AI / ML based bilateral model in multi-cell scenario

Abstract

Embodiments of the present disclosure relate to artificial intelligence (AI) / machine learning (ML)-based bilateral models in a multi-cell scenario. In an aspect, a terminal device sends, to a first network device, first information about outputs of a first one or more machine learning models available at the terminal device based on a change in channel conditions associated with the terminal device. Further, the terminal device receives, from the first network device, second information about performance of a second one or more machine learning models available at one or more further network devices. The performance is obtained based on the first information.

Description

Technical Field

[0001] Various example embodiments relate to the field of communications, and more specifically to devices, methods, apparatuses, and computer-readable storage media for two-sided models based on artificial intelligence (AI) / machine learning (ML) in multi-cell scenarios. Background Technology

[0002] A communication network can be viewed as a facility that enables communication between two or more communication devices or provides communication devices with access to a data network. Mobile or wireless communication networks are an example of communication networks.

[0003] Such communication networks operate according to standards such as those issued by the 3rd Generation Partnership Project (3GPP) or the European Telecommunications Standards Institute (ETSI). Examples of such standards include the so-called fifth-generation (5G) standard or other standards issued by 3GPP. Summary of the Invention

[0004] Generally, the exemplary embodiments of this disclosure provide solutions related to AI / ML-based two-sided models in multi-cell scenarios.

[0005] In a first aspect, a terminal device is provided. The terminal device includes at least one processor and at least one memory storing instructions, wherein when executed by the at least one processor, the terminal device at least: sends first information to a first network device regarding the output of one or more first machine learning models available at the terminal device; receives from the first network device second information regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information; and performs cell selection or cell reselection at least based on the second information.

[0006] In a second aspect, a first network device is provided. The first network device includes at least one processor and at least one memory storing instructions, the instructions, when executed by the at least one processor, causing the first network device to at least: receive first information from a terminal device regarding the output of one or more first machine learning models available at the terminal device; and send to the terminal device second information regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0007] In a third aspect, a terminal device is provided. The terminal device includes at least one processor and at least one memory storing instructions, the instructions, when executed by the at least one processor, causing the terminal device to at least: send first information to a first network device regarding the output of one or more first machine learning models available at the terminal device; receive second information from the first network device regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and perform a handover from the first network device to the second network device.

[0008] In a fourth aspect, a first network device is provided. The first network device includes at least one processor and at least one memory storing instructions, which, when executed by the at least one processor, cause the first network device to at least: receive first information from a terminal device regarding the output of one or more first machine learning models available at the terminal device; send second information to the terminal device regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and perform a handover process from the terminal device to the second network device.

[0009] In a fifth aspect, a terminal device is provided. The terminal device includes at least one processor and at least one memory storing instructions, the instructions, when executed by the at least one processor, causing the terminal device to at least: transmit first information to a first network device regarding the output of one or more first machine learning models available at the terminal device, based on a change in channel conditions associated with the terminal device; and receive second information from the first network device regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0010] In a sixth aspect, a first network device is provided. The first network device includes at least one processor and at least one memory storing instructions, the instructions, when executed by the at least one processor, causing the first network device to at least: receive from a terminal device at least one of the following: first information regarding the output of one or more first machine learning models available at the terminal device; or an indication of a change in channel conditions associated with the terminal device; and send to the terminal device second information regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0011] In a seventh aspect, a terminal device is provided. The terminal device includes at least one processor and at least one memory storing instructions, the instructions, when executed by the at least one processor, cause the terminal device to at least: send first information to a first network device regarding the output of a first or more machine learning models available at the terminal device, based on an update of a machine learning model at the terminal device; and receive second information from the first network device regarding the performance of a second or more machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0012] In an eighth aspect, a first network device is provided. The first network device includes at least one processor and at least one memory storing instructions, the instructions, when executed by the at least one processor, causing the first network device to at least: receive from a terminal device at least one of the following: first information regarding the output of one or more first machine learning models available at the terminal device; or an indication of updating the machine learning models at the terminal device; and send to the terminal device second information regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0013] In a ninth aspect, a method is provided. The method includes: at a terminal device, sending first information to a first network device regarding the output of one or more first machine learning models available at the terminal device; at the terminal device, receiving from the first network device second information regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information; and at the terminal device, performing cell selection or cell reselection based at least on the second information.

[0014] In a tenth aspect, a method is provided. The method includes: receiving, at a first network device, first information from a terminal device regarding the output of one or more first machine learning models available at the terminal device; and sending, at the first network device, second information to the terminal device regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0015] In an eleventh aspect, a method is provided. The method includes: at a terminal device, sending first information to a first network device regarding the output of one or more first machine learning models available at the terminal device; at the terminal device, receiving second information from the first network device regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and performing a handover from the first network device to the second network device at the terminal device.

[0016] In a twelfth aspect, a method is provided. The method includes: at a first network device, receiving from a terminal device first information regarding the output of one or more first machine learning models available at the terminal device; at the first network device, sending to the terminal device second information regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and at the first network device, performing a handover process from the terminal device to the second network device.

[0017] In a thirteenth aspect, a method is provided. The method includes: transmitting, at the terminal device, first information about the output of one or more first machine learning models available at the terminal device to a first network device based on a change in channel conditions associated with the terminal device; and receiving, at the terminal device, second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0018] In a fourteenth aspect, a method is provided. The method includes: receiving, at a first network device, at a terminal device, at at least one of the following: first information regarding the output of one or more first machine learning models available at the terminal device; or an indication of a change in channel conditions associated with the terminal device; and transmitting, at the first network device, to the terminal device, second information regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0019] In a fifteenth aspect, a method is provided. The method includes: sending, at the terminal device, first information about the output of one or more first machine learning models available at the terminal device to a first network device based on machine learning model updates at the terminal device; and receiving, at the terminal device, second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0020] In a sixteenth aspect, a method is provided. The method includes: receiving, at a first network device, at a terminal device, at at least one of the following: first information regarding the output of a first or more machine learning models available at the terminal device; or an instruction to update a machine learning model at the terminal device; and sending, at the first network device, second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0021] In a seventeenth aspect, an apparatus is provided. The apparatus includes: means for transmitting, at a terminal device, first information about the output of one or more first machine learning models available at the terminal device to a first network device; means for receiving, at the terminal device, second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information; and means for performing cell selection or cell reselection at the terminal device based at least on the second information.

[0022] In an eighteenth aspect, an apparatus is provided. The apparatus includes: means for receiving, at a first network device, first information from a terminal device regarding the output of one or more first machine learning models available at the terminal device; and means for transmitting, at the first network device, second information from the terminal device regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0023] In a nineteenth aspect, an apparatus is provided. The apparatus includes: means for transmitting, at a terminal device, first information about the output of one or more first machine learning models available at the terminal device to a first network device; means for receiving, at the terminal device, second information about the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and means for performing a handover from the first network device to the second network device at the terminal device.

[0024] In a twentieth aspect, an apparatus is provided. The apparatus includes: means for receiving, at a first network device, first information from a terminal device regarding the output of one or more first machine learning models available at the terminal device; means for sending, at the first network device, second information from the terminal device regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and means for performing a handover process from the terminal device to the second network device at the first network device.

[0025] In a twenty-first aspect, an apparatus is provided. The apparatus includes components for: transmitting, at the terminal device, first information regarding the output of one or more first machine learning models available at the terminal device to a first network device based on a change in channel conditions associated with the terminal device; and receiving, at the terminal device, second information regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0026] In a twenty-second aspect, an apparatus is provided. The apparatus includes components for: receiving, at a first network device, at a terminal device, at at least one of the following: first information regarding the output of one or more first machine learning models available at the terminal device; or an indication of a change in channel conditions associated with the terminal device; and transmitting, at the first network device, to the terminal device, second information regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0027] In a twenty-third aspect, an apparatus is provided. The apparatus includes components for: sending first information at the terminal device to a first network device regarding the output of one or more first machine learning models available at the terminal device, based on machine learning model updates at the terminal device; and receiving, at the terminal device, second information from the first network device regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0028] In a twenty-fourth aspect, an apparatus is provided. The apparatus includes: components for receiving, at a first network device, at a terminal device, at at least one of the following: first information regarding the output of a first or more machine learning models available at the terminal device; or an instruction to update a machine learning model at the terminal device; and components for transmitting, at the first network device, to the terminal device, second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0029] In a twenty-fifth aspect, a non-transitory computer-readable medium is provided that, when executed by a device, causes the device to perform at least the method according to any one of the ninth to sixteenth aspects above.

[0030] In a twenty-sixth aspect, a computer program product is provided, comprising program instructions for at least executing the method according to any one of the ninth to sixteenth aspects above.

[0031] In the twenty-seventh aspect, a computer program including instructions is provided, which, when executed by a device, cause the device to perform at least the method according to any one of the ninth to sixteenth aspects above.

[0032] In a twenty-eighth aspect, a terminal device is provided. The terminal device includes: a transmitting circuitry configured to: transmit first information to a first network device regarding the output of a first or more machine learning models available at the terminal device; a receiving circuitry configured to receive from the first network device second information regarding the performance of a second or more machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information; and an execution circuitry configured to perform cell selection or cell reselection based at least on the second information.

[0033] In a twenty-ninth aspect, a first network device is provided. The first network device includes: a receiving circuitry configured to receive from a terminal device first information regarding the output of one or more first machine learning models available at the terminal device; and a transmitting circuitry configured to transmit to the terminal device second information regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0034] In a thirtieth aspect, a terminal device is provided. The terminal device includes: a transmitting circuitry configured to: transmit first information to a first network device regarding the output of one or more first machine learning models available at the terminal device; a receiving circuitry configured to receive from the first network device second information regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and an execution circuitry configured to perform a handover from the first network device to the second network device.

[0035] In a thirty-first aspect, a first network device is provided. The first network device includes: a receiving circuitry configured to receive from a terminal device first information regarding the output of one or more first machine learning models available at the terminal device; a transmitting circuitry configured to transmit to the terminal device second information regarding the performance of one or more second machine learning models available at a second network device, wherein the performance is obtained based on the first information; and an execution circuitry configured to execute a handover process from the terminal device to the second network device.

[0036] In a thirty-second aspect, a terminal device is provided. The terminal device includes: a transmitting circuitry configured to transmit first information to a first network device regarding the output of one or more first machine learning models available at the terminal device based on a change in channel conditions associated with the terminal device; and a receiving circuitry configured to receive from the first network device second information regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0037] In a thirty-third aspect, a first network device is provided. The first network device includes: a receiving circuitry configured to receive from a terminal device at least one of the following: first information regarding the output of a first or more machine learning models available at the terminal device; or an indication of a change in channel conditions associated with the terminal device; and a transmitting circuitry configured to transmit to the terminal device second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0038] In a thirty-fourth aspect, a terminal device is provided. The terminal device includes: a transmitting circuitry configured to transmit first information about the output of one or more first machine learning models available at the terminal device to a first network device based on updates to machine learning models at the terminal device; and a receiving circuitry configured to receive from the first network device second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0039] In a thirty-fifth aspect, a first network device is provided. The first network device includes a receiving circuitry configured to receive from a terminal device at least one of the following: first information regarding the output of a first or more machine learning models available at the terminal device; or an indication of an update to a machine learning model at the terminal device; and a transmitting circuitry configured to transmit to the terminal device second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0040] It should be understood that the summary portion is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0041] Some exemplary embodiments will now be described with reference to the accompanying drawings, in which: Figure 1A An example environment in which example embodiments of this disclosure may be implemented is shown; Figure 1B An example autoencoder architecture associated with some example embodiments of this disclosure is shown; Figure 2 A first signaling flow between a terminal device and a first network device according to some example embodiments of the present disclosure is shown; Figure 3A Example illustrations illustrating the permissibility of model / function sharing according to some exemplary embodiments of this disclosure are shown; Figure 3Band 3C Example communication processes that allow model / function sharing according to some example embodiments of this disclosure are shown; Figure 3D Example illustrations showing that model / function sharing is not permitted according to some example embodiments of this disclosure; Figure 3E Example communication processes that do not allow model / function sharing are shown according to some example embodiments of this disclosure; Figure 3F Example communication processes related to cell reselection are shown according to some example embodiments of this disclosure; Figure 4 A second signaling flow between a terminal device and a first network device according to some example embodiments of the present disclosure is shown; Figure 5 A third signaling flow between a terminal device and a first network device is shown according to some example embodiments of the present disclosure; Figure 6 A fourth signaling flow between a terminal device and a first network device according to some example embodiments of the present disclosure is shown; Figure 7 A flowchart is shown illustrating a method implemented at a terminal device according to some embodiments of the present disclosure; Figure 8 A flowchart is shown illustrating a method implemented at a first network device according to some embodiments of the present disclosure; Figure 9 A flowchart is shown illustrating a method implemented at a terminal device according to some embodiments of the present disclosure; Figure 10 A flowchart is shown illustrating a method implemented at a first network device according to some embodiments of the present disclosure; Figure 11 A flowchart is shown illustrating a method implemented at a terminal device according to some embodiments of the present disclosure; Figure 12 A flowchart is shown illustrating a method implemented at a first network device according to some embodiments of the present disclosure; Figure 13 A flowchart is shown illustrating a method implemented at a terminal device according to some embodiments of the present disclosure; Figure 14 A flowchart is shown illustrating a method implemented at a first network device according to some embodiments of the present disclosure; Figure 15 A simplified block diagram of a device suitable for implementing some example embodiments of this disclosure is shown; and Figure 16 A block diagram illustrating an example of a computer-readable medium according to some exemplary embodiments of the present disclosure is shown.

[0042] Throughout the accompanying drawings, the same or similar reference numerals denote the same or similar elements. Detailed Implementation

[0043] The principles of this disclosure will now be described with reference to some exemplary embodiments. It should be understood that these embodiments are described for illustrative purposes and to assist those skilled in the art in understanding and implementing this disclosure, without implying any limitation on the scope of this disclosure. The disclosure described herein can be implemented in various ways other than those described below.

[0044] In the following description and claims, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

[0045] References to "an embodiment," "embodiment," "example embodiment," etc., in this disclosure indicate that the described embodiment may include a particular feature, structure, or characteristic, but not every embodiment needs to include that particular feature, structure, or characteristic. Furthermore, such phrases do not necessarily refer to the same embodiment. Moreover, when a particular feature, structure, or characteristic is described in connection with an embodiment, whether explicitly described or not, it is believed that its influence on such feature, structure, or characteristic in conjunction with other embodiments is within the knowledge of those skilled in the art.

[0046] It should be understood that although the terms “first” and “second” may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used only to distinguish one element from another. For example, without departing from the scope of the exemplary embodiments, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element. As used herein, the term “and / or” includes any and all combinations of one or more of the listed terms.

[0047] The terminology used herein is for the purpose of describing particular embodiments and is not intended to limit the exemplary embodiments. As used herein, the singular forms “a,” “an,” and “the” are also intended to include the plural forms unless the context clearly indicates otherwise. It will be further understood that the terms “comprising,” “including,” “having,” “containing,” and / or “comprising” as used herein specify the presence of stated features, elements, and / or components, etc., but do not exclude the presence or addition of one or more other features, elements, components, and / or combinations thereof. As used herein, “at least one of the following: ” and “at least one of ” and similar wording, wherein the list of two or more elements is connected by “and” or “or”, means at least any one of the elements, or at least any two or more of the elements, or at least all of the elements.

[0048] As used in this application, the term "circuit system" may refer to one or more of the following: (a) Hardware circuit implementation only (such as implementation in analog and / or digital circuits only), and (b) A combination of hardware circuitry and software, such as (if applicable): (i) A combination of (multiple) analog and / or digital hardware circuits having software / firmware, and (ii) Any part of the (multiple) hardware processors having software (including (multiple) digital signal processors, software, and (multiple) memories working together to enable a device (such as a mobile phone or server) to perform various functions), and (c) (Multiple) hardware circuits and / or (multiple) processors, such as (multiple) microprocessors or a portion thereof, which require software (e.g., firmware) for operation, but the software may not exist when it is not required for operation.

[0049] This definition of circuit system applies to all uses of the term in this application, including in any claim. As another example, as used in this application, the term circuit also covers only hardware circuitry or a processor (or multiple processors) or a portion of hardware circuitry or a processor and its accompanying software and / or firmware implementation. For example, and if applicable to a particular claim element, the term circuit also covers baseband integrated circuits or processor integrated circuits for mobile devices or similar integrated circuits in servers, cellular network devices or other computing or network devices.

[0050] As used herein, the term "communication network" refers to a network that conforms to any suitable communication standard, such as New Radio (NR), Long Term Evolution (LTE), LTE-A Advanced (LTE-A), Wideband Code Division Multiple Access (WCDMA), High-Speed ​​Packet Access (HSPA), Narrowband Internet of Things (NB-IoT), etc. Furthermore, communication between terminal devices and network devices in a communication network can be performed according to any suitable generation of communication protocols, including but not limited to third-generation (3G), 4G, 4.5G, 5G, or 6G communication protocols, and / or any other currently known or future-developed protocols. Embodiments of this disclosure can be applied to various communication systems. Given the rapid development of communications, there will naturally be future types of communication technologies and systems that can implement this disclosure. This disclosure should not be construed as limiting its scope to the systems described above.

[0051] As used herein, the term "network device" refers to a node in a communication network through which terminal devices can access the network and receive services. Network devices can refer to base stations (BS) or access points (APs), such as Node B (NodeB or NB), Radio Access Network (RAN) nodes, Evolved Node B (eNodeB or eNB), NR NB (also known as gNB), Remote Radio Unit (RRU), Radio Headend (RH), infrastructure equipment for V2X (Vehicle-to-Everything) communication, Transmit and Receive Point (TRP), Receive Point (RP), Remote Radio Headend (RRH), repeater, Integrated Access and Backhaul (IAB) node, low-power nodes (such as femtoBS, picoBS, etc.), depending on the terminology and technology used.

[0052] The term "terminal device" refers to any terminal device capable of wireless communication. As an example and not a limitation, a terminal device may also be referred to as a communication device, user equipment (UE), subscriber station (SS), portable subscriber station, mobile station (MS), or access terminal (AT). Terminal devices can include, but are not limited to, mobile phones, cellular phones, smartphones, Voice over IP (VoIP) phones, wireless local loop phones, tablets, wearable terminal devices, personal digital assistants (PDAs), portable computers, desktop computers, image capture terminal devices (such as digital cameras), gaming terminal devices, music storage and playback devices, in-vehicle wireless terminal devices, wireless endpoints, mobile stations, laptop embedded devices (LEEs), laptop devices (LMEs), USB dongles, smart devices, wireless customer premises equipment (CPEs), Internet of Things (IoT) devices, watches or other wearable devices, head-mounted displays (HMDs), vehicles, drones, medical devices and applications (e.g., remote surgery), industrial devices and applications (e.g., robots and / or other wireless devices operating in the context of industrial and / or automated processing chains), consumer electronics devices, devices operating on commercial and / or industrial wireless networks, etc. In the following description, the terms "terminal equipment", "communication equipment", "terminal", "user equipment" and "UE" are used interchangeably.

