method according to any embodiment of the present disclosure. Attached Figure Description

[0037] Figure 1 This is an exemplary process for the UE side.

[0038] Figure 2 This is an exemplary flowchart for the gNB side. Detailed Implementation

[0039] The detailed description set forth below with reference to the accompanying drawings is intended to describe various configurations and is not intended to represent only the configurations in which the concepts described herein may be practiced. The detailed description includes specific details for the purpose of providing a thorough understanding of the various concepts. However, it will be apparent to those skilled in the art that these concepts may be practiced without these specific details. Specifically, although terms from 3GPP 5G NR may be used in this disclosure to illustrate embodiments herein, this should not be construed as limiting the scope of the invention.

[0040] Some of the embodiments contemplated herein will now be described more fully with reference to the accompanying drawings. However, other embodiments are also included within the scope of the subject matter disclosed herein, and the disclosed subject matter should not be construed as being limited to the embodiments set forth herein; rather, these embodiments are provided by way of example to convey the scope of the subject matter to those skilled in the art.

[0041] Generally, all terms used herein should be interpreted according to their ordinary meaning in the relevant art, unless a different meaning is expressly given and / or implied in the context of their use. Unless otherwise expressly stated, all references to elements, devices, components, means, steps, etc., should be openly interpreted as referring to at least one instance of that element, device, component, means, step, etc. The steps of any method disclosed herein need not be performed in the exact order disclosed, unless a step is explicitly described as occurring after or before another step, and / or it is implied that a step must occur after or before another step. Where appropriate, any feature of any embodiment of the embodiments disclosed herein may be applied to any other embodiment. Similarly, any advantage of any embodiment in the embodiments may be applied to any other embodiment, and vice versa. Other objects, features, and advantages of the appended embodiments will become apparent from the following description.

[0042] In some embodiments, the more general term "network node" may be used, and this term may correspond to any type of radio network node or any network node that communicates with the UE (directly or via another node) and / or with another network node. Examples of network nodes are NodeB, MeNB, ENB, network nodes belonging to MCG or SCG, base stations (BS), multi-standard radio (MSR) radio nodes (such as MSR BS), eNodeB, gNodeB, network controller, radio network controller (RNC), base station controller (BSC), repeater, donor node controlling repeater, base transceiver station (BTS), access point (AP), transport point, transport node, RRU, RRH, nodes in distributed antenna system (DAS), core network nodes (e.g., mobile switching center (MSC), mobility management entity (MME), etc.), operations and maintenance (O&M), operations support system (OSS), self-optimizing network (SON), location nodes (e.g., evolved servicing mobile location center (E-SMLC)), minimized drive test (MDT), test equipment (physical node or software), etc.

[0043] In some embodiments, the non-limiting terms User Equipment (UE) or Wireless Device may be used, and the term may refer to any type of wireless device that communicates with a network node and / or with another UE in a cellular or mobile communication system. Examples of UEs are target devices, device-to-device (D2D) UEs, machine-type UEs or UEs capable of machine-to-machine (M2M) communication, PDAs, PADs, tablet computers, mobile terminals, smartphones, laptop embedded devices (LEE), laptop mounted devices (LME), USB dongles, UE class M1, UE class M2, ProSe UE, V2V UE, V2X UE, etc.

[0044] Additionally, terms such as base station / gNodeB and UE should be considered non-restrictive and, in particular, do not imply any hierarchical relationship between them; generally, "gNodeB" can be considered device 1 and "UE" can be considered device 2, and the two devices communicate with each other via a radio channel. Furthermore, in the following text, the sender or receiver can be either a gNodeB (gNB) or a UE.

[0045] As those skilled in the art will understand, aspects of the embodiments may be embodied as a system, device, method, or program product. Therefore, embodiments may take the form of an all-hardware embodiment, an all-software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects.

[0046] For example, the disclosed embodiments can be implemented as hardware circuitry, including custom-designed very large-scale integration (“VLSI”) circuitry or gate arrays, off-the-shelf semiconductors (such as logic chips, transistors, or other discrete components). The disclosed embodiments can also be implemented in programmable hardware devices such as field-programmable gate arrays, programmable array logic, programmable logic devices, etc. As another example, the disclosed embodiments may include one or more physical or logical blocks of executable code, which may, for example, be organized as objects, procedures, or functions.

