UE-driven sequential training

Abstract

This disclosure provides systems, methods, and devices for wireless communication that support UE-driven sequential training. In a first aspect, a method of wireless communication includes obtaining channel state information data associated with a second network node; training a shared UE encoder based on the channel state information data and based on a decoder to generate a sequential training data set; and sending the sequential training data set to a third network node. Other aspects and features are also claimed and described.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of International Patent Application No. PCT / CN2022 / 112155, filed on August 12, 2022, entitled “UE-DRIVEN SEQUENTIAL TRAINING”, which is expressly incorporated herein by reference in its entirety. Technical Field

[0003] Aspects of the present disclosure relate generally to wireless communication systems, and more particularly to sequential training for encoding and decoding.Some features may enable and provide improved communications, including generating a shared or common encoder and decoder for cross-node channel state feedback. Background Art

[0004] Wireless communication networks are widely deployed to provide various communication services such as voice, video, packet data, messaging, broadcast, etc. These wireless networks may be multiple-access networks capable of supporting multiple users by sharing the available network resources. Such networks may be multiple-access networks capable of supporting communication for multiple users by sharing the available network resources.

[0005] A wireless communication network may include several components. These components may include wireless communication devices, such as a base station (or Node B) that may support communication for several user equipments (UEs). A UE may communicate with a base station via a downlink and an uplink. A downlink (or forward link) refers to a communication link from a base station to a UE, and an uplink (or reverse link) refers to a communication link from a UE to a base station.

[0006] A base station may send data and control information to a UE on a downlink, or receive data and control information from a UE on an uplink. On the downlink, transmissions from a base station may encounter interference caused by transmissions from neighboring base stations or other wireless radio frequency (RF) transmitters. On the uplink, transmissions from a UE may encounter interference from uplink transmissions from other UEs communicating with neighboring base stations or from other wireless RF transmitters. Such interference may degrade performance on both the downlink and uplink.

[0007] As the demand for mobile broadband access continues to grow, the potential for interference and congested networks grows with more UEs accessing long-range wireless communication networks and more short-range wireless systems deployed in communities. Research and development continues to advance wireless technologies to not only meet the growing demand for mobile broadband access, but also to improve and enhance the user experience with mobile communications. Summary of the invention

[0008] The following summarizes some aspects of the present disclosure to provide a basic understanding of the technology discussed. This summary is not an exhaustive overview of all expected features of the present disclosure, and is neither intended to identify key or important elements of all aspects of the present disclosure, nor to delineate the scope of any or all aspects of the present disclosure. The sole purpose of this summary is to present some concepts of one or more aspects of the present disclosure in a summarized form as a prelude to more specific embodiments presented later.

[0009] In one aspect of the present disclosure, an apparatus includes at least one processor and a memory coupled to the at least one processor. The at least one processor is configured to obtain channel state information data associated with a second network node; train a shared UE encoder based on the channel state information data and based on a decoder to generate a sequential training data set; and send the sequential training data set to a third network node.

[0010] In an additional aspect of the present disclosure, an apparatus includes at least one processor and a memory coupled to the at least one processor. The at least one processor is configured to receive a sequential training data set from a second network node; train a base station decoder based on the sequential training data set to generate decoder model information; and send the decoder model information for the base station decoder to a third network node.

[0011] In an additional aspect of the present disclosure, an apparatus includes at least one processor and a memory coupled to the at least one processor. The at least one processor is configured to send channel state information data to a second network node; receive encoder model information from the second network node, the encoder model information being based on the channel state information; and send data to a third network node by encoding the data based on the encoder model information.

[0012] In an additional aspect of the present disclosure, an apparatus includes at least one processor and a memory coupled to the at least one processor. The at least one processor is configured to receive decoder model information for a shared base station decoder from a second network node; and receive encoded data from a third network node by decoding the encoded data based on the shared base station decoder.

[0013] The features and technical advantages of examples according to the present disclosure have been outlined quite broadly above so that the following specific embodiments may be better understood. Additional features and advantages will be described below. The disclosed concepts and specific examples may be easily used as a basis for modifying or designing other structures for achieving the same purpose of the present disclosure. Such equivalent constructions do not depart from the scope of the appended claims. When considered in conjunction with the accompanying drawings, the characteristics of the concepts disclosed herein (both their organization and methods of operation) and the associated advantages will be better understood from the following description. Each of the figures in the accompanying drawings is provided for the purpose of illustration and description, and not as a definition of limitations of the claims.