[0053] The principles and implementation of embodiments of this disclosure will now be described in detail with reference to the accompanying drawings. First, refer to... Figure 1A It illustrates an example environment 100 in which example embodiments of the present disclosure may be implemented.

[0054] Environment 100, which may be part of a communication network, includes terminal device 110 and network device 120 (also referred to as first network device 120) communicating with each other. Environment 100 may include one or more network devices 130-1, 130-2, ..., 130-N (collectively or individually referred to as network device 130 or second network device 130), where N represents any suitable positive integer. Terminal device 110 and / or first network device 120 may communicate with second network device 130. As an example, one or more of terminal device 110, first network device 120, and second network device 130 may communicate with... Figure 1A Communicates with one or more other devices not shown in the diagram.

[0055] To transmit data and / or control information, terminal device 110 may perform communication with first network device 120 or second network device 130. The link from first network device 120 or second network device 130 to terminal device 110 is referred to as a downlink (DL), while the link from terminal device 110 to first network device 120 or second network device 130 is referred to as an uplink (UL).

[0056] In some example embodiments, the first network device 120 may be the primary network device (also known as the serving network device) of the terminal device 110, and the second network device 130 may be a neighboring network device (also known as a non-serving network device), such as a target set of neighboring network devices. As an example implementation, the terminal device 110 may perform cell (re)selection, and the (re)selected cell may be provided by the second network device 130. As another example implementation, the first network device 120 may switch the terminal device 110 to the second network device 130.

[0057] Despite Figure 1A The communication environment 100 describes terminal device 110, first network device 120, and second network device 130; however, embodiments of this disclosure can be applied to any other suitable communication devices communicating with each other. That is, embodiments of this disclosure are not limited to... Figure 1A An exemplary scenario. In this regard, it should be noted that, although in Figure 1A The terminal device is schematically depicted as a mobile phone, and the first network device 120 and the second network device 130 are schematically depicted as base stations. It should be understood that these depictions are exemplary in nature and do not imply any limitation. In other embodiments, the first device 110, the first network device 120, and the second network device 130 may be any other communication device, such as any other wireless communication device.

[0058] It should be understood that, such as Figure 1A The specific number of various communication devices and communication links shown are for illustrative purposes only and do not imply any limitation. Communication environment 100 may include any suitable number of communication devices and any suitable number of communication links for implementing embodiments of this disclosure. Furthermore, it should be understood that various wireless and wired communications may be present among all communication devices (if desired).

[0059] Communications in Environment 100 may follow any suitable communication standards or protocols that are already in use or will be developed in the future, such as Universal Mobile Telecommunications System (UMTS), Long Term Evolution (LTE), LTE-Advanced (LTE-A), 5G New Radio (NR), 6G, Wi-Fi, and Global Microwave Access Interoperability (WiMAX) standards, and employ any suitable communication technologies, including, for example, Multiple-Input Multiple-Output (MIMO), Orthogonal Frequency Division Multiplexing (OFDM), Time Division Multiplexing (TDM), Frequency Division Multiplexing (FDM), Code Division Multiplexing (CDM), Bluetooth, ZigBee and Machine-Type Communications (MTC), Enhanced Mobile Broadband (eMBB), Massive Machine-Type Communications (mMTC), Ultra-Reliable Low-Latency Communications (URLLC), Carrier Aggregation (CA), Dual Connectivity (DC), and New Radio Unlicensed (NR-U) technologies.

[0060] At Radio Access Network Working Group 4 (RAN4) #109 meeting, Release 18 (Rel-18) of the AI / ML study for the NR air interface (FS_NR_AIML_Air) was completed, and the results, protocols, and outstanding issues are documented in Technical Report (TR) 38.843 (Version 18.0.0).

[0061] In addition, a new Release 19 (Rel-19) work item (WI) on AI / ML for the NR air interface (NR_AIML_Air) was approved at RAN#102 to begin normative work on a general AI / ML framework for the air interface and to implement the recommended use cases in previous studies.

[0062] Generalization of AI / ML models / features is identified as one of the major challenges supporting AI / ML-enabled use cases. Generalization refers to the ability of a model and / or feature to appropriately adapt to new or previously unseen data drawn from the same distribution used to create the model and / or feature. In other words, generalization checks how well a model / feature can digest new data (primarily corresponding to new environments and / or scenarios) and make correct predictions (for unseen / new environments or scenarios) after training on the training set. If a model is trained too well on the training data, it will fail to generalize and will eventually make incorrect predictions when given new data. This will render the model ineffective even if it is able to make correct predictions on the training dataset. This is known as overfitting. In practice, encountering different radio conditions or parameter settings can severely impact the performance of a model and / or feature if it has already been trained for a given radio condition or parameter setting.

[0063] Furthermore, some objectives related to the use case of Channel State Information (CSI) compression have been discussed. In the current 3GP PCSI feedback framework, the Rank Indicator (RI), Precoding Matrix Indicator (PMI), and Channel Quality Indicator (CQI) can be jointly reported to the gNB according to the configuration provided by the gNB. In the case of subband reporting, CQI feedback may require more resources. For codebook-based solutions, the UE determines the CQI used for reporting based on the precoding matrix indicated by the PMI and its associated receiver.

[0064] With the introduction of multi-user multiple-input multiple-output (MU-MIMO) systems, the overhead required to send high-resolution CSI feedback at high rank in the uplink has increased many times over. Therefore, AI / ML-based solutions can help reduce overhead by compressing CSI reports. Various approaches exist that utilize two-sided models (e.g., autoencoders) to achieve this using AI / ML-based solutions.

[0065] Autoencoders are based on unsupervised learning techniques, where bottlenecks are imposed on the network to force a compressed knowledge representation of the original input. The main challenge remains the reconstruction of the original input. Figure 1B The architecture of an autoencoder is shown in the diagram. By definition, an autoencoder addresses the problem of CSI feedback compression.

[0066] Before training the autoencoder, four hyperparameters and other parameters need to be set: 1) codeword / bottleneck size, 2) number of layers, 3) number of nodes per layer, and 4) the loss function to be used, such as mean squared error (MSE) or cosine similarity. The number of nodes per layer typically decreases with each subsequent layer of the encoder and increases back in the decoder. The decoder is symmetrical to the encoder in terms of layer structure.

[0067] like Figure 1B As described, an autoencoder consists of three parts: 1) an encoder, 2) a bottleneck (here, the codeword), and 3) a decoder. The encoder is designed to process the input data (e.g., the channel matrix) into a decoder. The data (or feature vectors) is compressed into codewords with a dimension smaller than the original information. The bottleneck (e.g., the codeword) is the compressed representation of the original information. Following the bottleneck is the decoder, which decompresses the codewords and reconstructs the data (i.e., recovers the information). The module. Then, and The comparison shows that this is a lossy process, and the recovered matrix... It is possible to not be with same.

[0068] The aforementioned AI / ML-enabled CSI compression use case for cellular communications features a cascaded, two-sided model. One model operates on the UE side to compress CSI, while the other model operates on the gNB side to decompress the compressed CSI received from the UE. When deploying AI / ML-enabled features involving a two-sided model, the models at the UE and / or gNB can be updated frequently and independently at different times. Alignment between the UE and gNB-side models is necessary whenever there is an update to the model or a change to the associated gNB to ensure interoperability and performance.

[0069] Specifically, in UE mobility scenarios, multiple UE-side models may need to interoperate with multiple different gNB-side models that the UE will have to operate after mobility activities. However, the UE may typically not have any prior information about the multiple network-side models. Therefore, interoperability and joint performance between the UE-side and network-side models / functions cannot be guaranteed, as the UE and / or network-side models / functions can vary across different gNBs based on deployment and vendor implementation.

[0070] Therefore, for UE mobility scenarios, it becomes crucial for the UE and gNB (i.e., the primary gNB and (multiple) neighboring gNBs) to know / evaluate in advance the interoperability and performance characteristics of their corresponding models / functions for dual-side AI / ML enabling. However, currently, there is no efficient method that allows the UE and gNB to determine the interoperability and performance characteristics of their corresponding models / functions. Given the above, how to support the UE and gNB in ​​determining the interoperability and joint performance of UE and network-side models / functions is an important problem to be solved.

[0071] According to embodiments of this disclosure, a first solution to the above-mentioned problems is provided. Using this solution, a terminal device sends first information to a first network device regarding the output of one or more first machine learning models available at the terminal device. Furthermore, the terminal device receives from the first network device second information regarding the performance of one or more second machine learning models available at one or more other network devices. The performance is obtained based on the first information. Additionally, the terminal device performs cell selection or cell reselection based at least on the second information.

[0072] According to embodiments of this disclosure, a second solution to the above-mentioned problems is provided. Using this solution, a terminal device sends first information to a first network device regarding the output of one or more first machine learning models available at the terminal device. Furthermore, the terminal device receives second information from the first network device regarding the performance of one or more second machine learning models available at a second network device. The performance is obtained based on the first information. Additionally, the terminal device performs a handover from the first network device to the second network device.

[0073] According to embodiments of this disclosure, a third solution to the above-described problems is provided. Using this solution, the terminal device sends first information about the outputs of one or more first machine learning models available at the terminal device to a first network device based on changes in channel conditions associated with the terminal device. Furthermore, the terminal device receives second information from the first network device about the performance of one or more second machine learning models available at one or more other network devices. The performance is obtained based on the first information.

[0074] According to embodiments of this disclosure, a fourth solution to the above-mentioned problems is provided. Using this solution, a terminal device, based on updates to a machine learning model at the terminal device, sends first information to a first network device regarding the output of one or more first machine learning models available at the terminal device. Furthermore, the terminal device receives from the first network device second information regarding the performance of one or more second machine learning models available at one or more other network devices. The performance is obtained based on the first information.

[0075] By feeding back model / function performance information to the terminal device, the terminal device can know the performance of the models / functions available at a second network device (e.g., a neighboring cell (gNB)). In this way, the terminal device can optimize mobility processes across different / target network devices based on the second information, for example, making better mobility decisions in idle mode cell reselection / handover scenarios. Therefore, it allows for enhanced communication performance and improved communication efficiency.

[0076] In this disclosure, the terms “model” and “function” are used interchangeably in some cases.

[0077] In this disclosure, the terms “machine learning (ML) model” and “artificial intelligence (AI) model” are used interchangeably in some cases.

[0078] In this disclosure, the terms “performance,” “inference,” and “test results” are used interchangeably in some cases.

[0079] In this disclosure, the terms "interoperability check" and "performance verification" may be used interchangeably in some cases.

[0080] Figure 2 A first signaling flow 200 between a terminal device and a first network device according to some example embodiments of the present disclosure is shown. For discussion purposes, reference will be made to... Figure 1A Describe signaling flow 200.

[0081] like Figure 2As shown, terminal device 110 sends (205) information (also known as first information) to first network device 120 regarding the output of one or more machine learning models (also known as the first one or more machine learning models) available at terminal device 110. First network device 120 may include serving terminal network device 110 or a master network device to which terminal device 110 is locked.

[0082] In some example embodiments, the first one or more machine learning models can be used as encoders for encoding CSI, and in this case, the output can correspond to the CSI encoded by the first one or more machine learning models. In other words, the output can include one or more compressed CSIs corresponding to the first one or more machine learning models. For example, if the terminal device 110 is capable of storing / using multiple AI / ML models / functions for two-sided use cases, the output can include multiple compressed CSIs corresponding to multiple terminal-side models / functions. The first one or more machine learning models can also be used for any other purpose (e.g., as an encoder for encoding any other information, as a terminal-side model portion of another two-sided model use case, etc.), and the scope of this disclosure is not limited in this respect.

[0083] In some example embodiments, terminal device 110 may send an indication to first network device 120 regarding whether terminal device 110 is capable of using a first or more machine learning models (e.g., for a two-sided model use case). For example, terminal device 110 may indicate this advanced capability to first network device 120 via, for example, a UE capability message before sending first information to first network device 120.

[0084] In some example embodiments, the transmission of first information can be used to trigger the first network device 120 to obtain the performance of one or more machine learning models (also referred to as second or more machine learning models) available at the second network device 130. In this case, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models. The performance of the second or more machine learning models can be obtained based on the first information. The second network device 130 may include adjacent network devices for the terminal device 110. Alternatively or additionally, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models and one or more machine learning models available at the first network device 120. It should be understood that although embodiments of this disclosure have been discussed in the case where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130 relative to the first or more machine learning models, embodiments of this disclosure can also be applied to cases where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130 and one or more machine learning models available at the first network device 120 relative to the first or more machine learning models. The scope of this disclosure is not limited in this respect. Further details will be discussed in more detail below.

[0085] In some example embodiments, performance may include the combined (i.e., overall) performance / compatibility (e.g., interoperability characteristics) of a second or more machine learning models of the second network device 130 and a first or more machine learning models at the terminal device 110. As an example implementation, performance may be relative to the current model performance of the first network device 120, for example, relative to the performance of the current cell(s) models(s).

[0086] In an example embodiment where the first or more machine learning models are used as an encoder for encoding CSI, a second or more machine learning models can be used as a decoder for decoding the CSI encoded by the first or more machine learning models. In this case, performance can include the joint performance (inference) of both the terminal-side model / function and the network-side model / function (i.e., the encoder-decoder pair). For example, performance can be associated with performance indicators for decoding the encoded CSI, such as decoding accuracy, decoded CQI, etc.

[0087] The transmission of the first information used to trigger the initiation of model / functional validation tests associated with a second or more machine learning models can be triggered in various ways. For example, the transmission of the first information can be triggered from the terminal side. As another example, the transmission of the first information can be triggered from the network side.

[0088] In some example embodiments where the transmission of first information is triggered from the terminal side, terminal device 110 may send the first information based on an update of the machine learning model at terminal device 110. For example, when a model update occurs at terminal device 110, terminal device 110 may trigger first network device 120 to obtain the performance of a second or more machine learning models available at second network device 130; in other words, trigger a model / function verification test. In this case, when a model / function update exists at terminal device 110, terminal device 110 may need to use the updated model / function. Terminal device 110 may then send an instruction to first network device 120 (e.g., a message carrying the first information, for example, along with additional reasoning information, which will be discussed in detail below) to trigger a model / function verification test against the updated model(s) available at terminal device 110 and the network-side model(s) available at second network device 130.

[0089] In some example embodiments where the transmission of the first information is triggered from the terminal side, terminal device 110 may transmit the first information based on a change in channel conditions associated with terminal device 110. The change in radio conditions may be associated with the mobility of terminal device 110. In some implementations, terminal device 110 may perform cell / beam-level radio measurements (e.g., Layer 1 (L1) / Layer 3 (L3) Reference Signal Received Power (RSRP)) in the primary / serving cell. When terminal device 110 determines a decrease in the measured RSRP level (e.g., a decrease in the RSRP level associated with the primary cell) (e.g., a decrease compared to a previous RSRP measurement, falling below a threshold), it may send an indication (e.g., a message carrying the first information, such as additional cause information, which will be discussed in detail below) to first network device 120 to trigger model / functional verification testing. In some implementations, terminal device 110 may perform radio measurements (e.g., RSRP measurements) associated with second network device 130. When terminal device 110 determines that the measured RSRP level corresponding to a neighboring cell of second network device 130 has increased, terminal device 110 may send an indication (e.g., a message carrying first information, such as additional cause information, which will be discussed in detail below) to first network device 120 to trigger a model / functional verification test. Alternatively or additionally, this trigger may also be generated by the determination that terminal device 110 is moving toward the cell edge.

[0090] In some example embodiments where the transmission of the first information is triggered from the terminal side, the terminal device 110 may send the first information based on moving to a new cell. For example, whenever the terminal device 110 moves to a new cell, the new cell may become the primary cell, and the old primary cell may become one of one or more neighboring cells. In this scenario, the old set of neighboring cells may change because some cells in the old set may no longer be valid neighbors, and some new neighboring cells are added to the set of neighboring cells. In this case, the terminal device 110 may send an indication (e.g., a message carrying the first information, for example, along with additional reason information, which will be discussed in detail below) to the first network device 120 to trigger model / functional verification testing.

[0091] In some example embodiments where the transmission of the first information is triggered from the network side, the first network device 120 may send an instruction to the terminal device 110 to send the first information, and then the terminal device 110 may send the first information accordingly, as will be discussed below.

[0092] In some implementations, when a machine learning model / feature update / change occurs at a first network device 120 or a second network device 130 (e.g., one or more network devices 130-1, 130-2, ..., 130-N), the first network device 120 may instruct the terminal device 110 to send a first message to trigger the initiation of a model / feature verification test. In this case, when a model / feature update exists at the first network device 120 or the second network device 130, the corresponding network device may need to use the updated network-side model / feature(s). Then, the network device (e.g., the first network device 120 or the second network device 130) can trigger a model / feature verification test for the updated network-side model / feature(s) relative to the terminal-side model / feature(s).

[0093] In some implementations, when the first network device 120 determines that the terminal device 110 is moving toward the cell edge, it can instruct the terminal device 110 to send a first message to trigger the initiation of a model / functional verification test.

[0094] In some example embodiments, terminal device 110 may send reason information to first network device 120, such as an indication of the reason for sending the first indication. The reason may include at least one of the following: an update of the machine learning model at terminal device 110; a change in channel conditions associated with terminal device 110; or terminal device 110 moving to a new cell.

[0095] In some example embodiments, the first information and the aforementioned cause information (e.g., an indication of cause) may be included in a signaling message, such as a higher-level signaling message, like Radio Access Control (RRC) signaling. As an example, the message may include a network_GNB_MODEL_PER_Request(input_payload, cause). This message may be intended to be used by terminal device 110 to trigger the first network device 120 to obtain the performance of a second or more machine learning models, or in other words, to trigger a model / function performance test on the pool of models / functions available at the second network device 130. The "input_payload" field may include the input payload to the second or more machine learning models / functions available at the second network device 130. For example, the input payload may include the output of a terminal-side model (e.g., compressed CSI provided as output by a terminal-side encoder) or a processed version of the terminal-side model's output. The “Reason” field can be intended to indicate the reason for initiating NEIGHBOR_GNB_MODEL_PER_REQUEST, such as machine learning model update at terminal device 110 (e.g., UE-side model update), change of channel conditions associated with terminal device 110, or terminal device 110 moving to a new cell (e.g., notification area (RNA) based on a new RAN).

[0096] Upon receiving the first information from terminal device 110, the first network device 120 can be triggered to obtain the performance (i.e., inference / test results) of a second or more machine learning models. Then, as... Figure 2 As shown, the first network device 120 sends (210) information (also referred to as second information) about the performance of a second or more machine learning models to the terminal device 110. For example, the performance of the second or more machine learning models may include the ranked joint performance of the terminal-network side model pairs. Alternatively or additionally, the information about the performance of the second or more machine learning models may be sent to the second network device 130 (i.e., shared with the second network device 130).

[0097] In some example embodiments, the second information may be included in a signaling message, such as a higher-level signaling message, like RRC signaling. As an example, the message may include NEIGHBOR_GNB_MODEL_PERFORMANCE_REPORT(gnb_model_perf_info). This message may be intended to be used by the first network device 120 to send a merged / related report of the performance of a second or more machine learning models / features in the pool of the second network device 130 (and / or the first network device 120) to the terminal device 110. The “gnb_model_perf_info” field may include model / feature performance information for the second or more machine learning models (and / or one or more models of the first network device 120) of the second network device 130. The merged report can then be used to optimize AI / ML model / feature-related model lifecycle management (LCM) signaling during mobility scenarios.