[0047] Furthermore, embodiments may take the form of a program product embodied in one or more computer-readable storage devices, which store machine-readable code, computer-readable code, and / or program code, hereinafter referred to as code. The storage device may be tangible, non-transitory, and / or non-transferable. The storage device may not contain signals. In one embodiment, the storage device uses only signals to access the code.

[0048] Any combination of one or more computer-readable media may be used. A computer-readable medium may be a computer-readable storage medium. A computer-readable storage medium may be a storage device for storing code. A storage device may be, for example, but not limited to, electronic, magnetic, optical, electromagnetic, infrared, holographic, micromechanical, or semiconductor systems, devices, or apparatuses, or any suitable combination of the foregoing.

[0049] More specific examples of storage devices (a non-exhaustive list) will include the following: electrical connections having one or more wires; portable computer floppy disks; hard disks; random access memory (“RAM”); read-only memory (“ROM”); erasable programmable read-only memory (“EPROM” or flash memory); portable optical disc read-only memory (“CD-ROM”); optical storage devices; magnetic storage devices; or any suitable combination of the foregoing. In the context of this document, a computer-readable storage medium can be any tangible medium that can contain or store programs for use by or in conjunction with an instruction execution system, device, or apparatus.

[0050] The code used to implement the operations of the embodiments can be any number of lines and can be written in any combination of one or more programming languages, including object-oriented programming languages ​​(such as Python, Ruby, Java, Smalltalk, C++, etc.), as well as conventional procedural programming languages ​​(such as the "C" programming language, etc.) and / or machine languages ​​(such as assembly language). The code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer via any type of network (including a local area network ("LAN"), a wireless LAN ("WLAN"), or a wide area network ("WAN"), or can be connected to an external computer (e.g., via the Internet through an Internet Service Provider ("ISP").

[0051] Furthermore, the features, structures, or characteristics described in the embodiments can be combined in any suitable manner. Numerous specific details, such as examples of programming, software modules, user selection, network transactions, database queries, database structures, hardware modules, hardware circuits, hardware chips, etc., are provided in the following description to provide a thorough understanding of the embodiments. However, those skilled in the art will recognize that the embodiments can be practiced without one or more of these specific details or using other methods, components, materials, etc. In other instances, well-known structures, materials, or operations are not shown or described in detail to avoid obscuring aspects of the embodiments. Throughout the specification, references to “an embodiment,” “embodiment,” or similar language mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Therefore, unless expressly stated otherwise, the phrases “in one embodiment,” “in an embodiment,” and similar language appearing throughout the specification may, but do not necessarily all refer to the same embodiment, but rather mean “one or more, but not all, embodiments.” Unless expressly stated otherwise, the terms “comprising,” “including,” “having,” and variations thereof mean “including, but not limited to,” “including.” Unless expressly stated otherwise, the enumeration list of items does not imply that any or all of the items are mutually exclusive. Unless otherwise expressly stated, the terms “a,” “one,” and “the / said” also mean “one or more.”

[0052] The following description refers to schematic flowcharts and / or block diagrams of methods, apparatus, systems, and program products according to embodiments. It should be understood that each block of the schematic flowcharts and / or block diagrams, and combinations of blocks of the schematic flowcharts and / or block diagrams, can be implemented by code. This code can be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus to produce a machine such that instructions executable via the processor of the computer or other programmable data processing apparatus establish components for implementing the functions / actions specified in the flowcharts and / or block diagrams.

[0053] The code may also be stored in a storage device that can instruct a computer, other programmable data processing device or other device to operate in a particular manner, such that the instructions stored in the storage device produce an article of art, which includes instructions that implement the functions / actions specified in the flowchart and / or block diagram.

[0054] The code may also be loaded onto a computer, other programmable data processing apparatus or other device such that a series of operational steps to be performed on the computer, other programmable apparatus or other device produce a computer-implemented process, such that the code to be executed on the computer or other programmable apparatus provides a process for implementing the functions / actions specified in the flowchart and / or block diagram.

[0055] The flowcharts and / or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, systems, methods, and program products according to various embodiments. In this regard, each block in the flowcharts and / or block diagrams may represent a module, segment, or portion of code, which includes one or more executable instructions of that code for implementing a specified logical function.