[0098] Whether sharing of AI / ML model / function related information (e.g., model / function identifier (ID) used to identify multiple models / functions) / files (e.g., one or more model / function parameters or complete model / function files) (e.g., sharing of AI / ML model / function related information / files between gNBs) is permitted / supported between the first network device 120 and the second network device 130, the methods by which the first network device 120 obtains the performance of the second or more machine learning models can be different.

[0099] In some example embodiments, AI / ML model / function-related information / file sharing may be permitted between the first network device 120 and the second network device 130, and in this case, the first network device 120 may be aware of a second or more machine learning models running at the second network device 130. For example, AI / ML model / function-related information / file sharing may be supported if the first network device 120 and the second network device 130 are from the same vendor. As another example, if the first network device 120 and the second network device 130 are from different vendors, AI / ML model / function-related information / file sharing may be supported if the model / function-related information / file sharing between the different vendors is agreed upon, for example, in 3GPP. As yet another example, AI / ML model / function-related information / file sharing may be supported if the AI / ML model / function is standardized in 3GPP to assist the terminal device 110 during mobility scenarios.

[0100] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the performance of the second or more machine learning models can be determined by the first network device 120 itself. For example, the second or more machine learning models / features in the pool of the second network device 130 can be the same as one or more in the first network device 120, and therefore the required one or more models / features of the second network device 130 may already be available at the first network device 120. As another example, the model / feature pools at the second network device 130 and the first network device 120 are different, and in this case, the model / feature(s) files of the second network device 130 can be shared with the first network device 120. When AI / ML model sharing is allowed, the first network device 120 can perform corresponding interoperability checks (i.e., performance verification) at its end. For example, the first network device 120 can evaluate / verify the joint performance of the first or more machine learning models / features (e.g., encoders) with each of the second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130.

[0101] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the first network device 120 may query and / or collect information and / or files of a second or more machine learning models / features present in a pool of the second network device 130. The first network device 120 may run inference on the second or more machine learning models / features to evaluate the interoperability performance of the terminal-side(multiple) models / features with the network-side(multiple) models / features available at the second network device 130. Alternatively or additionally, the first network device 120 may evaluate the interoperability performance of the terminal-side(multiple) models / features with the network-side(multiple) models / features available at the first network device 120.

[0102] Figure 3A An example illustration showing how model / feature sharing is allowed is shown. For example... Figure 3A The UE shown can be an example implementation of terminal device 110, such as... Figure 3A The gNB1 shown can be an example implementation of the first network device 120, such as... Figure 3A The gNB2 and gNB3 shown can be example implementations of network devices 130-1 and 130-2, respectively. In this case, gNB1 can be the primary gNB for the UE, while gNB2 and gNB3 are adjacent gNBs.

[0103] like Figure 3AAs shown, the UE is using the activated model / function E1 (e.g., encoder E1) and can trigger the main gNB (i.e., gNB1) with AI / ML models / functions (e.g., decoders) D1, D2, D3 in its pool to run / initiate (multiple) joint performance (inference) tests for dual-side models / functions (e.g., encoder-decoder pairs) for UE-side models / functions and network-side models / functions, wherein the network-side model / function pool contains the models / functions of the main gNB1 and / or the models / functions existing in the pools of adjacent gNBs (i.e., gNB2 and gNB3), wherein the model / function pools of gNB2 and gNB3 contain (D4, D5, D6) and (D7, D8, D9) respectively.

[0104] Models / functions or related information (e.g., model / function identifiers (IDs) used to identify models / functions) can be sent from each of the neighboring gNBs (i.e., gNB2 and gNB3) to the primary gNB (i.e., gNB1). For example, gNB2 can send the complete file of a model in its model pool or the ID of a model in its model pool (i.e., D5, D6, D4) to gNB1. For example, gNB3 can send the complete file of a model in its model pool or the ID of a model in its model pool (i.e., D7, D8, D9) to gNB1. gNB1 can then perform inference on the output of the UE-side model / function using each of the models / functions in the primary gNB1 and the neighboring gNBs (i.e., gNB2, gNB3) locally in the primary gNB1. The primary gNB1 can run inference locally, and the corresponding inference / performance reports(s) can be sent back to the neighboring gNBs (i.e., gNB2, gNB3) and / or the UE.

[0105] Now for reference Figure 3B and 3C Let's discuss an example communication process that allows for model / feature sharing. For example... Figure 3B and 3C The UE shown can be an example implementation of terminal device 110, such as... Figure 3B and 3C The gNB1 shown can be an example implementation of the first network device 120, such as... Figure 3B and 3C The gNB2 and gNB3 shown can be example implementations of network devices 130-1 and 130-2, respectively. gNB1 can be the UE's primary gNB, while gNB2 and gNB3 are neighboring gNBs.

[0106] First refer to Figure 3BIn this scenario, the model / function (i.e., decoder) pools at adjacent gNBs (i.e., gNB2 and gNB3) and the primary gNB (i.e., gNB1) are identical. Specifically, all gNBs in gNB1, gNB2, and gNB3 have the same decoder pools – D1, D2, and D3 – and the UE locked to gNB1 has encoder models / functions E1 and E2. The dual-side models / functions on both the terminal and network sides are associated with AI / ML-enabled CSI compression and CSI decompression.

[0107] like Figure 3B As shown, in step 1, to evaluate the interoperability performance of the UE-side model / function (i.e., encoders E1 and E2) with the network-side model / function (i.e., decoders D1, D2, and D3) of the primary gNB1 and / or adjacent gNB2 and gNB3, the UE uses NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST to trigger a model / function verification test. As part of the test, the UE uses its available encoders E1 and E2 to compress the CSI feedback and sends the compressed outputs of E1 and E2 to gNB1.

[0108] In step 2, gNB1 receives the compressed CSI sent by the UE. Since the models / functions in gNB1's pool are the same as those in adjacent gNB2 and gNB3, the required models / functions are already available at gNB1.

[0109] In step 3, gNB1 runs model / function reasoning / verification tests on the same compressed CSI feedback received from the UE in step 1 for all / related models / functions in the pool by applying each encoded CSI as input to the AI / ML decoders of the target / neighboring gNBs (i.e., gNB2 and gNB3, whose AI / ML decoders are the same as gNB1's). Based on the decoded CSI, gNB1 evaluates the (joint) performance / compatibility of the AI / ML decoders of the target gNB2 and gNB3 with the encoder at the UE.

[0110] In step 4, once the (multiple) inference performance results / reports are available, gNB1 sends the (multiple) joint performance / compatibility reports to the UE using NEIGHBOR_GNB_MODEL_PERFORMANCE_REPORT, and / or sends the (multiple) joint performance / compatibility reports to neighboring gNB2 and gNB3.

[0111] Now for reference Figure 3CIn this scenario, adjacent gNB2 and gNB3 have different model / function (decoder) pools at the primary gNB1; however, model / function file sharing between adjacent gNB2 and gNB3 and the primary gNB1 is allowed. For example, this scenario can focus on deployments with localized models / functions as well as deployments from gNBs from different vendors. The model / function pools for gNB1, gNB2, and gNB3 are (D1, D2, D3), (D5, D6, D4), and (D7, D8, D9), respectively. UEs locked to gNB1 have encoder models / functions E1 and E2. The dual-side models / functions on both the terminal and network sides are associated with AI / ML-enabled CSI compression and CSI decompression.

[0112] like Figure 3C Step 1 shown is similar to... Figure 3B Step 1 shown is the same. For the sake of simplicity, details will be omitted.

[0113] like Figure 3C Step 2 shown Figure 3B Step 2 is the same as shown, except that the model / function (i.e., one or more specific model / function parameters, such as a complete model / function file) needs to be shared by adjacent gNB2 and gNB3 to the main gNB1, because the corresponding pools of adjacent gNB2 and gNB3 are different from the corresponding pool of the main gNB1, but model / function file sharing is allowed.

[0114] like Figure 3C Step 3 shown is similar to... Figure 3B Step 3 is the same as shown. For the sake of simplicity, details will be omitted.

[0115] like Figure 3C Step 4 shown is similar to... Figure 3B Step 4 is the same as shown. For the sake of simplicity, details will be omitted.

[0116] In some other example embodiments, sharing of AI / ML model / feature-related information / files may not be permitted between the first network device 120 and the second network device 130, and in such cases, the first network device 120 may be unaware of the model / features of the second network device 130. For example, if the first network device 120 and the second network device 130 are from two or more vendors, sharing of AI / ML model / feature-related information / files may not be supported, thereby limiting inter-vendor model sharing for reasons such as proprietary and security.

[0117] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may query / request the performance (i.e., performance verification / inference) of a second or more machine learning models from the second network device 130. For example, the first network device 120 may query / request the joint performance of a first or more machine learning models / features (e.g., encoders) with each of a second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130. In this case, the first network device 120 may query / request a corresponding interoperability check (i.e., joint performance verification) between the first or more machine learning models available on the terminal side and the second or more machine learning models available from the second network device 130 on the network side.

[0118] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may forward compressed input received from the terminal device 110 to the second network device 130 to allow the second network device 130 to determine the interoperability performance of(multiple) terminal-side models / features with their corresponding(multiple) network-side models / features. The first network device 120 may collect inference / test results and then share them with the terminal device 110.

[0119] Now, let's refer to 3D to discuss example illustrations illustrating the disallowed sharing of models / features. For example... Figure 3D As shown, the UE can be an example implementation of terminal device 110, gNB1 can be an example implementation of the first network device 120, and gNB2 and gNB3 can be example implementations of network devices 130-1 and 130-2, respectively. The UE is using an activated model / function E1 (e.g., encoder E1) and can trigger the main gNB (i.e., gNB1) with AI / ML models / functions (e.g., decoders) D1, D2, D3 in its pool to run / initiate (multiple) joint performance (inference) tests for dual-side models / functions (e.g., encoder-decoder pairs) of the UE-side model / function and the network-side model / function, wherein the network-side model / function pool contains the models / functions of the main gNB1 and / or the models / functions existing in the pools of adjacent gNBs (i.e., gNB2 and gNB3), wherein the model / function pools of gNB2 and gNB3 contain (D4, D5, D6) and (D7, D8, D9), respectively.

[0120] like Figure 3DAs shown, the primary gNB (i.e., gNB1) can provide outputs received from the UE (e.g., compressed CSI) to neighboring gNBs (gNB2 and gNB3). Joint inference can be performed locally at the corresponding gNB2 and gNB3 using models / functions from the corresponding model / function pools of the neighboring gNBs (i.e., D4, D5, D6 for gNB2 and D7, D8, D9 for gNB3). Inference / performance reports can be shared back to gNB1. Once the inference / performance reports are available at the primary gNB1, the merged reports can be sent to the UE.

[0121] Figure 3E An example communication process that does not allow model / function sharing is shown. Figure 3E The UE shown can be an example implementation of terminal device 110. Figure 3E The gNB1 shown can be an example implementation of the first network device 120. Figure 3E The gNB2 and gNB3 shown can be example implementations of network devices 130-1 and 130-2, respectively. In this case, gNB1 can be the UE's primary gNB, while gNB2 and gNB3 are neighboring gNBs.

[0122] exist Figure 3E In this scenario, model / function information sharing may or may not be permitted (i.e., which one or more models are available at gNB2 and gNB3 and can be shared with the main gNB1), however, actual model / function sharing (i.e., sharing of specific model parameters / files) is not permitted between adjacent gNBs and the main gNB1. In this case, the model / function pools at adjacent gNB2 and gNB3 are different from the model / function pool at gNB1. Each gNB uses a different decoder model / function pool - gNB1 (D1, D2, D3); gNB2 (D5, D6, D4); gNB3 (D7, D8, D9). The UE locked to gNB1 has encoder models / functions E1 and E2. The dual-side models / functions on the terminal side and network side are associated with AI / ML-enabled CSI compression and CSI decompression.

[0123] like Figure 3E Step 1 shown is similar to... Figure 3B Step 1 shown is the same. For the sake of simplicity, details will be omitted.

[0124] In step 2, gNB1 receives the compressed CSI sent by the UE and requests neighboring gNB2 and gNB3 to perform inference on the compressed CSI received in step 1 by forwarding the compressed CSI received from the UE to neighboring gNB2 and gNB3.

[0125] In step 3, adjacent gNB2 and gNB3 use the received input to run inference using the decoders in their model / function pools and send performance feedback back to the main gNB1.

[0126] like Figure 3E Step 4 shown is similar to... Figure 3B Step 4 is the same as shown. For the sake of simplicity, details will be omitted.

[0127] Return to reference Figure 2 After receiving the second information from the first network device 120, the terminal device 110 performs (215) cell selection or cell reselection based at least on the second information. In this case, based on the second information, the terminal device 110 can know the performance of a second or more machine learning models available at the second network device 130 for a given channel condition, and then, based on this, the terminal device can perform cell selection or cell reselection accordingly.

[0128] In some example embodiments, if terminal device 110 determines that the performance of at least one machine learning model available at at least one network device, as indicated by the second information, is acceptable, then terminal device 110 may select or reselect a cell of at least one of the network devices 130-1, 130-2, ..., 130-N for connection based on performance. Whether the performance of at least one machine learning model available at at least one network device is acceptable can be determined based on an acceptable level or threshold. For example, if the joint performance is above an acceptable level, then the joint performance is acceptable.

[0129] In some example embodiments, a corresponding network device among one or more network devices 130-1, 130-2, ..., 130-N may perform a system information broadcast indicating one or more machine learning models available at the corresponding network device. In this case, a list of models / functions (model / function IDs) available for the cell may be broadcast in the system information broadcast (e.g., system information block (SIB4), system information block (SIB5), etc.). Therefore, terminal device 110 may receive the system information broadcast and then select or reselect a cell based at least on the received system information broadcast. For example, if the machine learning model available at the corresponding network device among one or more network devices 130-1, 130-2, ..., 130-N is acceptable and has better joint performance relative to the terminal-side model(s) than one or more other network devices, and if the signal strength of the corresponding network device and one or more other network devices is the same or similar, then terminal device 110 may select or reselect a cell of the corresponding network device.

[0130] After cell selection or cell reselection, terminal device 110 may need to determine whether to use the activated machine learning model from the first or one more machine learning models. Depending on the different second pieces of information, the terminal's utilization behavior of the joint performance report can differ, as will be discussed below.

[0131] In some example embodiments, if the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models, then terminal device 110 may stop using the first or more machine learning models after cell selection or cell reselection. In this case, if no alternative / relatively better model is available for the first or more machine learning models, terminal device 110 may fall back to non-AI / ML behavior to avoid potential performance degradation after cell change. For example, in the case where terminal device 110 is a simple / basic UE equipped with only one or a few AI / ML models / functions, it may first indicate its limited model / function capabilities to first network device 120 through UE capabilities. Based on this capability, first network device 120 may decide to provide terminal device 110 with limited performance reports, such as a binary report mentioning whether joint execution is acceptable or unacceptable, to help terminal device 110 fall back to non-AI / ML behavior in advance (if needed) in a network-transparent or opaque manner.

[0132] In some example embodiments, if the second information indicates that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model, the terminal device 110 may continue to use the activated machine learning model after cell selection or cell reselection, and / or store the second information. In this case, if the joint performance is above an acceptable level, the same activated model / function may continue to be used after a cell change, and / or the performance results provided by the first network device 120 may be stored for future mobility optimization purposes. As an example implementation, when identifying the aforementioned acceptable level of joint performance, the terminal device 110 may store the ID / information of the terminal-side-network-side model pair (e.g., encoder-decoder model pair) and the ID(s) of the second network device 130, such that if the terminal device 110 will need to access the second network device 130 in the future (e.g., the same set / subset of one or more network devices 130-1, 130-2, ..., 130-N), it will not be necessary to re-evaluate the joint performance, and the stored information can be reused to save evaluation signaling and resources. For example, the model / function ID can be stored in, for example, a terminal device 110 or a subscriber identity module (SIM) card for future use.

[0133] In some example embodiments, if the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the performance of the activated machine learning model, and the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to at least one of the first or more machine learning models, then the terminal device 110 may stop using the activated machine learning model after cell selection or cell reselection, and begin using at least one of the first or more machine learning models after cell selection or cell reselection. In this case, based on the joint performance provided by the first network device 120, the terminal device 110 may switch its activated AI / ML model / function to another alternative / relatively better model / function (if available) of the first or more machine learning models to avoid potential performance degradation after cell change. For example, if the terminal device 110 is an advanced UE equipped with multiple AI / ML models / functions, it may first indicate its model / function switching capability to the first network device 120 through UE capabilities, and the first network device 120 may decide to provide the terminal device 110 with a detailed performance report so that it can switch to a better performing model / function (if needed) in a network-transparent or non-transparent manner.

[0134] Now for reference Figure 3F Let's discuss an example communication process 300 in which cell reselection occurs. It should be understood that process flow 300 can be considered as follows: Figure 2 A more specific example of signaling flow 200 is shown. For example... Figure 3F The UE shown can be an example implementation of terminal device 110, such as... Figure 3F The gNB1 shown can be an example implementation of the first network device 120, such as... Figure 3F The gNB2 and gNB3 shown can be example implementations of network devices 130-1 and 130-2, respectively. gNB1 and gNB2 have the same decoder model / function pool - D1, D2 and D3, while gNB3 has a different decoder model / function pool - D7, D8 and D9.

[0135] like Figure 3F As shown, in step 1, there is a trigger for any change relative to the two-sided model, such as a model update at the UE, a model update at the gNB, or an RNA change. In step 2, the UE sends a NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST message with compressed CSI input from the encoder to gNB1.

[0136] In step 3, gNB1 obtains model information from neighboring gNB2 and gNB3. If model sharing is allowed, inference can be performed locally on gNB1; otherwise, inference can be performed in the respective gNB. Responses can be evaluated, and responses from neighboring gNBs can be merged. gNB1 then prepares its response to the UE. In step 4, gNB1 sends a NEIGHBOR_GNB_MODEL_PERFORMANCE_REPORT message to the UE, displaying the performance of the model pool available at neighboring gNB2 and gNB3 (e.g., together with gNB1).

[0137] In step 5, the UE learns about the models available in neighboring gNB2 and gNB3 and their performance for a given channel condition. In step 6, during periodic updates or cell reselection processes, the UE reads SIB4 for neighboring cell information. In steps 7 and 8, gNB2 and gNB3 perform SIB broadcasts using the Physical Cell Identifier (PCI) and model information, respectively.

[0138] In step 9, after reading the SIB, the UE identifies two cells with similar signal strengths, PCI=2 and 4. In step 10, since models D7 and D8 associated with the cell with PCI=4 perform better, the UE selects the cell with PCI=4 instead of the cell with PCI=2.

[0139] The above reference Figure 2 The described operations and features also apply to process 300 and have similar effects. For the sake of simplicity, details will be omitted.

[0140] Figure 4 A second signaling flow 400 between a terminal device and a first network device according to some example embodiments of the present disclosure is illustrated. For discussion purposes, reference will be made to... Figure 1A Description of signaling flow 400.

[0141] like Figure 4 As shown, terminal device 110 sends (405) information (also known as first information) to first network device 120 regarding the output of one or more machine learning models (also known as the first one or more machine learning models) available at terminal device 110. First network device 120 may include a main network device serving terminal network device 110 or a main network device to which terminal device 110 is locked.

[0142] In some example embodiments, the first one or more machine learning models can be used as encoders for encoding CSI, and in this case, the output can correspond to the CSI encoded by the first one or more machine learning models. In other words, the output can include one or more compressed CSIs corresponding to the first one or more machine learning models. For example, if the terminal device 110 is capable of storing / using multiple AI / ML models / functions for two-sided use cases, the output can include multiple compressed CSIs corresponding to multiple terminal-side models / functions. The first one or more machine learning models can also be used for any other purpose (e.g., as an encoder for encoding any other information, as a terminal-side model portion of another two-sided model use case, etc.), and the scope of this disclosure is not limited in this respect.

[0143] In some example embodiments, terminal device 110 may send an indication to first network device 120 whether terminal device 110 is capable of using a first or more machine learning models (e.g., for a two-sided model use case). For example, terminal device 110 may indicate the advanced capability to first network device 120 via, for example, a UE capability message before sending first information to first network device 120.

[0144] In some example embodiments, the transmission of first information can be used to trigger the first network device 120 to obtain the performance of one or more machine learning models (also referred to as second or more machine learning models) available at the second network device 130. In this case, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models. The performance of the second or more machine learning models can be obtained based on the first information. The second network device 130 may include adjacent network devices for the terminal device 110. Alternatively or additionally, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models and one or more machine learning models available at the first network device 120. It should be understood that although embodiments of this disclosure have been discussed in the case where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130, embodiments of this disclosure can also be applied to cases where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130 and one or more machine learning models available at the first network device 120. The scope of this disclosure is not limited in this respect. Further details will be discussed in more detail below.

[0145] In some example embodiments, performance may include the combined (i.e., overall) performance / compatibility (e.g., interoperability characteristics) of a second or more machine learning models of the second network device 130 and a first or more machine learning models at the terminal device 110. As an example implementation, performance may be relative to the current model performance of the first network device 120, for example, relative to the performance of the current cell(s) model(s).

[0146] In an example embodiment where the first or more machine learning models are used as an encoder for encoding CSI, a second or more machine learning models can be used as a decoder for decoding the CSI encoded by the first or more machine learning models. In this case, performance can include the joint performance (inference) of both the terminal-side model / function and the network-side model / function (i.e., the encoder-decoder pair). For example, performance can be associated with performance indicators for decoding the encoded CSI, such as decoding accuracy, decoded CQI, etc.

[0147] The transmission of the first information used to trigger the initiation of model / functional validation tests associated with a second or more machine learning models can be triggered in various ways. For example, the transmission of the first information can be triggered from the terminal side. As another example, the transmission of the first information can be triggered from the network side.

[0148] In some example embodiments where the transmission of the first information is triggered from the terminal side, the terminal device 110 may send the first information based on an update of the machine learning model at the terminal device 110. For example, when a model update occurs at the terminal device 110, the terminal device 110 may trigger the first network device 120 to obtain the performance of a second or more machine learning models available at the second network device 130; in other words, trigger a model / function verification test. In this case, when a model / function update exists at the terminal device 110, the terminal device 110 may need to use the updated model / function. The terminal device 110 may then send an instruction to the first network device 120 (e.g., a message carrying the first information, for example, along with additional reasoning information, which will be discussed in detail below) to trigger a model / function verification test against the updated model(s) available at the terminal device 110 and the network-side models / functions of the second network device 130.

[0149] In some example embodiments where the transmission of the first information is triggered from the terminal side, terminal device 110 may send the first information based on a change in channel conditions associated with terminal device 110. The change in radio conditions may be associated with the mobility of terminal device 110. In some implementations, terminal device 110 may perform cell / beam level radio measurements (e.g., L1 / L3 RSRP) in the primary / serving cell. When terminal device 110 determines that the measured RSRP level has decreased (e.g., a decrease in the RSRP level associated with the primary cell) (e.g., a decrease compared to a previous RSRP measurement, a decrease below a threshold), terminal device 110 may send an indication (e.g., a message carrying the first information, e.g., along with additional cause information, which will be discussed in detail below) to first network device 120 to trigger model / functional verification testing. In some implementations, terminal device 110 may perform radio measurements (e.g., RSRP measurements) associated with second network device 130. When terminal device 110 determines an increase in the measured RSRP level corresponding to one or more neighboring cells of second network device 130, terminal device 110 may send an indication (e.g., a message carrying first information, such as additional cause information, which will be discussed in detail below) to first network device 120 to trigger a model / functional verification test. Alternatively or additionally, this trigger may also be generated by the determination that terminal device 110 is moving toward the cell edge.

[0150] In some example embodiments where the transmission of the first information is triggered from the terminal side, the terminal device 110 may send the first information based on moving to a new cell. For example, whenever the terminal device 110 moves to a new cell, the new cell may become the primary cell, and the old primary cell may become one of one or more neighboring cells. In this scenario, the old set of neighboring cells may change because some cells in the old set may no longer be valid neighbors, and some new neighboring cells are added to the set of neighboring cells. In this case, the terminal device 110 may send an indication (e.g., a message carrying the first information, for example, along with additional reason information, which will be discussed in detail below) to the first network device 120 to trigger model / functional verification testing.

[0151] In some example embodiments where the transmission of the first information is triggered from the network side, the first network device 120 may send an instruction to the terminal device 110 to send the first information, and then the terminal device 110 may send the first information accordingly, as will be discussed below.

[0152] In some implementations, when a machine learning model / feature update / change occurs at the first network device 120 or the second network device 130, the first network device 120 may instruct the terminal device 110 to send a first message to trigger the initiation of a model / feature verification test. In this case, when a model / feature update exists at the first network device 120 or the second network device 130, the corresponding network device may need to use the updated network-side model / feature. Then, the network device (e.g., the first network device 120 or the second network device 130) may trigger a model / feature verification test for the network-side updated(multiple) models / features relative to the terminal-side(multiple) models / features.

[0153] In some implementations, when the first network device 120 determines that the terminal device 110 is moving toward the cell edge, it can instruct the terminal device 110 to send a first message to trigger the initiation of a model / functional verification test.

[0154] In some implementations, the first network device 120 may instruct the terminal device 110 to send a first message to trigger the initiation of a model / functional verification test, while preparing for a possible handover from the terminal device 110 to a second network device 130 (e.g., to one of one or more network devices 130-1, 130-2, ..., 130-N). In this case, when the first network device 120 determines that a handover is about to occur, it may instruct the terminal device 110 to send a first message to trigger the initiation of a model / functional verification test.

[0155] In some example embodiments, terminal device 110 may send reason information to first network device 120, such as an indication of the reason for sending the first indication. The reason may include at least one of the following: an update of the machine learning model at terminal device 110; a change in channel conditions associated with terminal device 110; or terminal device 110 moving to a new cell.

[0156] In some example embodiments, the first information and the aforementioned cause information (e.g., an indication of cause) may be included in a signaling message, such as a higher-level signaling message, like Radio Access Control (RRC) signaling. As an example, the message may include NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST(input_payload, cause). This message may be intended by terminal device 110 to trigger the first network device 120 to obtain the performance of a second or more machine learning models, or in other words, to trigger a model / function performance test on the pool of models / functions(s) available at the second network device 130. The "input_payload" field may include the input payload to the second or more machine learning models / functions available at the second network device 130. For example, the input payload may include the output of a terminal-side model (e.g., compressed CSI provided as output by a terminal-side encoder) or a processed version of the terminal-side model's output. The “Reason” field can be designed to indicate the reason for initiating NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST, such as machine learning model update at terminal device 110 (e.g., UE-side model update), change of channel conditions associated with terminal device 110, or terminal device 110 moving to a new cell (e.g., notification area (RNA) based on a new RAN).

[0157] Upon receiving the first information from terminal device 110, the first network device 120 can be triggered to obtain the performance (i.e., inference / test results) of a second or more machine learning models. Then, as... Figure 4 As shown, the first network device 120 sends (410) information (also referred to as second information) about the performance of a second or more machine learning models to the terminal device 110. For example, the performance of the second or more machine learning models may include the ranked joint performance of the terminal-network side model pairs. Alternatively or additionally, the information about the performance of the second or more machine learning models may be sent to the second network device 130 (i.e., shared with the second network device 130).

[0158] In some example embodiments, the second information may be included in a signaling message, such as a higher-level signaling message, like RRC signaling. As an example, the message may include NEIGHBOR_GNB_MODEL_PERFORMANCE_REPORT(gnb_model_perf_info). This message may be intended for use by the first network device 120 to send a merged / related report of the performance of a second or more machine learning models / features in the pool of the second network device 130 (and / or the first network device 120) to the terminal device 110. The “gnb_model_perf_info” field may include model / feature performance information for the second or more machine learning models (and / or one or more models of the first network device 120) of the second network device 130. The merged report can then be used to optimize LCM signaling related to AI / ML models / features during mobility scenarios.

[0159] Whether sharing of AI / ML model / function related information (e.g., model / function ID used to identify the model / function) / files (e.g., one or more model / function parameters or complete model / function files) between the first network device 120 and the second network device 130 is permitted / supported, the method by which the first network device 120 obtains the performance of the second or more machine learning models can be different.

[0160] In some example embodiments, AI / ML model / function-related information / file sharing may be permitted between the first network device 120 and the second network device 130, and in this case, the first network device 120 may be aware of a second or more machine learning models running at the second network device 130. For example, AI / ML model / function-related information / file sharing may be supported if the first network device 120 and the second network device 130 are from the same vendor. As another example, if the first network device 120 and the second network device 130 are from different vendors, AI / ML model / function-related information / file sharing may be supported if the model / function-related information / file sharing between the different vendors is agreed upon, for example, in 3GPP. As yet another example, AI / ML model / function-related information / file sharing may be supported if the AI / ML model / function is standardized in 3GPP to assist the terminal device 110 during mobility scenarios.

[0161] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the performance of the second or more machine learning models can be determined by the first network device 120 itself. For example, the second or more machine learning models / features in the pool of the second network device 130 can be the same as one or more in the first network device 120, and therefore the required one or more models / features of the second network device 130 may already be available at the first network device 120. As another example, the model / feature pools at the second network device 130 and the first network device 120 are different, and in this case, the model / feature(s) files of the second network device 130 can be shared with the first network device 120. Because AI / ML model sharing is allowed, the first network device 120 can perform corresponding interoperability checks (i.e., performance verification) at its end. For example, the first network device 120 can evaluate / verify the joint performance of the first or more machine learning models / features (e.g., encoders) with each of the second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130.

[0162] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the first network device 120 may query and / or collect information and / or files of a second or more machine learning models / features present in a pool at the second network device 130. The first network device 120 may run inference on the second or more machine learning models / features to evaluate the interoperability performance of (multiple) endpoint-side models / features with (multiple) network-side models / features available at the second network device 130. Alternatively or additionally, the first network device 120 may evaluate the interoperability performance of (multiple) endpoint-side models / features with (multiple) network-side models / features available at the first network device 120.

[0163] The above reference Figures 3A to 3C The described operations and features also apply to signaling flow 400 and have similar effects. For the sake of simplicity, details will be omitted.

[0164] In some other example embodiments, sharing of AI / ML model / feature-related information / files may not be permitted between the first network device 120 and the second network device 130, and in such cases, the first network device 120 may be unaware of the model / features of the second network device 130. For example, if the first network device 120 and the second network device 130 are from two or more vendors, sharing of AI / ML model / feature-related information / files may not be supported, thereby limiting inter-vendor model sharing for reasons such as proprietary and security.

[0165] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may query / request the performance (i.e., performance verification / inference) of a second or more machine learning models from the second network device 130. For example, the first network device 120 may query / request the joint performance of a first or more machine learning models / features (e.g., encoders) with each of a second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130. In this case, the first network device 120 may query / request a corresponding interoperability check (i.e., joint performance verification) between the first or more machine learning models available on the terminal side and the second or more machine learning models available on the network side from the second network device 130.

[0166] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may forward compressed input received from the terminal device 110 to the second network device 130 to allow the second network device 130 to determine the interoperability performance of(multiple) terminal-side models / features with their corresponding(multiple) network-side models / features. The first network device 120 may collect inference / test results and then share them with the terminal device 110.

[0167] As referenced above Figure 3D and 3E The operations and features described also apply to signaling flow 400 and have similar effects. For the sake of simplicity, details will be omitted.

[0168] Return to reference Figure 4 Based on the performance of one or more second machine learning models available at the second network device 130, the first network device 120 can determine a network device (e.g., network device 130-1) from the second network device 130 (i.e., from one or more network devices 130-1, 130-2, ..., 130-N) to switch terminal device 110. Then, the first network device 120 performs (415) a handover process from terminal device 110 to the second network device 130 (i.e., the determined network device, such as network device 130-1). Therefore, terminal device 110 performs (420) a handover from the first network device 120 to the second network device 130 (i.e., the determined network device, such as network device 130-1).

[0169] After the switch, terminal device 110 may need to determine whether to use the activated machine learning model from the first or one more machine learning models. Depending on the different second pieces of information, the terminal's utilization behavior of the joint performance report can differ, as will be discussed below.

[0170] In some example embodiments, if the second information indicates that the performance of a second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs a handover) is unacceptable relative to any of the first or more machine learning models, then the terminal device 110 may stop using the first or more machine learning models after the handover. In this case, if no alternative / relatively better model of the first or more machine learning models is available, the terminal device 110 may fall back to non-AI / ML behavior to avoid potential performance degradation after the cell change. For example, if the terminal device 110 is a simple / basic UE equipped with only one / few AI / ML models / functions, it may first indicate its model / function limitations to the first network device 120 through UE capabilities. Based on this capability, the first network device 120 may decide to provide the terminal device 110 with limited performance reports, such as a binary report mentioning whether the joint performance is acceptable or unacceptable, to help the terminal device 110 fall back to non-AI / ML behavior in advance (if needed) in a network-transparent or opaque manner.

[0171] In some example embodiments, if the second information indicates that at least one of the second or more machine learning models (e.g., one or more models available at network device 130-1 where the handover to terminal device 110 is performed) has acceptable performance relative to the activated machine learning model among the first or more machine learning models, then terminal device 110 may continue to use the activated machine learning model after the handover and / or store the second information. In this case, if the joint performance is above an acceptable level, the same activated model / function may continue to be used after the cell change, and / or the performance results provided by the first network device 120 may be stored for future mobility optimization purposes. As an example implementation, when the aforementioned acceptable level of joint performance is identified, terminal device 110 may store the ID / information of the terminal-network-side model pair (e.g., encoder-decoder model pair) and the ID(s) of the second network device 130, such that if terminal device 110 will need to access the second network device 130 in the future (e.g., the same set / subset of one or more network devices 130-1, 130-2, ..., 130-N), it will not be necessary to re-evaluate the joint performance, and the stored information can be reused, saving evaluation signaling and resources. For example, the model / function ID may be stored, for example, in terminal device 110 or in a subscriber identity module (SIM) card for future use.

[0172] In some example embodiments, if the second information indicates that the performance of a second or more machine learning models (e.g., one or more models available at network device 130-1 to which terminal device 110 performs the handover) is unacceptable relative to the activated machine learning model, and at least one of the second or more machine learning models (e.g., one or more models available at network device 130-1 to which terminal device 110 performs the handover) is acceptable relative to at least one of the first or more machine learning models, then terminal device 110 may stop using the activated machine learning model after the handover and start using at least one of the first or more machine learning models after the handover. In this case, based on the joint performance provided by the first network device 120, terminal device 110 may switch its activated AI / ML model / feature to another alternative / relatively better model / feature of the first or more machine learning models (if available) to avoid potential performance degradation after cell change. For example, if terminal device 110 is an advanced UE equipped with multiple AI / ML models / functions, it can first indicate its model / function switching capability to the first network device 120 through the UE capability. The first network device 120 can then decide to provide terminal device 110 with a detailed performance report so that it can switch to a better performing model / function in a network-transparent or non-transparent manner (if needed).

[0173] Figure 5 A third signaling flow 500 between a terminal device and a first network device according to some example embodiments of the present disclosure is illustrated. For discussion purposes, reference will be made to... Figure 1A Describe signaling flow 500.

[0174] like Figure 5 As shown, based on a change in channel conditions associated with terminal device 110, terminal device 110 sends (505) information (also called first information) to first network device 120 regarding the outputs of one or more machine learning models (also called the first one or more machine learning models) available at terminal device 110. The change in radio conditions may be associated with the mobility of terminal device 110. First network device 120 may include a main network device serving terminal network device 110 or to which terminal device 110 is locked.

[0175] In some example embodiments, the first one or more machine learning models can be used as encoders for encoding CSI, and in this case, the output can correspond to the CSI encoded by the first one or more machine learning models. In other words, the output can include one or more compressed CSIs corresponding to the first one or more machine learning models. For example, if the terminal device 110 is capable of storing / using multiple AI / ML models / functions for two-sided use cases, the output can include multiple compressed CSIs corresponding to multiple terminal-side models / functions. The first one or more machine learning models can also be used for any other purpose (e.g., as an encoder for encoding any other information, as a terminal-side model portion of another two-sided model use case, etc.), and the scope of this disclosure is not limited in this respect.

[0176] In some example embodiments, terminal device 110 may send an indication to first network device 120 whether terminal device 110 is capable of using a first or more machine learning models (e.g., for a two-sided model use case). For example, terminal device 110 may indicate the advanced capability to first network device 120 via, for example, a UE capability message before sending first information to first network device 120.

[0177] In some example embodiments, the transmission of first information can be used to trigger the first network device 120 to obtain the performance of one or more machine learning models (also referred to as second or more machine learning models) available at the second network device 130. In this case, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models. The performance of the second or more machine learning models can be obtained based on the first information. The second network device 130 may include adjacent network devices for the terminal device 110. Alternatively or additionally, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models and one or more machine learning models available at the first network device 120. It should be understood that although embodiments of this disclosure have been discussed in the case where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130, embodiments of this disclosure can also be applied to cases where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130 and one or more machine learning models available at the first network device 120. The scope of this disclosure is not limited in this respect. Further details will be discussed in more detail below.

[0178] In some example embodiments, performance may include the joint (i.e., overall) performance / compatibility (e.g., interoperability characteristics) of a second or more machine learning models of the second network device 130 and a first or more machine learning models at the terminal device 110. As an example implementation, performance may be relative to the current model performance of the first network device 120, for example, relative to the performance of the model of the current cell.

[0179] In an example embodiment where the first or more machine learning models are used as an encoder for encoding CSI, a second or more machine learning models can be used as a decoder for decoding the CSI encoded by the first or more machine learning models. In this case, performance can include the joint performance (inference) of both the terminal-side model / function and the network-side model / function (i.e., the encoder-decoder pair). For example, performance can be associated with performance indicators for decoding the encoded CSI, such as decoding accuracy, decoded CQI, etc.