[0056] It should also be noted that in some alternative implementations, the functions mentioned in the boxes may not occur in the order shown in the figures. For example, depending on the functionality involved, two boxes shown consecutively may actually be executed substantially concurrently, or the boxes may sometimes be executed in reverse order. Other steps and methods that are functionally, logically, or effectively equivalent to one or more boxes or portions thereof in the figures shown are conceivable.

[0057] While various arrow and line types may be used in flowcharts and / or block diagrams, they are not intended to limit the scope of the corresponding embodiments. In fact, some arrows or other connectors may be used only to indicate the logical flow of the depicted embodiment. For example, an arrow may indicate a waiting or monitoring period of unspecified duration between enumerated steps in the depicted embodiment. It should also be noted that each block in the block diagram and / or flowchart, as well as combinations of blocks in the block diagram and / or flowchart, may be implemented by a system based on dedicated hardware or a combination of dedicated hardware and code that performs the specified function or action.

[0058] The description of the elements in each figure can be referenced to the elements in the preceding figures. Throughout all figures, the same numbers refer to the same elements, including alternative embodiments of the same elements.

[0059] The detailed description set forth below with reference to the accompanying drawings is intended as a description of various configurations and is not intended to represent only the configurations in which the concepts described herein may be practiced. The detailed description includes specific details for the purpose of providing a thorough understanding of the various concepts. However, it will be apparent to those skilled in the art that these concepts may be practiced without these specific details. For example, although 3GPP terminology from, for example, 5G NR may be used in this disclosure to illustrate embodiments thereof, this should not be construed as limiting the scope of this disclosure.

[0060] This disclosure relates to a wireless communication system, which may be, for example, a 5G NR wireless communication system. More specifically, it refers to a RAN (Radio Access Network) within the wireless communication system for exchanging data with a UE via radio signals. For example, the RAN may transmit data (downlink DL) to the UE, such as instance data received from the core network (CN). The RAN may also receive data (uplink UL) from the UE, which may be forwarded to the CN.

[0061] In the example shown, the RAN includes a base station (BS). Of course, the RAN may include more than one BS to increase the coverage of the wireless communication system. Depending on the implemented wireless communication standard, each of these BSs may be referred to as an NB, eNodeB (or eNB), gNodeB (or gNB in ​​the case of a 5G NR wireless communication system), access point, etc. The UE is located within the coverage area of ​​the BS. The coverage area of ​​the BS corresponds, for example, to an area where the UE can decode PDCCH transmitted by the BS.

[0062] Examples of wireless devices suitable for implementing any of the methods discussed in this disclosure at the UE correspond to devices that provide wireless connectivity to a RAN of a wireless communication system and can be used to exchange data with that RAN. Such wireless devices may be included in the UE. The UE may be, for example, a cellular phone, a wireless modem, a wireless communication device, a handheld device, a laptop computer, etc. The UE may also be an Internet of Things (IoT) device, such as a wireless camera, a smart sensor, a smart meter, smart glasses, a vehicle (manned or unmanned), a GPS device, etc., or any other device capable of running applications that require the exchange of data with a remote receiver via a wireless device.

[0063] The wireless device includes one or more processors and one or more memories. The one or more processors may include, for example, a central processing unit (CPU), a digital signal processor (DSP), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), etc. The one or more memories may include any type of computer-readable volatile and non-volatile memory (hard disk, solid-state drive, optical disk, electronic storage, etc.). The one or more memories may store a computer program product in the form of a set of program code instructions, which are to be executed by the one or more processors to implement all or part of the steps of a method for exchanging data on the UE side according to any embodiment of the embodiments disclosed herein.

[0064] The wireless device may further include a main radio (MR) unit. This MR unit corresponds to the main wireless communication unit of the wireless device, which is used to exchange data with the BS of the RAN using radio signals. The MR unit may implement one or more wireless communication protocols and may be, for example, a 3G, 4G, 5G, NR, WiFi, WiMax, or other transceivers. In a preferred embodiment, the MR unit corresponds to a 5G NR wireless communication unit.

[0065] The following explanation provides a detailed description of the mechanism for pre-configuring and signaling specific information about online model training by configuring a set of UE behaviors. AI / ML-based technologies are currently applied in many different applications, and based on observed potential benefits, 3GPP has also begun its technology research to apply them to multiple use cases. The AI / ML lifecycle can be broken down into several phases, such as data collection / preprocessing, model training, model testing / validation, model deployment / update, and model monitoring, each of which is equally important for achieving the target performance of any particular model.