[0180] In some example embodiments, changes in channel conditions associated with a terminal device can be determined in various ways. In some implementations, terminal device 110 can perform cell / beam level radio measurements (e.g., L1 / L3 RSRP) in the primary / serving cell. When terminal device 110 determines a decrease in the measured RSRP level (e.g., a decrease compared to a previous RSRP measurement, a decrease below a threshold), it can send an indication (e.g., a message carrying first information, such as additional cause information, which will be discussed in detail below) to first network device 120 to trigger a model / functional verification test. In some implementations, terminal device 110 can perform radio measurements (e.g., RSRP measurements) associated with a second network device 130. When terminal device 110 determines an increase in the RSRP level for measurements corresponding to the neighboring cells(s) of the second network device 130, terminal device 110 can send an indication (e.g., a message carrying first information, such as additional cause information, which will be discussed in detail below) to first network device 120 to trigger a model / functional verification test. Alternatively, or additionally, the trigger can also be generated by determining that the terminal device 110 is moving toward the cell edge.

[0181] In some example embodiments, terminal device 110 may send reason information to first network device 120, such as an indication of the reason for sending the first indication, i.e., a change in channel conditions associated with terminal device 110.

[0182] In some example embodiments, the first information and the aforementioned cause information (e.g., indication of cause) may be included in a signaling message, such as a higher-level signaling message, such as Radio Access Control (RRC) signaling. As an example, the message may include NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST(input_payload, cause). This message may be intended by terminal device 110 to trigger the first network device 120 to obtain the performance of a second or more machine learning models, or in other words, to trigger a model / function performance test on the pool of models / functions(s) available at the second network device 130. The "input_payload" field may include the input payload to the second or more machine learning models / functions available at the second network device 130. For example, the input payload may include the output of a terminal-side model (e.g., compressed CSI provided as output by a terminal-side encoder) or a processed version of the terminal-side model's output. The "cause" field may be intended to indicate the reason for initiating NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST, i.e., a change in channel conditions associated with terminal device 110.

[0183] Upon receiving the first information from terminal device 110, the first network device 120 can be triggered to obtain the performance (i.e., inference / test results) of a second or more machine learning models. Then, as... Figure 5 As shown, the first network device 120 sends (510) information (also referred to as second information) about the performance of a second or more machine learning models to the terminal device 110. For example, the performance of the second or more machine learning models may include the joint performance of the terminal-network side model pairs based on their rankings. Alternatively or additionally, the information about the performance of the second or more machine learning models may be sent to the second network device 130 (i.e., shared with the second network device 130).

[0184] In some example embodiments, the second information may be included in a signaling message, such as a higher-level signaling message, like RRC signaling. As an example, the message may include F1232748P-US (gnb_model_perf_info). This message may be intended to be used by the first network device 120 to send a merged / related report of the performance of a second or more machine learning models / features in the pool of the second network device 130 (and / or the first network device 120) to the terminal device 110. The “gnb_model_perf_info” field may include model / feature performance information for the second or more machine learning models of the second network device 130 (and / or one or more models of the first network device 120). The merged report can then be used to optimize LCM signaling related to AI / ML models / features during mobility scenarios.

[0185] Whether sharing of AI / ML model / function related information (e.g., model / function ID used to identify the model / function) / files (e.g., one or more model / function parameters or complete model / function files) between the first network device 120 and the second network device 130 is permitted / supported, the method by which the first network device 120 obtains the performance of the second or more machine learning models can be different.

[0186] In some example embodiments, AI / ML model / function-related information / file sharing may be permitted between the first network device 120 and the second network device 130, and in this case, the first network device 120 may be aware of a second or more machine learning models running at the second network device 130. For example, AI / ML model / function-related information / file sharing may be supported if the first network device 120 and the second network device 130 are from the same vendor. As another example, if the first network device 120 and the second network device 130 are from different vendors, AI / ML model / function-related information / file sharing may be supported if, for example, model / function-related information / file sharing between different vendors is agreed upon in 3GPP. As yet another example, AI / ML model / function-related information / file sharing may be supported if the AI / ML model / function is standardized in 3GPP to assist the terminal device 110 during mobility scenarios.

[0187] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the performance of the second or more machine learning models can be determined by the first network device 120 itself. For example, the second or more machine learning models / features in the pool of the second network device 130 can be the same as one or more in the first network device 120, and therefore the required one or more models / features of the second network device 130 may already be available at the first network device 120. As another example, the model / feature pools at the second network device 130 and the first network device 120 are different, and in this case, the model / feature(s) files of the second network device 130 can be shared with the first network device 120. When AI / ML model sharing is allowed, the first network device 120 can perform corresponding interoperability checks (i.e., performance verification) at its end. For example, the first network device 120 can evaluate / verify the joint performance of the first or more machine learning models / features (e.g., encoders) with each of the second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130.

[0188] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the first network device 120 may query and / or collect information and / or files of a second or more machine learning models / features present in a pool at the second network device 130. The first network device 120 may run inference on the second or more machine learning models / features to evaluate the interoperability performance of(multiple) endpoint-side models / features with(multiple) network-side models / features available at the second network device 130. Alternatively or additionally, the first network device 120 may evaluate the interoperability performance of(multiple) endpoint-side models / features with(multiple) network-side models / features available at the first network device 120.

[0189] The above reference Figures 3A to 3C The described operations and features also apply to Signaling Flow 500 and have similar effects. For the sake of simplicity, details will be omitted.

[0190] In some other example embodiments, sharing of AI / ML model / feature-related information / files between the first network device 120 and the second network device 130 may not be permitted, and in such cases, the first network device 120 may be unaware of the model / features of the second network device 130. For example, if the first network device 120 and the second network device 130 are from two or more vendors, sharing of AI / ML model / feature-related information / files may not be supported, thereby restricting inter-vendor model sharing for reasons such as proprietary and security.

[0191] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may query / request the performance (i.e., performance verification / inference) of a second or more machine learning models from the second network device 130. For example, the first network device 120 may query / request the joint performance of a first or more machine learning models / features (e.g., encoders) with each of a second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130. In this case, the first network device 120 may query / request a corresponding interoperability check (i.e., joint performance verification) between the first or more machine learning models available on the terminal side and the second or more machine learning models available on the network side from the second network device 130.

[0192] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may forward compressed input received from the terminal device 110 to the second network device 130 to allow the second network device 130 to determine the interoperability performance of(multiple) terminal-side models / features with their corresponding(multiple) network-side models / features. The first network device 120 may collect inference / test results and then share them with the terminal device 110.

[0193] As referenced above Figure 3D and 3E The operations and features described also apply to signaling flow 500 and have similar effects. For the sake of simplicity, details will be omitted.

[0194] In some example embodiments, after receiving second information from the first network device 120, the terminal device 110 may perform cell selection or cell reselection based at least on the second information. In this case, based on the second information, the terminal device 110 may know one or more second machine learning models available at the second network device 130 and their performance for a given channel condition, and then, based on this, the terminal device may perform cell selection or cell reselection accordingly.

[0195] In some example embodiments, if terminal device 110 determines that the performance of at least one machine learning model available at at least one network device, as indicated by the second information, is acceptable, then terminal device 110 may select or reselect a cell of at least one of the network devices 130-1, 130-2, ..., 130-N for connection based on performance. Whether the performance of at least one machine learning model available at at least one network device is acceptable can be determined based on an acceptable level or threshold. For example, if the joint performance is above an acceptable level, then the joint performance is acceptable.

[0196] In some example embodiments, a corresponding network device among one or more network devices 130-1, 130-2, ..., 130-N may perform a system information broadcast indicating one or more machine learning models available at the corresponding network device. In this case, a broadcast of a list of models / functions (model / function IDs) available for the cell may be introduced in the system information broadcast (e.g., System Information Block (SIB4), System Information Block (SIB5), etc.). Therefore, terminal device 110 may receive the system information broadcast and then select or reselect a cell based at least on the received system information broadcast. For example, if the machine learning model available at the corresponding network device among one or more network devices 130-1, 130-2, ..., 130-N is acceptable and has better joint performance relative to the terminal-side model(s) than one or more other network devices, and if the signal strength of the corresponding network device and one or more other network devices is the same or similar, then terminal device 110 may select or reselect a cell of the corresponding network device.

[0197] After cell selection or cell reselection, terminal device 110 may need to determine whether to use the activated machine learning model from the first or one more machine learning models. Depending on the different second pieces of information, the terminal's behavior using the joint performance report can differ, as will be discussed below.

[0198] In some example embodiments, if the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models, the terminal device 110 may stop using the first or more machine learning models after cell selection or cell reselection.

[0199] In some example embodiments, if the second information indicates that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model, the terminal device 110 may continue to use the activated machine learning model after cell selection or cell reselection, and / or store the second information.

[0200] In some example embodiments, if the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the performance of the activated machine learning model, and the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to at least one of the first or more machine learning models, then the terminal device 110 may stop using the activated machine learning model after cell selection or cell reselection, and start using at least one of the first or more machine learning models after cell selection or cell reselection.

[0201] As per the above reference Figure 3F The operations and features described also apply to signaling flow 500 and have similar effects. For the sake of simplicity, details will be omitted.

[0202] In some example embodiments, based on the performance of a second or more machine learning models available at the second network device 130, the first network device 120 may determine a network device (e.g., network device 130-1) from the second network device 130 (i.e., from one or more network devices 130-1, 130-2, ..., 130-N) to switch terminal device 110. The first network device 120 may then perform a handover process from terminal device 110 to the second network device 130 (i.e., the determined network device, such as network device 130-1). Therefore, terminal device 110 may perform a handover from the first network device 120 to the second network device 130 (i.e., the determined network device, such as network device 130-1).

[0203] After the switch, terminal device 110 may need to determine whether to use the activated machine learning model from the first or one more machine learning models. Depending on the different second pieces of information, the terminal's behavior using the joint performance report can differ, as will be discussed below.

[0204] In some example embodiments, if the second information indicates that the performance of a second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs a handover to it) is unacceptable relative to any of the first or more machine learning models, the terminal device 110 may stop using the first or more machine learning models after the handover.

[0205] In some example embodiments, if the second information indicates that the performance of at least one machine learning model in the second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs a handover to it) is acceptable relative to the performance of the activated machine learning model in the first or more machine learning models, the terminal device 110 may continue to use the activated machine learning model after the handover and / or store the second information.

[0206] In some example embodiments, if the second information indicates that the performance of a second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs the handover) is unacceptable relative to the activated machine learning model, and the performance of at least one of the second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs the handover) is acceptable relative to at least one of the first or more machine learning models, then the terminal device 110 may stop using the activated machine learning model after the handover and start using at least one of the first or more machine learning models after the handover.

[0207] In an example embodiment where terminal device 110 ceases using one or more machine learning models after a cell change (e.g., after cell selection, cell reselection, or handover), if no alternative / relatively better model is available for the first one or more machine learning models, terminal device 110 may fall back to non-AI / ML behavior to avoid potential performance degradation after the cell change. For example, in the case where terminal device 110 is a simple / basic UE equipped with only one / few AI / ML models / functions, it may first indicate its model / function limitations to first network device 120 via UE capabilities. Based on this capability, first network device 120 may decide to provide terminal device 110 with limited performance reports, such as a binary report mentioning whether the joint performance is acceptable or unacceptable, to help terminal device 110 fall back to non-AI / ML behavior in advance (if needed) in a network-transparent or opaque manner.

[0208] In an example embodiment where terminal device 110 continues to use an activated machine learning model after a cell change (e.g., after cell selection, cell reselection, or handover), if the joint performance is above an acceptable level, the same activated model / function can continue to be used after the cell change, and / or the performance results provided by the first network device 120 can be stored for future mobility optimization purposes. As an example implementation, when identifying the aforementioned acceptable level of joint performance, terminal device 110 can store the ID / information of the terminal-network-side model pair (e.g., encoder-decoder model pair) and the ID(s) of the second network device 130, such that if terminal device 110 will need to access the second network device 130 in the future (e.g., the same set / subset of one or more network devices 130-1, 130-2, ..., 130-N), it will not be necessary to re-evaluate the joint performance, and the stored information can be reused to save evaluation signaling and resources. For example, the model / function ID can be stored, for example, in terminal device 110 or a subscriber identity module (SIM) card for future use.

[0209] In an example embodiment where terminal device 110 stops using the activated machine learning model and starts using at least one other machine learning model among the first or more machine learning models after a cell change (e.g., after cell selection, cell reselection, or handover), based on the joint performance provided by the first network device 120, terminal device 110 can switch its activated AI / ML model / function to another alternative / relatively better model / function (if available) of the first or more machine learning models to avoid potential performance degradation after the cell change. For example, if terminal device 110 is an advanced UE equipped with multiple AI / ML models / functions, it can first indicate its model / function switching capability to the first network device 120 through UE capabilities. The first network device 120 can then decide to provide terminal device 110 with a detailed performance report so that it can switch to a better performing model / function (if needed) in a network-transparent or non-transparent manner.

[0210] Figure 6 A fourth signaling flow 600 between a terminal device and a first network device according to some example embodiments of the present disclosure is illustrated. For discussion purposes, reference will be made to... Figure 1A Describe signaling flow 600.

[0211] like Figure 6As shown, based on the machine learning model update at terminal device 110, terminal device 110 sends (605) information (also called first information) to first network device 120 regarding the output of one or more machine learning models (also called the first one or more machine learning models) available at terminal device 110. Changes in radio conditions may be associated with the mobility of terminal device 110. First network device 120 may include a main network device serving terminal network device 110 or to which terminal device 110 is locked.

[0212] In some example embodiments, the first one or more machine learning models can be used as encoders for encoding CSI, and in this case, the output can correspond to the CSI encoded by the first one or more machine learning models. In other words, the output can include one or more compressed CSIs corresponding to the first one or more machine learning models. For example, if the terminal device 110 is capable of storing / using multiple AI / ML models / functions for two-sided use cases, the output can include multiple compressed CSIs corresponding to multiple terminal-side models / functions. The first one or more machine learning models can also be used for any other purpose (e.g., as an encoder for encoding any other information, as a terminal-side model portion of another two-sided model use case, etc.), and the scope of this disclosure is not limited in this respect.

[0213] In some example embodiments, terminal device 110 may send an indication to first network device 120 regarding whether terminal device 110 is capable of using a first or more machine learning models (e.g., for a two-sided model use case). For example, terminal device 110 may indicate this advanced capability to first network device 120 via, for example, a UE capability message before sending first information to first network device 120.

[0214] In some example embodiments, the transmission of first information can be used to trigger the first network device 120 to obtain the performance of one or more machine learning models (also referred to as second or more machine learning models) available at the second network device 130. In this case, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models. The performance of the second or more machine learning models can be obtained based on the first information. The second network device 130 may include adjacent network devices for the terminal device 110. Alternatively or additionally, the transmission of first information can be used to trigger the initiation of model / functionality verification tests associated with the second or more machine learning models and one or more machine learning models available at the first network device 120. It should be understood that although embodiments of this disclosure have been discussed in the case where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130, embodiments of this disclosure can also be applied to cases where the transmission of the first information is used to trigger the acquisition of the performance of the second or more machine learning models available at the second network device 130 and one or more machine learning models available at the first network device 120. The scope of this disclosure is not limited in this respect. Further details will be discussed in more detail below.

[0215] In some example embodiments, performance may include the joint (i.e., overall) performance / compatibility (e.g., interoperability characteristics) of a second or more machine learning models of the second network device 130 and a first or more machine learning models at the terminal device 110. As an example implementation, performance may be relative to the current model performance of the first network device 120, for example, relative to the performance of the model of the current cell.

[0216] In an example embodiment where the first or more machine learning models are used as an encoder for encoding CSI, a second or more machine learning models can be used as a decoder for decoding the CSI encoded by the first or more machine learning models. In this case, performance can include the joint performance (inference) of both the terminal-side model / function and the network-side model / function (i.e., the encoder-decoder pair). For example, performance can be associated with performance indicators for decoding the encoded CSI, such as decoding accuracy, decoded CQI, etc.

[0217] In some example embodiments, when a model update occurs at terminal device 110, terminal device 110 may trigger first network device 120 to obtain the performance of a second or more machine learning models available at second network device 130; in other words, trigger model / feature verification testing. In this case, when a model / feature update exists at terminal device 110, terminal device 110 may need to use the updated model / feature. Terminal device 110 may then send an instruction to first network device 120 (e.g., a message carrying first information, such as additional reasoning information, which will be discussed in detail below) to trigger model / feature verification testing against the updated model(s) available at terminal device 110 and the network-side models / features of second network device 130.

[0218] In some example embodiments, terminal device 110 may send reason information to first network device 120, such as an indication of the reason for sending a first indication, i.e., an update of the machine learning model at terminal device 110.

[0219] In some example embodiments, the first information and the aforementioned cause information (e.g., an indication of cause) may be included in a signaling message, such as a higher-level signaling message, like Radio Access Control (RRC) signaling. As an example, the message may include NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST(input_payload, cause). This message may be intended by terminal device 110 to trigger the first network device 120 to obtain the performance of a second or more machine learning models, or in other words, to trigger a model / feature performance test on the pool of models / features(s) available at the second network device 130. The "input_payload" field may include the input payload to the second or more machine learning models / features available at the second network device 130. For example, the input payload may include the output of a terminal-side model (e.g., compressed CSI provided as output by a terminal-side encoder) or a processed version of the terminal-side model's output. The "cause" field may be intended to indicate the reason for initiating NEIGHBOR_GNB_MODEL_PERFORMANCE_REQUEST, i.e., an update to the machine learning model at terminal device 110.

[0220] Upon receiving the first information from terminal device 110, the first network device 120 can be triggered to obtain the performance (i.e., inference / test results) of a second or more machine learning models. Then, as... Figure 6As shown, the first network device 120 sends (610) information (also referred to as second information) about the performance of a second or more machine learning models to the terminal device 110. For example, the performance of the second or more machine learning models may include the ranked joint performance of the terminal-network side model pairs. Alternatively or additionally, the information about the performance of the second or more machine learning models may be sent to the second network device 130 (i.e., shared with the second network device 130).

[0221] In some example embodiments, the second information may be included in a signaling message, such as a higher-level signaling message, like RRC signaling. As an example, the message may include NEIGHBOR_GNB_MODEL_PERFORMANCE_REPORT(gnb_model_perf_info). This message may be intended for use by the first network device 120 to send a merged / related report of the performance of a second or more machine learning models / features in the pool of the second network device 130 (and / or the first network device 120) to the terminal device 110. The “gnb_model_perf_info” field may include model / feature performance information for the second or more machine learning models of the second network device 130 (and / or one or more models of the first network device 120). The merged report can then be used to optimize LCM signaling related to AI / ML models / features during mobility scenarios.

[0222] Whether sharing of AI / ML model / function related information (e.g., model / function ID used to identify the model / function) / files (e.g., one or more model / function parameters or complete model / function files) between the first network device 120 and the second network device 130 is permitted / supported, the method by which the first network device 120 obtains the performance of the second or more machine learning models can be different.

[0223] In some example embodiments, AI / ML model / function-related information / file sharing can be permitted between the first network device 120 and the second network device 130, and in this case, the first network device 120 can be aware of a second or more machine learning models running at the second network device 130. For example, AI / ML model / function-related information / file sharing can be supported if the first network device 120 and the second network device 130 are from the same vendor. As another example, if the first network device 120 and the second network device 130 are from different vendors, AI / ML model / function-related information / file sharing can be supported if, for example, model / function-related information / file sharing between different vendors is agreed upon in 3GPP. As yet another example, AI / ML model / function-related information / file sharing can be supported if the AI / ML model / function is standardized in 3GPP to assist the terminal device 110 during mobility scenarios.

[0224] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the performance of the second or more machine learning models can be determined by the first network device 120 itself. For example, the second or more machine learning models / features in the pool of the second network device 130 can be the same as one or more in the first network device 120, and therefore the required one or more models / features of the second network device 130 may already be available at the first network device 120. As another example, the model / feature pools at the second network device 130 and the first network device 120 are different, and in this case, the model / feature files of the second network device 130 can be shared with the first network device 120. When AI / ML model sharing is allowed, the first network device 120 can perform corresponding interoperability checks (i.e., performance verification) at its end. For example, the first network device 120 can evaluate / verify the joint performance of the first or more machine learning models / features (e.g., encoders) with each of the second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130.

[0225] In an example embodiment that allows sharing of AI / ML model / feature related information / files, the first network device 120 may query and / or collect information and / or files of a second or more machine learning models / features present in the pool of the second network device 130. The first network device 120 may run inference on the second or more machine learning models / features to evaluate the interoperability performance of (multiple) endpoint-side models / features with (multiple) network-side models / features available at the second network device 130. Alternatively or additionally, the first network device 120 may evaluate the interoperability performance of endpoint-side models / features with network-side models / features available at the first network device 120.

[0226] The above reference Figures 3A to 3C The described operations and features also apply to Signaling Flow 600 and have similar effects. For the sake of simplicity, details will be omitted.

[0227] In some other example embodiments, sharing of AI / ML model / feature-related information / files may not be permitted between the first network device 120 and the second network device 130, and in such cases, the first network device 120 may be unaware of the model / features of the second network device 130. For example, if the first network device 120 and the second network device 130 are from two or more vendors, sharing of AI / ML model / feature-related information / files may not be supported, thereby limiting inter-vendor model sharing for reasons such as proprietary and security.

[0228] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may query / request the performance (i.e., performance verification / inference) of a second or more machine learning models from the second network device 130. For example, the first network device 120 may query / request the joint performance of a first or more machine learning models / features (e.g., encoders) with each of a second or more machine learning models / features (e.g., decoders) in the model / feature pool at the second network device 130. In this case, the first network device 120 may query / request a corresponding interoperability check (i.e., joint performance verification) between the first or more machine learning models available on the terminal side and the second or more machine learning models available on the network side from the second network device 130.

[0229] In an example embodiment where sharing of AI / ML model / feature-related information / files is not permitted, the first network device 120 may forward compressed input received from the terminal device 110 to the second network device 130 to allow the second network device 130 to determine the interoperability performance of(multiple) terminal-side models / features with their corresponding(multiple) network-side models / features. The first network device 120 may collect inference / test results and then share them with the terminal device 110.

[0230] As referenced above Figure 3D and 3E The operations and features described also apply to signaling flow 600 and have similar effects. For the sake of simplicity, details will be omitted.

[0231] In some example embodiments, after receiving second information from the first network device 120, the terminal device 110 may perform cell selection or cell reselection based at least on the second information. In this case, based on the second information, the terminal device 110 may know one or more second machine learning models available at the second network device 130 and their performance for a given channel condition, and then, based on this, the terminal device may perform cell selection or cell reselection accordingly.

[0232] In some example embodiments, if terminal device 110 determines that the performance of at least one machine learning model available at at least one network device, as indicated by the second information, is acceptable, then terminal device 110 may select or reselect a cell of at least one of the network devices 130-1, 130-2, ..., 130-N for connection based on performance. Whether the performance of at least one machine learning model available at at least one network device is acceptable can be determined based on an acceptable level or threshold. For example, if the joint performance is above an acceptable level, then the joint performance is acceptable.

[0233] In some example embodiments, a corresponding network device among one or more network devices 130-1, 130-2, ..., 130-N may perform a system information broadcast indicating one or more machine learning models available at the corresponding network device. In this case, a broadcast of a list of models / functions (model / function IDs) available for the cell may be introduced in the system information broadcast (e.g., System Information Block (SIB4), System Information Block (SIB5), etc.). Therefore, terminal device 110 may receive the system information broadcast and then select or reselect a cell based at least on the received system information broadcast. For example, if the machine learning model available at the corresponding network device among one or more network devices 130-1, 130-2, ..., 130-N is acceptable and has better joint performance relative to the terminal-side model(s) than one or more other network devices, and if the signal strength of the corresponding network device and one or more other network devices is the same or similar, then terminal device 110 may select or reselect a cell of the corresponding network device.

[0234] After cell selection or cell reselection, terminal device 110 may need to determine whether to use the activated machine learning model from the first or one more machine learning models. Depending on the different second pieces of information, the terminal's behavior using the joint performance report can differ, as will be discussed below.

[0235] In some example embodiments, if the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models, the terminal device 110 may stop using the first or more machine learning models after cell selection or cell reselection.

[0236] In some example embodiments, if the second information indicates the performance of at least one machine learning model available for the selected or reselected cell relative to the activated machine learning model, the terminal device 110 may continue to use the activated machine learning model after cell selection or cell reselection, and / or store the second information.

[0237] In some example embodiments, if the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the performance of the activated machine learning model, and the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to at least one of the first or more machine learning models, then the terminal device 110 may stop using the activated machine learning model after cell selection or cell reselection, and start using at least one of the first or more machine learning models after cell selection or cell reselection.

[0238] As per the above reference Figure 3F The operations and features described also apply to signaling flow 600 and have similar effects. For the sake of simplicity, details will be omitted.

[0239] In some example embodiments, based on the performance of a second or more machine learning models available at the second network device 130, the first network device 120 may determine a network device (e.g., network device 130-1) from the second network device 130 (i.e., from one or more network devices 130-1, 130-2, ..., 130-N) to switch terminal device 110. The first network device 120 may then perform a handover process from terminal device 110 to the second network device 130 (i.e., the determined network device, such as network device 130-1). Therefore, terminal device 110 may perform a handover from the first network device 120 to the second network device 130 (i.e., the determined network device, such as network device 130-1).

[0240] After the switch, terminal device 110 may need to determine whether to use the activated machine learning model from the first or one more machine learning models. Depending on the different second pieces of information, the terminal's behavior using the joint performance report can differ, as will be discussed below.

[0241] In some example embodiments, if the second information indicates that the performance of a second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs a handover to it) is unacceptable relative to any of the first or more machine learning models, the terminal device 110 may stop using the first or more machine learning models after the handover.

[0242] In some example embodiments, if the second information indicates that at least one of the second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs a handover to it) has acceptable performance relative to the activated machine learning model among the first or more machine learning models, the terminal device 110 may continue to use the activated machine learning model after the handover and / or store the second information.

[0243] In some example embodiments, if the second information indicates that the performance of a second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs the handover) is unacceptable relative to the activated machine learning model, and the performance of at least one of the second or more machine learning models (e.g., one or more models available at network device 130-1 where the terminal device 110 performs the handover) is acceptable relative to at least one of the first or more machine learning models, then the terminal device 110 may stop using the activated machine learning model after the handover and start using at least one of the first or more machine learning models after the handover.

[0244] In an example embodiment where terminal device 110 ceases using one or more machine learning models after a cell change (e.g., after cell selection, cell reselection, or handover), if no alternative / relatively better model is available for the first one or more machine learning models, terminal device 110 may fall back to non-AI / ML behavior to avoid potential performance degradation after the cell change. For example, in the case where terminal device 110 is a simple / basic UE equipped with only one / few AI / ML models / functions, it may first indicate its model / function limitations to first network device 120 via UE capabilities. Based on this capability, first network device 120 may decide to provide terminal device 110 with limited performance reports, such as a binary report mentioning whether the joint performance is acceptable or unacceptable, to help terminal device 110 fall back to non-AI / ML behavior in advance (if needed) in a network-transparent or opaque manner.

[0245] In an example embodiment where terminal device 110 continues to use an activated machine learning model after a cell change (e.g., after cell selection, cell reselection, or handover), if the joint performance is above an acceptable level, the same activated model / function can continue to be used after the cell change, and / or the performance results provided by the first network device 120 can be stored for future mobility optimization purposes. As an example implementation, when identifying the aforementioned acceptable level of joint performance, terminal device 110 can store the ID / information of the terminal-network-side model pair (e.g., encoder-decoder model pair) and the ID(s) of the second network device 130, such that if terminal device 110 will need to access the second network device 130 in the future (e.g., the same set / subset of one or more network devices 130-1, 130-2, ..., 130-N), it will not be necessary to re-evaluate the joint performance, and the stored information can be reused to save evaluation signaling and resources. For example, the model / function ID can be stored, for example, in terminal device 110 or a subscriber identity module (SIM) card for future use.

[0246] In an example embodiment where, after a cell change (e.g., after cell selection, cell reselection, or handover), terminal device 110 stops using the activated machine learning model and begins using at least one other machine learning model among the first or more machine learning models, based on the joint performance provided by the first network device 120, terminal device 110 can switch its activated AI / ML model / function to another alternative / relatively better model / function (if available) of the first or more machine learning models to avoid potential performance degradation after the cell change. For example, if terminal device 110 is an advanced UE equipped with multiple AI / ML models / functions, it can first indicate its model / function switching capability to the first network device 120 through UE capabilities. The first network device 120 can then decide to provide terminal device 110 with a detailed performance report so that it can switch to a better performing model / function (if needed) in a network-transparent or non-transparent manner.

[0247] It should be understood that although embodiments of this disclosure are discussed with respect to CSI compression use cases with AI / ML enabled, embodiments of this disclosure may also be applied to other use cases.

[0248] According to the reference Figures 2 to 6Some embodiments allow the primary network device to make better mobility decisions to hand over a terminal device to a neighboring network device without impacting performance and without needing to re-evaluate the terminal network encoder-decoder pair performance after a cell change. For example, if the primary network device needs to choose between two destination network devices for handover at a terminal device, it can choose the network device with better performance for a given AI / ML model / function at the terminal device. Furthermore, if the terminal device's movement / trajectory is towards a specific network device, the primary network device can determine the optimal network-side model / function (decoder) from a pool of target network devices (e.g., a set of decoders) to ensure the terminal device will be paired with the best decoder after handover. Additionally, by feeding model / function performance information back to the terminal device, the terminal device is allowed to make better mobility decisions in idle-mode cell reselection / handover scenarios. For example, if the terminal device knows the performance of the models / functions available for neighboring cells (network devices), it can choose a cell with a better-performing model during idle-mode cell reselection or connection-mode handover when other cell reselection parameters are nearly identical between different neighboring cells. In this way, enhanced communication performance and improved communication efficiency are possible.

[0249] Figure 7 A flowchart 700 illustrating a method implemented at a terminal device according to some embodiments of the present disclosure is shown. For discussion purposes, references will be made to... Figure 1A Method 700 is described from the perspective of the terminal device 110.

[0250] At box 710, terminal device 110 sends first information to first network device 120 regarding the output of one or more first machine learning models available at terminal device 110.

[0251] At box 720, terminal device 110 receives from first network device 120 second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0252] At frame 730, terminal device 110 performs cell selection or cell reselection based at least on the second information.

[0253] In some example embodiments, performing cell selection or cell reselection may include selecting or reselecting a cell to connect to at least one network device based on performance, based on determining that the performance of at least one machine learning model available at at least one of one or more other network devices, according to second information indicating that the performance is acceptable.

[0254] In some example embodiments, terminal device 110 may also, based on determining that the performance of all one or more machine learning models available for the selected cell or the reselected cell is unacceptable relative to any of the first one or more machine learning models, stop using the first one or more machine learning models after cell selection or cell reselection. In some example embodiments, terminal device 110 may also, based on determining that the performance of at least one machine learning model available for the selected cell or the reselected cell is acceptable relative to the performance of the activated machine learning model among the first one or more machine learning models, continue using the activated machine learning model after cell selection or cell reselection, and / or store the second information, based on determining that the performance of all one or more machine learning models available for the selected cell or the reselected cell is unacceptable relative to the performance of the activated machine learning model, and that the performance of at least one machine learning model available for the selected cell or the reselected cell is acceptable relative to the performance of at least one of the first one or more machine learning models, stop using the activated machine learning model after cell selection or cell reselection, and start using at least one of the first one or more machine learning models after cell selection or cell reselection.

[0255] In some example embodiments, sending the first information may include at least one of the following: sending the first information based on a machine learning model update at the terminal device 110; sending the first information based on a change in channel conditions associated with the terminal device 110; sending the first information based on moving to a new cell; or sending the first information based on receiving an instruction from the first network device 120 instructing the terminal device 110 to send the first information.

[0256] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0257] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0258] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0259] In some example embodiments, terminal device 110 may also send an indication to first network device 120 of the reason for sending the first indication, wherein the reason includes at least one of the following: an update of the machine learning model at terminal device 110; a change in channel conditions associated with terminal device 110; or terminal device 110 moving to a new cell.

[0260] In some example embodiments, terminal device 110 may also receive system information broadcasts from another network device among one or more other devices, indicating one or more machine learning models available at the other network device.

[0261] In some example embodiments, terminal device 110 may also send an indication to first network device 120 regarding whether terminal device 110 is capable of using a first or more machine learning models.

[0262] Those skilled in the art will understand that the above references Figures 2 to 3F All the operations and features described also apply to method 700 and have similar effects.

[0263] Figure 8 A flowchart 800 illustrating a method implemented at a first network device according to some embodiments of the present disclosure is shown. For discussion purposes, reference will be made to... Figure 1A Method 800 is described from the perspective of the first network device 120.

[0264] At box 810, the first network device 120 receives from the terminal device 110 first information about the output of a first or more machine learning models available at the terminal device 110; and At box 820, the first network device 120 sends second information to the terminal device 110 regarding the performance of a second or more machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0265] In some example embodiments, the second information may indicate that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the second information may indicate that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model among the first or more machine learning models. In some example embodiments, the second information may indicate that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to at least one of the first or more machine learning models.

[0266] In some example embodiments, the first network device 120 may also send an instruction to the terminal device 110 to send first information based on at least one of the following: a machine learning model update at the network device at the first network device 120 or at one or more other network devices; or a determination that the terminal device 110 is moving toward the cell edge.

[0267] In some example embodiments, the first network device 120 may also receive from the terminal device 110 an indication of the reason for sending the first indication, wherein the reason includes at least one of the following: an update of the machine learning model at the terminal device 110; a change in the channel conditions associated with the terminal device 110; or the terminal device 110 moving to a new cell.

[0268] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0269] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0270] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0271] In some example embodiments, the first network device 120 may also receive an indication from the terminal device 110 regarding whether the terminal device 110 is capable of using a first or more machine learning models.

[0272] Those skilled in the art will understand that the above references Figures 2 to 3F All the operations and features described also apply to method 800 and have similar effects.

[0273] Figure 9 A flowchart 900 illustrating a method implemented at a terminal device according to some embodiments of the present disclosure is shown. For purposes of discussion, reference will be made to... Figure 1A Method 900 is described from the perspective of the terminal device 110.

[0274] At box 910, terminal device 110 sends first information to first network device 120 regarding the output of one or more first machine learning models available at terminal device 110.

[0275] At box 920, terminal device 110 receives from first network device 120 second information about the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0276] At box 930, terminal device 110 performs a handover from first network device 120 to second network device.

[0277] In some example embodiments, the terminal device 110 may also, based on determining that the performance of the second or more machine learning models is unacceptable relative to any of the first or more machine learning models, stop using the first or more machine learning models after a switch. In some example embodiments, the terminal device 110 may also, based on determining that the performance of at least one of the second or more machine learning models is acceptable relative to the activated machine learning model among the first or more machine learning models, continue using the activated machine learning model after a switch and / or store the second information, based on determining that the performance of the second or more machine learning models is unacceptable relative to the activated machine learning model, and that the performance of at least one of the second or more machine learning models is acceptable relative to at least one of the first or more machine learning models, stop using the activated machine learning model after a switch and start using at least one of the first or more machine learning models after a switch.

[0278] In some example embodiments, sending the first information may include at least one of the following: sending the first information based on a machine learning model update at the terminal device 110; sending the first information based on a change in channel conditions associated with the terminal device 110; sending the first information based on moving to a new cell; or sending the first information based on receiving an instruction from the first network device 120 instructing the terminal device 110 to send the first information.

[0279] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0280] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0281] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0282] In some example embodiments, terminal device 110 may also send an indication to first network device 120 of the reason for sending the first indication, wherein the reason includes at least one of the following: an update of the machine learning model at terminal device 110; a change in channel conditions associated with terminal device 110; or terminal device 110 moving to a new cell.

[0283] In some example embodiments, terminal device 110 may also send an indication to first network device 120 regarding whether terminal device 110 is capable of using a first or more machine learning models.

[0284] Those skilled in the art will understand that the above references Figure 4 All the operations and features described herein also apply to method 900 and have similar effects.

[0285] Figure 10 A flowchart 1000 illustrating a method implemented at a first network device according to some embodiments of the present disclosure is shown. For discussion purposes, reference will be made to... Figure 1A Method 1000 is described from the perspective of the first network device 120.

[0286] At box 1010, the first network device 120 receives first information from the terminal device 110 regarding the output of a first or more machine learning models available at the terminal device 110; At box 1020, the first network device 120 sends second information to the terminal device 110 regarding the performance of one or more second machine learning models available at the second network device, wherein the performance is obtained based on the first information; and At box 1030, the first network device 120 performs the handover process from the terminal device 110 to the second network device.

[0287] In some example embodiments, the second information may indicate that the performance of the second or more machine learning models relative to any of the first or more machine learning models is unacceptable. In some example embodiments, the second information may indicate that the performance of at least one of the second or more machine learning models relative to the activated machine learning model in the first or more machine learning models is acceptable. In some example embodiments, the second information may indicate that the performance of the second or more machine learning models relative to the activated machine learning model is unacceptable, and that the performance of at least one of the second or more machine learning models is acceptable relative to at least one of the first or more machine learning models.

[0288] In some example embodiments, the first network device 120 may also send an instruction to the terminal device 110 to send first information based on at least one of the following: a machine learning model update at the first network device 120 or at a network device in one or more other network devices; or determining that the terminal device 110 is moving toward the cell edge; or determining to switch the terminal device 110 from the first network device 120 to a network device in one or more other network devices.

[0289] In some example embodiments, the first network device 120 may also receive from the terminal device 110 an indication of the reason for sending the first indication, wherein the reason includes at least one of the following: an update of the machine learning model at the terminal device 110; a change in the channel conditions associated with the terminal device 110; or the terminal device 110 moving to a new cell.

[0290] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0291] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0292] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0293] In some example embodiments, the first network device 120 may also receive an indication from the terminal device 110 regarding whether the terminal device 110 is capable of using a first or more machine learning models.

[0294] Those skilled in the art will understand that the above references Figure 4 All the operations and features described herein also apply to method 1000 and have similar effects.