[0066] When deploying a wireless communication network that enables AI / ML models, it is then important to consider how to handle the activated AI / ML models in terms of reconfiguration for wireless devices during operations such as model training, inference, and updates. The model used for activation must be aligned between the network side and the UE side using model ID information. If misaligned, model activation may fail due to mismatches in ML conditions and / or ML operation configurations between the network side and the UE side.

[0067] Figure 1 This is an example process for the UE side.

[0068] The UE sends its distributed learning capabilities (e.g., UECapability messages) to the network, whereby the UE indicates capabilities based on device characteristics (e.g., computing power, supported AI / ML frameworks, or models). The UE indicates its intention to participate in distributed learning (e.g., federated learning, split learning), for example, using L1 / L2 / L3 UEAssistanceInformation messages. The UE determines whether to change its indications based on implementation (e.g., dataset thresholds, energy levels, application layer configuration) and sends updates periodically or aperiodically. The dataset threshold is configured by the gNB via UE-specific messages or system information messages.

[0069] When a gNB / network requests a UE to provide its capabilities, the UE indicates its capabilities, for example, based on its device characteristics (e.g., computing power, supported AI / ML frameworks or models).

[0070] When the gNB / network provides reporting configuration, the UE applies the received configuration to check whether the configured reporting criteria are met. For example, the gNB / network can configure the UE to provide periodic or non-periodic updates to its distributed learning capabilities.

[0071] When the configured reporting criteria are met, the UE provides distributed learning capability updates to the gNB / network.

[0072] according to Figure 1The method for resource-aware distributed learning in wireless communication systems is characterized by the following steps performed by the UE: a. The UE checks whether it has received a UE capability request; b. If the check in step a is correct, the UE indicates distributed learning capability; c. The UE checks whether it has received the capability report configuration; d. If the check in step c is correct, the UE checks whether the reporting criteria are met; e. If the check in step d is correct, the UE reports an update to the distributed learning capability.

[0073] The method of claim 1, wherein the distributed learning is federated or split-based, or any method that includes the method of training a global model in smaller, geographically dispersed parts.

[0074] Furthermore, the UE determines its indication for change based on an implementation method and sends updates periodically or aperiodically, wherein the implementation method uses a dataset threshold and / or energy level and / or application layer configuration. Additionally, the dataset threshold is configured by the gNB via a UE-specific message or system information message.

[0075] Figure 2 This is an exemplary flowchart for the gNB side. The method for resource-aware distributed learning in a wireless communication system is characterized by the gNB handling the following steps: gNB indicates support for distributed learning; gNB requests UE's support for distributed learning; gNB provides configuration for distributed learning and UE capability reporting; gNB checks whether it has received a report on the distributed learning UE capability;

[0076] If step d is correct, the gNB adjusts the distributed learning configuration and / or the configuration used for distributed learning UE capability reporting.

[0077] In addition, the gNB configures the UE to periodically report updates based on UE-specific messages or system information messages, and the periodic reporting is triggered based on the periodicity configured by the gNB.

[0078] Non-periodic reporting is triggered based on the UE implementation. One possible UE implementation is that it checks if the UE's energy state is low. Additionally, the UE implementation checks if the UE is unwilling to send any UL data, has no data to send, and / or checks whether it's based on UE mobility. This means the gNB configures the UE to report updates periodically, for example, based on UE-specific messages or system information messages. Periodic reporting is triggered based on the periodicity configured by the gNB. Non-periodic reporting is triggered based on the UE implementation (e.g., whether the energy state is low, whether the UE doesn't want to send any UL data, based on UE mobility).

[0079] The gNB / network indicates support for distributed learning and queries the UE for its distributed learning capabilities. Based on the received distributed learning UE capabilities, the gNB / network determines and provides the UE with the configuration for distributed learning and UE capability reporting. When the gNB / network receives a UE report updating its distributed learning capabilities, it adjusts the distributed learning configuration and / or the configuration for distributed learning UE capability reporting.

[0080] The dataset threshold can be defined, for example, as "the minimum number of data points that the device can collect to qualify for split learning".