[0295] Figure 11 A flowchart 1100 illustrating a method implemented at a terminal device according to some embodiments of the present disclosure is shown. For discussion purposes, references will be made to... Figure 1A Method 1100 is described from the perspective of terminal device 110.

[0296] At box 1110, terminal device 110 sends first information to first network device 120 about the output of one or more machine learning models available at terminal device 110 based on changes in channel conditions associated with terminal device 110.

[0297] At box 1120, terminal device 110 receives from first network device 120 second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0298] In some example embodiments, terminal device 110 may also perform cell selection or cell reselection based at least on second information. In some example embodiments, performing cell selection or cell reselection may include selecting or reselecting a cell to connect to at least one network device based on performance, determined by the second information indicating that the performance of at least one machine learning model available at at least one of one or more other network devices is acceptable. In some example embodiments, terminal device 110 may also stop using the first or more machine learning models after cell selection or cell reselection, determined by the second information indicating that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, terminal device 110 may also continue using the activated machine learning model after cell selection or cell reselection, and / or store the second information, determined by the second information indicating that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the activated machine learning model in the first or more machine learning models. In some example embodiments, the terminal device 110 may also, based on determining second information, indicate that the performance of all or more machine learning models available for the selected cell or the reselected cell is unacceptable relative to the performance of the activated machine learning model, and that the performance of at least one machine learning model available for the selected cell or the reselected cell is acceptable relative to at least one of the first or more machine learning models, stop using the activated machine learning model after cell selection or cell reselection, and start using at least one of the first or more machine learning models after cell selection or cell reselection.

[0299] In some example embodiments, terminal device 110 may also perform a handover from first network device 120 to a second network device among one or more other network devices. In some example embodiments, terminal device 110 may also, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to any of the first or more machine learning models, cease using the first or more machine learning models after cell selection or cell reselection. In some example embodiments, terminal device 110 may also, based on determining that the performance of at least one machine learning model available at the second network device is acceptable relative to the activated machine learning model among the first or more machine learning models, continue using the activated machine learning model after handover, and / or store the second information, based on determining that the performance of at least one machine learning model available at the second network device is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available at the second network device is acceptable relative to at least one of the first or more machine learning models, cease using the activated machine learning model after handover, and begin using at least one of the first or more machine learning models after handover, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available at the second network device is acceptable relative to at least one of the first or more machine learning models.

[0300] In some example embodiments, terminal device 110 may also determine a change in channel conditions based on at least one of the following: a decrease in the reference signal received power (RSRP) level associated with the primary cell; an increase in the RSRP level associated with a neighboring cell; or determining that terminal device 110 is moving toward the cell edge.

[0301] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0302] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0303] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0304] In some example embodiments, terminal device 110 may also send an indication of a change in channel conditions to first network device 120.

[0305] In some example embodiments, changes in radio conditions may be associated with the mobility of terminal device 110.

[0306] In some example embodiments, terminal device 110 may also send an indication to first network device 120 regarding whether terminal device 110 is capable of using a first or more machine learning models.

[0307] Those skilled in the art will understand that the above references Figure 5 All the operations and features described also apply to method 1100 and have similar effects.

[0308] Figure 12 A flowchart 1200 illustrating a method implemented at a first network device according to some embodiments of the present disclosure is shown. For discussion purposes, reference will be made to... Figure 1A Method 1200 is described from the perspective of the first network device 120.

[0309] At box 1210, the first network device 120 receives from the terminal device 110 at least one of the following: first information about the output of a first or more machine learning models available at the terminal device 110; or an indication of a change in channel conditions associated with the terminal device 110.

[0310] At box 1220, the first network device 120 sends second information to the terminal device 110 regarding the performance of a second or more machine learning models available at the second network device, wherein the performance is obtained based on the first information.

[0311] In some example embodiments, the second information may indicate that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the second information may indicate that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model among the first or more machine learning models. In some example embodiments, the second information may indicate that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of at least one of the first or more machine learning models.

[0312] In some example embodiments, the first network device 120 may also perform a handover process from the terminal device 110 to the second network device.

[0313] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0314] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0315] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0316] In some example embodiments, changes in radio conditions may be associated with the mobility of terminal device 110.

[0317] In some example embodiments, the first network device 120 may also receive an indication from the terminal device 110 regarding whether the terminal device 110 is capable of using a first or more machine learning models.

[0318] Those skilled in the art will understand that the above references Figure 5 All the operations and features described also apply to method 1200 and have similar effects.

[0319] Figure 13 A flowchart 1300 illustrating a method implemented at a terminal device according to some embodiments of the present disclosure is shown. For discussion purposes, reference will be made to... Figure 1A Method 1300 is described from the perspective of the terminal device 110.

[0320] At box 1310, terminal device 110, based on an update of the machine learning model at terminal device 110, sends first information to first network device 120 regarding the output of one or more first machine learning models available at terminal device 110; and At box 1320, terminal device 110 receives from first network device 120 second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0321] In some example embodiments, terminal device 110 may also perform cell selection or cell reselection based at least on second information. In some example embodiments, performing cell selection or cell reselection may include selecting or reselecting a cell to connect to at least one network device based on performance, determined by the second information indicating that the performance of at least one machine learning model available at at least one of one or more other network devices is acceptable. In some example embodiments, terminal device 110 may also stop using the first or more machine learning models after cell selection or cell reselection, determined by the second information indicating that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, terminal device 110 may also continue using the activated machine learning model after cell selection or cell reselection, and / or store the second information, determined by the second information indicating that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the activated machine learning model among the first or more machine learning models. In some example embodiments, the terminal device 110 may also, based on determining second information, indicate that the performance of all or more machine learning models available for the selected cell or the reselected cell is unacceptable relative to the performance of the activated machine learning model, and that the performance of at least one machine learning model available for the selected cell or the reselected cell is acceptable relative to at least one of the first or more machine learning models, stop using the activated machine learning model after cell selection or cell reselection, and start using at least one of the first or more machine learning models after cell selection or cell reselection.

[0322] In some example embodiments, terminal device 110 may also perform a handover from first network device 120 to a second network device among one or more other network devices. In some example embodiments, terminal device 110 may also, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to any of the first or more machine learning models, stop using the first or more machine learning models after the handover. In some example embodiments, terminal device 110 may also, based on determining that the performance of at least one machine learning model available at the second network device is acceptable relative to the activated machine learning model among the first or more machine learning models, continue using the activated machine learning model after the handover, and / or store the second information, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available at the second network device is acceptable relative to at least one of the first or more machine learning models. In some example embodiments, terminal device 110 may also, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available at the second network device is acceptable relative to at least one of the first or more machine learning models, stop using the activated machine learning model after the handover, and start using at least one of the first or more machine learning models after the handover.

[0323] In some example embodiments, terminal device 110 may also send an indication to first network device 120 regarding whether terminal device 110 is capable of using a first or more machine learning models.

[0324] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0325] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0326] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0327] In some example embodiments, terminal device 110 may also send an instruction to the first network device 120 indicating an update of the machine learning model at terminal device 110.

[0328] Those skilled in the art will understand that the above references Figure 6 All the operations and features described herein also apply to method 1300 and have similar effects.

[0329] Figure 14A flowchart 1400 illustrating a method implemented at a first network device according to some embodiments of the present disclosure is shown. For discussion purposes, reference will be made to... Figure 1A Method 1400 is described from the perspective of the first network device 120.

[0330] At box 1410, the first network device 120 receives from the terminal device 110 at least one of the following: first information about the output of a first or more machine learning models available at the terminal device 110; or an indication of an update to the machine learning model at the terminal device 110.

[0331] At box 1420, the first network device 120 sends second information to the terminal device 110 regarding the performance of a second or more machine learning models available at the second network device, wherein the performance is obtained based on the first information.

[0332] In some example embodiments, the second information may indicate that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the second information may indicate that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model in the first or more machine learning models. In some example embodiments, the second information may indicate that the performance of all or more machine learning models available for the locked-selected or reselected cell is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available for the locked-selected or reselected cell is acceptable relative to the performance of at least one of the first or more machine learning models.

[0333] In some example embodiments, the first network device 120 may also perform a handover process from the terminal device 110 to the second network device.

[0334] In some example embodiments, the output may correspond to channel state information encoded by a first or more machine learning models.

[0335] In some example embodiments, a second or more machine learning models can be used as a decoder for decoding encoded channel state information.

[0336] In some example embodiments, performance may be associated with a performance indicator for decoding encoded channel state information.

[0337] In some example embodiments, the first network device 120 may also receive an indication from the terminal device 110 regarding whether the terminal device 110 is capable of using a first or more machine learning models.

[0338] Those skilled in the art will understand that the above references Figure 6 All the operations and features described herein also apply to method 1400 and have similar effects.

[0339] In some example embodiments, the apparatus capable of performing method 700 (e.g., terminal device 110) may include components for performing the corresponding steps of method 700. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0340] In some example embodiments, the apparatus includes: means for sending first information to a first network device about the output of a first or more machine learning models available at a terminal device; means for receiving from the first network device second information about the performance of a second or more machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information; and means for performing cell selection or cell reselection based at least on the second information.

[0341] In some example embodiments, the components for performing cell selection or cell reselection include components for selecting or reselecting a cell to be connected to at least one network device based on performance, determined by second information indicating that the performance of at least one machine learning model available at at least one of one or more other network devices is acceptable.

[0342] In some example embodiments, the apparatus further includes at least one of the following: components for stopping the use of the first or more machine learning models after cell selection or cell reselection, based on determining that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any machine learning model in the first or more machine learning models; components for continuing to use the activated machine learning model after cell selection or cell reselection, based on determining that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model in the first or more machine learning models, based on the second information; or components for stopping the use of the activated machine learning model after cell selection or cell reselection, and starting to use at least one machine learning model in the first or more machine learning models, based on determining that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the performance of the activated machine learning model, and that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of at least one machine learning model in the first or more machine learning models, based on the second information, and starting to use at least one machine learning model in the first or more machine learning models after cell selection or cell reselection.

[0343] In some example embodiments, the component for sending the first information includes at least one of the following: a component for sending the first information based on a machine learning model update at the terminal device; a component for sending the first information based on a change in channel conditions associated with the terminal device; a component for sending the first information based on moving to a new cell; or a component for sending the first information based on receiving an instruction from a first network device instructing the terminal device to send the first information.

[0344] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0345] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0346] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0347] In some example embodiments, the apparatus further includes: a component for sending an indication to a first network device of a reason for sending the first indication, wherein the reason includes at least one of the following: an update of a machine learning model at the terminal device; a change in channel conditions associated with the terminal device; or the terminal device moving to a new cell.

[0348] In some example embodiments, the terminal device also includes a component for receiving system information broadcasts from another network device among one or more other devices, indicating that one or more machine learning models are available at the other network device.

[0349] In some example embodiments, the apparatus further includes components for sending an indication to a first network device whether a terminal device is capable of using a first or more machine learning models.

[0350] In some embodiments, the apparatus further includes components for performing other steps in some embodiments of method 700. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the apparatus together with the at least one processor.

[0351] In some example embodiments, the apparatus capable of performing method 800 (e.g., the first network device 120) may include components for performing the corresponding steps of method 800. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0352] In some example embodiments, the apparatus includes: components for receiving first information from a terminal device regarding the output of a first or more machine learning models available at the terminal device; and components for sending to the terminal device second information regarding the performance of a second or more machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0353] In some example embodiments, the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the second information indicates that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model among the first or more machine learning models. In some example embodiments, the second information indicates that the performance of all or more machine learning models available for the locked-selected or reselected cell is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available for the locked-selected or reselected cell is acceptable relative to the performance of at least one of the first or more machine learning models.

[0354] In some example embodiments, the first network device further includes components for sending an instruction to the terminal device to send first information based on at least one of the following: a machine learning model update at the first network device or at a network device in one or more other network devices; or determining that the terminal device is moving toward the cell edge.

[0355] In some example embodiments, the first network device further includes: a component for receiving from the terminal device an indication of the reason for sending the first indication, wherein the reason includes at least one of the following: an update of the machine learning model at the terminal device; a change in channel conditions associated with the terminal device; or the terminal device moving to a new cell.

[0356] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0357] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0358] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0359] In some example embodiments, the first network device further includes a component for receiving an indication from a terminal device as to whether the terminal device is capable of using a first or more machine learning models.

[0360] In some embodiments, the device further includes components for performing other steps in some embodiments of method 800. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the device together with the at least one processor.

[0361] In some example embodiments, the apparatus capable of performing method 900 (e.g., terminal device 110) may include components for performing the corresponding steps of method 900. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0362] In some example embodiments, the apparatus includes: components for sending first information to a first network device regarding the output of a first or more machine learning models available at a terminal device; components for receiving from the first network device second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information; and components for performing a handover from the first network device to the second network device.

[0363] In some example embodiments, the apparatus further includes at least one of the following: components for indicating, based on determining second information, that the performance of a second or more machine learning models relative to any of the first or more machine learning models is unacceptable, and stopping the use of the first or more machine learning models after a switch; components for indicating, based on determining second information, that the performance of at least one of the second or more machine learning models relative to the activated machine learning model in the first or more machine learning models is acceptable, and continuing to use the activated machine learning model after a switch and / or storing the second information; or components for indicating, based on determining second information, that the performance of a second or more machine learning models relative to the activated machine learning model is unacceptable, and that the performance of at least one of the second or more machine learning models relative to at least one of the first or more machine learning models is acceptable, stopping the use of the activated machine learning model after a switch, and starting to use at least one of the first or more machine learning models after a switch.

[0364] In some example embodiments, the component for sending the first information includes at least one of the following: a component for sending the first information based on a machine learning model update at the terminal device; a component for sending the first information based on a change in channel conditions associated with the terminal device; a component for sending the first information based on moving to a new cell; or a component for sending the first information based on receiving an instruction from a first network device instructing the terminal device to send the first information.

[0365] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0366] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0367] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0368] In some example embodiments, the terminal device further includes a component for sending an indication to the first network device of the reason for sending the first indication, wherein the reason includes at least one of the following: an update of the machine learning model at the terminal device; a change in the channel conditions associated with the terminal device; or the terminal device moving to a new cell.

[0369] In some example embodiments, the apparatus further includes components for sending an indication to a first network device whether a terminal device is capable of using a first or more machine learning models.

[0370] In some embodiments, the apparatus further includes components for performing other steps in some embodiments of method 900. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the apparatus together with the at least one processor.

[0371] In some example embodiments, the apparatus capable of performing method 1000 (e.g., the first network device 120) may include components for performing the corresponding steps of method 1000. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0372] In some example embodiments, the apparatus includes: components for receiving first information from a terminal device regarding the output of a first or more machine learning models available at the terminal device; components for sending to the terminal device second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information; and components for performing a handover process from the terminal device to the second network device.

[0373] In some example embodiments, the second information indicates that the performance of the second or more machine learning models relative to any of the first or more machine learning models is unacceptable. In some example embodiments, the second information indicates that the performance of at least one of the second or more machine learning models is acceptable relative to the performance of the activated machine learning model in the first or more machine learning models. In some example embodiments, the second information indicates that the performance of the second or more machine learning models relative to the activated machine learning model is unacceptable, and that the performance of at least one of the second or more machine learning models is acceptable relative to the performance of at least one of the first or more machine learning models.

[0374] In some example embodiments, the apparatus further includes components for sending an instruction to a terminal device to send first information based on at least one of the following: a machine learning model update at a first network device or at a network device in one or more other network devices; or determining that the terminal device is moving toward the cell edge; or determining to switch the terminal device from the first network device to a network device in one or more other network devices.

[0375] In some example embodiments, the apparatus further includes: a component for receiving from a terminal device an indication of a reason for sending a first indication, wherein the reason includes at least one of the following: an update of a machine learning model at the terminal device; a change in channel conditions associated with the terminal device; or the terminal device moving to a new cell.

[0376] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0377] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0378] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0379] In some example embodiments, the apparatus further includes a component for receiving an indication from the terminal device as to whether the terminal device is capable of using a first or more machine learning models.

[0380] In some embodiments, the apparatus further includes components for performing other steps in some embodiments of method 1000. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the apparatus together with the at least one processor.

[0381] In some example embodiments, the apparatus capable of performing method 1100 (e.g., terminal device 110) may include components for performing the corresponding steps of method 1100. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0382] In some example embodiments, the apparatus includes: means for transmitting first information to a first network device regarding the output of a first or more machine learning models available at the terminal device based on a change in channel conditions associated with the terminal device; and means for receiving from the first network device second information regarding the performance of a second or more machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0383] In some example embodiments, the apparatus further includes components for performing cell selection or cell reselection at least based on second information. In some example embodiments, the components for performing cell selection or cell reselection include: components for selecting or reselecting a cell to be connected to at least one network device based on performance, determined by the second information indicating that the performance of at least one machine learning model available at at least one of one or more other network devices is acceptable. In some example embodiments, the apparatus further includes components for stopping the use of one or more machine learning models after cell selection or cell reselection, determined by the second information indicating that the performance of all one or more machine learning models available for the selected or reselected cell is unacceptable relative to any machine learning model in the first one or more machine learning models. In some example embodiments, the apparatus further includes components for continuing to use the activated machine learning model and / or storing the second information after cell selection or cell reselection, determined by the second information indicating that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model in the first one or more machine learning models. In some example embodiments, the apparatus further includes components for determining, based on second information indicating that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the performance of the activated machine learning model, and that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to at least one of the first or more machine learning models, stopping the use of the activated machine learning model after cell selection or cell reselection, and starting the use of at least one of the first or more machine learning models after cell selection or cell reselection.

[0384] In some example embodiments, the apparatus further includes components for performing a handover from a first network device to a second network device among one or more other network devices. In some example embodiments, the apparatus further includes components for stopping the use of the first or more machine learning models after cell selection or cell reselection, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the apparatus further includes components for continuing to use the activated machine learning model after handover and / or storing the second information, based on determining that the performance of at least one machine learning model available at the second network device is acceptable relative to the activated machine learning model among the first or more machine learning models. In some example embodiments, the apparatus further includes components for stopping the use of the activated machine learning model after handover and starting to use at least one of the first or more machine learning models after handover, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available at the second network device is acceptable relative to at least one of the first or more machine learning models.

[0385] In some example embodiments, the apparatus further includes: a component for determining the change in channel conditions based on at least one of the following: a decrease in the reference signal received power (RSRP) level associated with the primary cell; an increase in the RSRP level associated with a neighboring cell; or determining that the terminal device is moving toward the cell edge.

[0386] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0387] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0388] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0389] In some example embodiments, the apparatus further includes a component for sending an indication of a change in channel conditions to a first network device.

[0390] In some example embodiments, changes in radio conditions are associated with the mobility of the terminal device.

[0391] In some example embodiments, the apparatus further includes components for sending an indication to a first network device whether a terminal device is capable of using a first or more machine learning models.

[0392] In some embodiments, the apparatus further includes components for performing other steps in some embodiments of method 1100. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the apparatus together with the at least one processor.

[0393] In some example embodiments, the apparatus capable of performing method 1200 (e.g., the first network device 120) may include components for performing the corresponding steps of method 1200. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0394] In some example embodiments, the apparatus includes: components for receiving from a terminal device at least one of the following: first information regarding the output of a first or more machine learning models available at the terminal device; or an indication of a change in channel conditions associated with the terminal device; and components for sending to the terminal device second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0395] In some example embodiments, the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the second information indicates that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model among the first or more machine learning models. In some example embodiments, the second information indicates that the performance of all or more machine learning models available for the locked-selected or reselected cell is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available for the locked-selected or reselected cell is acceptable relative to the performance of at least one of the first or more machine learning models.

[0396] In some example embodiments, the apparatus also includes components for performing a handover process from a terminal device to a second network device.

[0397] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0398] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0399] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0400] In some example embodiments, changes in radio conditions are associated with the mobility of the terminal device.

[0401] In some example embodiments, the apparatus further includes a component for receiving an indication from the terminal device as to whether the terminal device is capable of using a first or more machine learning models.

[0402] In some embodiments, the apparatus further includes components for performing other steps in some embodiments of method 1200. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the apparatus together with the at least one processor.

[0403] In some example embodiments, the apparatus capable of performing method 1300 (e.g., terminal device 110) may include components for performing the corresponding steps of method 1300. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0404] In some example embodiments, the apparatus includes: means for sending first information to a first network device about the output of a first or more machine learning models available at the terminal device based on machine learning model updates at the terminal device; and means for receiving from the first network device second information about the performance of a second or more machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0405] In some example embodiments, the apparatus further includes components for performing cell selection or cell reselection at least based on second information. In some example embodiments, the components for performing cell selection or cell reselection include components for selecting or reselecting a cell to be connected to at least one network device based on performance, determined by the second information indicating that the performance of at least one machine learning model available at at least one of one or more other network devices is acceptable. In some example embodiments, the apparatus further includes components for stopping the use of one or more machine learning models after cell selection or cell reselection, determined by the second information indicating that the performance of all one or more machine learning models available for the selected or reselected cell is unacceptable relative to any machine learning model in the first one or more machine learning models. In some example embodiments, the apparatus further includes components for continuing to use the activated machine learning model and / or storing the second information after cell selection or cell reselection, determined by the second information indicating that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model in the first one or more machine learning models. In some example embodiments, the apparatus further includes components for determining, based on second information indicating that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to the performance of the activated machine learning model, and that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to at least one of the first or more machine learning models, stopping the use of the activated machine learning model after cell selection or cell reselection, and starting the use of at least one of the first or more machine learning models after cell selection or cell reselection.

[0406] In some example embodiments, the apparatus further includes components for performing a handover from a first network device to a second network device among one or more other network devices. In some example embodiments, the apparatus further includes components for stopping the use of the first or more machine learning models after the handover, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the apparatus further includes components for continuing to use the activated machine learning model and / or storing the second information after the handover, based on determining that the performance of at least one machine learning model available at the second network device is acceptable relative to the activated machine learning model among the first or more machine learning models. In some example embodiments, the apparatus further includes components for stopping the use of the activated machine learning model after the handover, and starting to use at least one of the first or more machine learning models after the handover, based on determining that the performance of all or more machine learning models available at the second network device is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available at the second network device is acceptable relative to at least one of the first or more machine learning models.

[0407] In some example embodiments, the apparatus further includes components for sending an indication to a first network device whether a terminal device is capable of using a first or more machine learning models.

[0408] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0409] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0410] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0411] In some example embodiments, the apparatus further includes components for sending instructions on updating the machine learning model to a first network device at the terminal device.

[0412] In some embodiments, the apparatus further includes components for performing other steps in some embodiments of method 1300. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the apparatus together with the at least one processor.

[0413] In some example embodiments, the apparatus capable of performing method 1400 (e.g., the first network device 120) may include components for performing the corresponding steps of method 1400. The components may be implemented in any suitable form. For example, the components may be implemented in a circuit or software module.

[0414] In some example embodiments, the apparatus includes: components for receiving from a terminal device at least one of the following: first information regarding the output of a first or more machine learning models available at the terminal device; or an indication of an update to a machine learning model at the terminal device; and components for sending to the terminal device second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0415] In some example embodiments, the second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models. In some example embodiments, the second information indicates that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model in the first or more machine learning models. In some example embodiments, the second information indicates that the performance of all or more machine learning models available for the locked-selected or reselected cell is unacceptable relative to the activated machine learning model, and that the performance of at least one machine learning model available for the locked-selected or reselected cell is acceptable relative to the performance of at least one of the first or more machine learning models.

[0416] In some example embodiments, the apparatus also includes components for performing a handover process from a terminal device to a second network device.

[0417] In some example embodiments, the output corresponds to channel state information encoded by a first or more machine learning models.

[0418] In some example embodiments, a second or more machine learning models are used as decoders for decoding encoded channel state information.

[0419] In some example embodiments, performance is associated with a performance indicator for decoding encoded channel state information.

[0420] In some example embodiments, the apparatus further includes a component for receiving an indication from the terminal device as to whether the terminal device is capable of using a first or more machine learning models.

[0421] In some embodiments, the apparatus further includes components for performing other steps in some embodiments of method 1400. In some embodiments, the components include at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to cause execution of the apparatus together with the at least one processor.

[0422] Figure 15 A simplified block diagram of a device 1500 suitable for implementing some example embodiments of the present disclosure is shown. The device 1500 can be provided to implement a communication device, such as... Figure 1A The terminal device 110, the first network device 120, or the second network device 130 are shown. As shown, device 1500 includes one or more processors 1510, one or more memories 1520 coupled to processor 1510, and one or more communication modules 1540 coupled to processor 1510.

[0423] Communication module 1540 is used for bidirectional communication. Communication module 1540 has at least one antenna to support communication. The communication interface can represent any interface necessary for communication with other network elements.

[0424] As a non-limiting example, processor 1510 can be any type suitable for a local technology network and can include one or more of the following: general-purpose computer, special-purpose computer, microprocessor, digital signal processor (DSP), and processor based on a multi-core processor architecture. Device 1500 can have multiple processors, such as application-specific integrated circuit chips that are time-dependent on a clock that synchronizes with the main processor.

[0425] Memory 1520 may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, read-only memory (ROM) 1524, electrically programmable read-only memory (EPROM), flash memory, hard disk, compact disc (CD), digital video disc (DVD), and other magnetic and / or optical storage. Examples of volatile memories include, but are not limited to, random access memory (RAM) 1522 and other volatile memories that do not persist during power outages.

[0426] Computer program 1530 includes computer-executable instructions that are executed by the associated processor 1510. Program 1530 may be stored in ROM 1524. Processor 1510 may perform any suitable actions and processes by loading program 1530 into RAM 1522.

[0427] Embodiments of this disclosure can be implemented via program 1530, enabling device 1500 to perform as described in the reference. Figures 2 to 6Any process discussed in this disclosure. Embodiments of this disclosure may also be implemented by hardware or by a combination of software and hardware.

[0428] In some example embodiments, program 1530 may be tangibly contained in a computer-readable medium, which may be included in device 1500 (such as in memory 1520) or other storage device accessible by device 1500. Device 1500 may load program 1530 from the computer-readable medium into RAM 1522 for execution. The computer-readable medium may include any type of tangible non-volatile storage device, such as ROM, EPROM, flash memory, hard disk, CD, DVD, etc.

[0429] Figure 16 A block diagram of an example of a computer-readable medium 1600 according to some exemplary embodiments of the present disclosure is shown. The computer-readable medium 1600 has a program 1530 stored thereon. It should be noted that, although in Figure 16 The computer-readable medium 1600 is depicted in the form of a CD or DVD, but the computer-readable medium 1600 may be any other form suitable for carrying or storing the program 1530.

[0430] Generally, the various embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of this disclosure are shown and described as block diagrams, flowcharts, or using some other graphical representation, it should be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware, or controllers or other computing devices, or some combination thereof, as non-limiting examples.

[0431] This disclosure also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer-executable instructions, such as those included in program modules, which are executed in a device on a target real or virtual processor to perform the functions described above. Figures 7 to 14 Any of the methods described herein. Typically, a program module includes routines, programs, libraries, objects, classes, components, data structures, etc., that perform a specific task or implement a specific abstract data type. In various embodiments, the functionality of a program module can be combined or split among program modules as needed. The machine-executable instructions for a program module can execute on a local or distributed device. In a distributed device, a program module can reside on both local and remote storage media.

[0432] Program code used to perform the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a stand-alone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0433] In the context of this disclosure, computer program code or related data may be carried by any suitable carrier wave to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carrier waves include signals, computer-readable media, etc.

[0434] Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable media can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any suitable combination thereof. More specific examples of computer-readable storage media will include electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable optical disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. As used herein, the term "non-transitory" is a limitation of the medium itself (i.e., tangible, not signaling), not a limitation of the persistence of data storage (e.g., RAM versus ROM).

[0435] Furthermore, although operations are described in a specific order, this should not be construed as requiring that such operations be performed in the specific order shown or sequentially, or requiring that all shown operations be performed to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure, but rather as a description of features that may be specific to particular embodiments. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0436] Although this disclosure has been described in language specific to structural features and / or methodological actions, it should be understood that the disclosure as defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are disclosed as exemplary forms for implementing the claims.

[0437] The embodiments disclosed herein provide the following examples.

[0438] Example 1. A terminal device for communication, comprising: At least one processor; and At least one memory stores instructions that, when executed by the at least one processor, cause the terminal device to at least: Based on changes in channel conditions associated with the terminal device, first information regarding the outputs of one or more first machine learning models available at the terminal device is sent to a first network device; and Receive second information from the first network device regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0439] Example 2. The terminal device according to Example 1, wherein the terminal device is further configured to: Cell selection or cell reselection is performed based at least on the second information.

[0440] Example 3. The terminal device according to Example 2, wherein performing the cell selection or the cell reselection includes: Based on determining that the performance of at least one machine learning model available at at least one of the one or more other network devices, indicated by the second information, is acceptable, the cell to be connected to the at least one network device is selected or reselected based on the performance.

[0441] Example 4. The terminal device according to Example 2, wherein the terminal device is further configured to include at least one of the following: Based on the determination that the performance of all one or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first one or more machine learning models, the first one or more machine learning models are discontinued after the cell selection or the cell reselection. Based on the determination that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model among the first or more machine learning models, the activated machine learning model continues to be used after the cell selection or cell reselection, and / or the second information is stored; or Based on the determination that the performance of all or more machine learning models available for the selected or reselected cell relative to the activated machine learning model is unacceptable, and the performance of at least one machine learning model available for the selected or reselected cell relative to at least one of the first or more machine learning models is acceptable, the activated machine learning model is stopped after the cell selection or the cell reselection, and the at least one of the first or more machine learning models is started after the cell selection or the cell reselection.

[0442] Example 5. The terminal device according to Example 1, wherein the terminal device is further configured to: Perform a handover from the first network device to a second network device among the one or more other network devices.

[0443] Example 6. The terminal device according to Example 5, wherein the terminal device is further configured to include at least one of the following: Based on the determination that the performance of all one or more machine learning models available at the second network device according to the second information is unacceptable relative to any of the first one or more machine learning models, the first one or more machine learning models are discontinued after the cell selection or the cell reselection. Based on the determination that the performance of at least one machine learning model available at the second network device, indicated by the second information, is acceptable relative to the performance of the activated machine learning model among the first one or more machine learning models, the activated machine learning model continues to be used after the switch, and / or the second information is stored; or Based on the determination that the performance of all or more machine learning models available at the second network device relative to the activated machine learning model is unacceptable, and that the performance of at least one machine learning model available at the second network device relative to at least one of the first or more machine learning models is acceptable, the activated machine learning model is stopped after the switch, and the at least one of the first or more machine learning models is started after the switch.

[0444] Example 7. The terminal device according to any one of Examples 1 to 6, wherein the terminal device is further configured to: The change in the channel conditions is determined based on at least one of the following: A decrease in the level of the reference signal received power (RSRP) associated with the primary cell; An increase in RSRP levels associated with neighboring cells; or It is determined that the terminal device is moving toward the edge of the cell.

[0445] Example 8. A terminal device according to any one of Examples 1 to 7, wherein the output corresponds to channel state information encoded by the first one or more machine learning models.

[0446] Example 9. A terminal device according to any one of Examples 1 to 8, wherein the second or more machine learning models are used as a decoder for decoding encoded channel state information.

[0447] Example 10. A terminal device according to any one of Examples 1 to 9, wherein the performance is associated with a performance indicator for decoding encoded channel state information.

[0448] Example 11. The terminal device according to any one of Examples 1 to 10, wherein the terminal device is further configured to: Send an indication of the change in the channel conditions to the first network device.

[0449] Example 12. A terminal device according to any one of Examples 1 to 11, wherein the change in radio conditions is associated with the mobility of the terminal device.

[0450] Example 13. The terminal device according to any one of Examples 1 to 12, wherein the terminal device is further configured to: Send an indication to the first network device regarding whether the terminal device can use the first one or more machine learning models.

[0451] Example 14. A first network device for communication, comprising: At least one processor; and At least one memory stores instructions that, when executed by the at least one processor, cause the first network device to at least: Receive at least one of the following from the terminal device: First information regarding the output of the first or more machine learning models available at the terminal device; or Instructions for changes in channel conditions associated with the terminal device; and Send to the terminal device second information regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0452] Example 15. The first network device according to Example 14, wherein one of the following: The second information indicates that the performance of all or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first or more machine learning models; or The second information indicates that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model among the first or more machine learning models; or The second information indicates that the performance of all or more machine learning models available for the selected cell or the reselected cell is unacceptable relative to the performance of the activated machine learning model, and that the performance of at least one machine learning model available for the selected cell or the reselected cell is acceptable relative to the performance of at least one of the first or more machine learning models.

[0453] Example 16. The first network device according to Example 14 or 15, wherein the first network device is further configured to: Perform the handover process from the terminal device to the second network device.

[0454] Example 17. A first network device according to any one of Examples 14 to 16, wherein the output corresponds to channel state information encoded by the first one or more machine learning models.

[0455] Example 18. A first network device according to any one of Examples 14 to 17, wherein the second or more machine learning models are used as a decoder for decoding encoded channel state information.

[0456] Example 19. A first network device according to any one of Examples 14 to 18, wherein the performance is associated with a performance indicator for decoding encoded channel state information.

[0457] Example 20. A first network device according to any one of Examples 14 to 19, wherein the change in radio conditions is associated with the mobility of the terminal device.

[0458] Example 21. A first network device according to any one of Examples 14 to 20, wherein the first network device is further configured to: Receive an indication from the terminal device regarding whether the terminal device can use the first one or more machine learning models.

[0459] Example 22. A method for communication, comprising: Based on changes in channel conditions associated with the terminal device, first information regarding the outputs of one or more first machine learning models available at the terminal device is transmitted from the terminal device to the first network device; and The terminal device receives second information from the first network device regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0460] Example 23. A method for communication, comprising: At the first network device, at least one of the following is received from the terminal device: First information regarding the output of the first or more machine learning models available at the terminal device; or Indication of a change in channel conditions associated with the terminal device; and The first network device sends second information to the terminal device regarding the performance of a second or more machine learning models available at a second network device, wherein the performance is obtained based on the first information.

[0461] Example 24. An apparatus for communication, comprising: Components for transmitting, at the terminal device, first information about the outputs of one or more machine learning models available at the terminal device, to a first network device based on changes in channel conditions associated with the terminal device; and A component for receiving, at the terminal device, second information about the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

[0462] Example 25. An apparatus for communication, comprising: Components for receiving at least one of the following from a terminal device at a first network device: First information regarding the output of the first or more machine learning models available at the terminal device; or Indication of a change in channel conditions associated with the terminal device; and Components for sending, at the first network device, second information about the performance of one or more second machine learning models available at a second network device to the terminal device, wherein the performance is obtained based on the first information.

[0463] Example 26. A non-transitory computer-readable medium comprising program instructions that, when executed by a device, cause the device to perform at least the method according to Example 22 or 23.

Claims

1. A terminal device for communication, comprising: At least one processor; as well as At least one memory stores instructions that, when executed by the at least one processor, cause the terminal device to at least: Based on a change in channel conditions associated with the terminal device, first information about the output of a first or more machine learning models available at the terminal device is sent to a first network device. as well as Receive second information from the first network device regarding the performance of one or more second machine learning models available at one or more other network devices, wherein the performance is obtained based on the first information.

2. The terminal device according to claim 1, wherein the terminal device is further configured to: Cell selection or cell reselection is performed based at least on the second information.

3. The terminal device according to claim 2, wherein performing the cell selection or the cell reselection includes: Based on determining that the performance of at least one machine learning model available at at least one of the one or more other network devices, indicated by the second information, is acceptable, the cell to be connected to the at least one network device is selected or reselected based on the performance.

4. The terminal device according to claim 2, wherein the terminal device is further configured to include at least one of the following: Based on the determination that the performance of all one or more machine learning models available for the selected or reselected cell is unacceptable relative to any of the first one or more machine learning models, the first one or more machine learning models are discontinued after the cell selection or the cell reselection. Based on the determination that the performance of at least one machine learning model available for the selected or reselected cell is acceptable relative to the performance of the activated machine learning model among the first or more machine learning models, the activated machine learning model continues to be used after the cell selection or cell reselection, and / or the second information is stored; or Based on the determination that the performance of all or more machine learning models available for the selected or reselected cell relative to the activated machine learning model is unacceptable, and the performance of at least one machine learning model available for the selected or reselected cell relative to at least one of the first or more machine learning models is acceptable, the activated machine learning model is stopped after the cell selection or the cell reselection, and the at least one of the first or more machine learning models is started after the cell selection or the cell reselection.

5. The terminal device according to claim 1, wherein the terminal device is further configured to: Perform a handover from the first network device to a second network device among the one or more other network devices.

6. The terminal device according to claim 5, wherein the terminal device is further configured to include at least one of the following: Based on the determination that the performance of all one or more machine learning models available at the second network device according to the second information is unacceptable relative to any of the first one or more machine learning models, the first one or more machine learning models are discontinued after the cell selection or the cell reselection. Based on the determination that the performance of at least one machine learning model available at the second network device, indicated by the second information, is acceptable relative to the performance of the activated machine learning model among the first one or more machine learning models, the activated machine learning model continues to be used after the switch, and / or the second information is stored; or Based on the determination that the performance of all or more machine learning models available at the second network device relative to the activated machine learning model is unacceptable, and that the performance of at least one machine learning model available at the second network device relative to at least one of the first or more machine learning models is acceptable, the activated machine learning model is stopped after the switch, and the at least one of the first or more machine learning models is started after the switch.

7. The terminal device according to any one of claims 1 to 6, wherein the terminal device is further configured to: The change in the channel conditions is determined based on at least one of the following: A decrease in the level of the reference signal received power (RSRP) associated with the primary cell; An increase in RSRP levels associated with neighboring cells; or It is determined that the terminal device is moving toward the edge of the cell.

8. The terminal device according to any one of claims 1 to 6, wherein the output corresponds to channel state information encoded by the first one or more machine learning models.

9. The terminal device according to any one of claims 1 to 6, wherein the second or more machine learning models are used as decoders for decoding encoded channel state information.

10. The terminal device according to any one of claims 1 to 6, wherein the performance is associated with a performance indicator for decoding encoded channel state information.

Images