[0081] UE signaling is performed as follows: For L1, one bit in the existing AI / ML-related physical UL control channel (e.g., PUCCH) (participating in distributed learning); for L2, a new logical channel ID carrying this information in the MAC CE; and for L3, in the RRC (e.g., UEAssistanceInformation) message. This means the network is aware of the UE's capabilities and can configure distributed learning accordingly. Signaling overhead can be significantly reduced. This translates to UE energy savings.

[0082] A preferred embodiment is an apparatus for a method of resource-aware distributed learning in a wireless communication system, the apparatus comprising a wireless transceiver, a processor coupled to a memory therein storing computer program instructions configured to implement the steps of the method as defined in claims 1 to 4.

[0083] Another preferred embodiment is an apparatus for a method of resource-aware distributed learning in a wireless communication system, the apparatus comprising a wireless transceiver, a processor coupled to a memory storing computer program instructions therein configured to implement the steps of the method as defined in claims 5 to 10.

[0084] Another preferred embodiment is a user equipment comprising the device according to claim 11.

[0085] Another embodiment is a gNB that includes the device according to claim 12.

[0086] Another embodiment is a wireless communication system for resource-aware distributed learning, wherein the wireless communication system includes at least one user equipment according to claim 13 and at least one gNB according to claim 14, wherein the user equipment and the gNB each include a processor coupled to a memory in which computer program instructions are stored.

Claims

1. A method for allocating signaling in a high-level model in a wireless communication system, wherein, The UE performs the following steps: a) The UE checks whether it has received a UE capability request; b) If the check in step a is correct, the UE indicates distributed learning capability; c) The UE checks whether it has received a capability report configuration; d) If the check in step c is correct, does the UE check meet the reporting criteria? e) If the check in step d is correct, the UE reports an update to the distributed learning capability.

2. The method according to claim 1, wherein, Distributed learning is either federated or split-based, or any approach that includes training a global model in smaller, geographically dispersed parts.

3. The method according to any one of the preceding claims, wherein, The UE determines its indication to change based on an implementation method and sends updates periodically or aperiodically, wherein the implementation method uses a dataset threshold, dataset availability, energy level and / or application layer configuration.

4. The method according to any one of the preceding claims, wherein, The dataset threshold is configured by the gNB via UE-specific messages or system information messages.

5. A method for assigning signaling to advanced models in a wireless communication system, wherein, gNB performs the following steps: a. The gNB indicates support for distributed learning; b. gNB requests support for distributed learning from UE; c. gNB provides configuration for distributed learning and UE capability reporting; d. The gNB checks whether it has received a report on the distributed learning UE capability; e. If step d is correct, the gNB adjusts the distributed learning configuration and / or the configuration used for distributed learning UE capability reporting.

6. The method according to claim 5, wherein, The gNB configures the UE to, for example, periodically report updates based on UE-specific messages or system information messages.

7. The method according to any one of claims 5 or 6, wherein, Periodic reports are triggered based on the periodicity configured by gNB.

8. The method according to any one of claims 5 or 7, wherein, Non-periodic reports are triggered based on the UE implementation method.

9. The method according to any one of claims 5 to 8, wherein, The UE implementation checks whether the UE's energy state is low or whether the UE's buffer is full.

10. The method according to any one of claims 5 to 9, wherein, The UE implementation checks whether the UE is unwilling to send any UL data, whether it has data to send, and / or whether the UE check is based on UE mobility.

11. An apparatus for a method of resource-aware distributed learning in a wireless communication system, the apparatus comprising a wireless transceiver, a processor coupled to a memory therein storing computer program instructions configured to implement the steps of claims 1 to 4.

12. An apparatus for a method of a resource-aware distributed learning communication system, the apparatus comprising a wireless transceiver and a processor coupled to a memory therein storing computer program instructions configured to implement the steps of claims 5 to 10.

13. A user equipment comprising the device according to claim 11.

14. A gNB comprising the device according to claim 12.

15. A wireless communication for advanced model allocation signaling that indicates different model ID allocations by configuring model alignment ID or index information, wherein, The wireless communication system includes at least one user equipment according to claim 13 and at least one gNB according to claim 14, wherein the user equipment and the gNB each include a processor coupled to a memory in which computer program instructions are stored.

Images