[0014] Although various aspects and specific implementations are described in this application by way of illustration of some examples, it will be understood by those skilled in the art that additional specific implementations and use cases may be generated in many different arrangements and scenarios. The innovations described herein may be implemented across many different platform types, devices, systems, shapes, sizes, and packaging arrangements. For example, aspects and / or uses may be generated via integrated chip specific implementations and other devices based on non-module components (e.g., end-user devices, vehicles, communication devices, computing devices, industrial equipment, retail / purchase devices, medical devices, devices that enable artificial intelligence (AI), etc.). Although some examples may or may not specifically point to use cases or applications, the applicability of a wide range of described innovations may occur. The scope of specific implementations may range from chip-level or modular components to non-modular, non-chip-level specific implementations, and further to the scope of aggregated, distributed, or original equipment manufacturer (OEM) devices or systems in conjunction with one or more aspects of the described innovations. In some practical environments, the devices in conjunction with the various aspects and features described may also necessarily include additional components and features for implementing and practicing the various aspects claimed and described. For example, the transmission and reception of wireless signals necessarily include multiple components for analog and digital purposes (e.g., hardware components including antennas, radio frequency (RF) chains, power amplifiers, modulators, buffers, processors, interleavers, adders / summers, etc.). The innovations described herein are intended to be practiced in a variety of devices, chip-level components, systems, distributed arrangements, end-user devices, etc., having different sizes, shapes, and configurations. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] A further understanding of the nature and advantages of the present disclosure may be achieved by reference to the following drawings. In the drawings, similar components or features may have the same reference label. In addition, various components of the same type may be distinguished by adding a dash and a second label to distinguish between similar components after the reference label. If only the first reference label is used in the specification, the description applies to any one of the similar components having the same first reference label, regardless of the second reference label.

[0016] Figure 1 is a block diagram illustrating details of an example wireless communication system in accordance with one or more aspects.

[0017] Figure 2 is a block diagram illustrating an example of a base station and a user equipment (UE) according to one or more aspects.

[0018] Figure 3A is a block diagram illustrating an example of encoder-decoder operation for channel state feedback according to one or more aspects.

[0019] Figure 3B is a block diagram illustrating an example of concurrent training in accordance with one or more aspects.

[0020] Figure 3C is a block diagram illustrating an example of sequential training in accordance with one or more aspects.

[0021] Figure 4 is a block diagram illustrating an example wireless communication system that supports UE-driven sequential training in accordance with one or more aspects.

[0022] Figure 5 is a timing diagram illustrating an example wireless communication system that supports UE-driven sequential training in accordance with one or more aspects.

[0023] Figure 6 is a timing diagram illustrating an example wireless communication system that supports UE-driven sequential training in accordance with one or more aspects.

[0024] Figure 7 is a timing diagram illustrating an example wireless communication system that supports UE-driven sequential training in accordance with one or more aspects.

[0025] Fig. 8A is a block diagram illustrating an example of sequential training with a reference decoder according to one or more aspects.

[0026] Figure 8B is a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects.

[0027] Fig. 9A is a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects.

[0028] Fig. 9B is a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects.

[0029] Fig. 9C is a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects.

[0030] Fig.9D is a block diagram illustrating an example of a common pre-processing layer for a decoder according to one or more aspects.

[0031] Fig.9E is a block diagram illustrating an example of a split architecture decoder in accordance with one or more aspects.

[0032] Fig.10 is a flow chart illustrating an example process for supporting UE-driven sequential training in accordance with one or more aspects.

[0033] Fig.11 is a flow chart illustrating an example process for supporting UE-driven sequential training in accordance with one or more aspects.

[0034] Fig.12 is a flow chart illustrating an example process for supporting UE-driven sequential training in accordance with one or more aspects.

[0035] Fig.13 is a flow chart illustrating an example process for supporting UE-driven sequential training in accordance with one or more aspects.

[0036] Fig.14 is a block diagram of an example UE supporting UE-driven sequential training according to one or more aspects.

[0037] Fig.15 is a block diagram of an example base station supporting UE-driven sequential training according to one or more aspects.

[0038] The same reference numbers and designations in different drawings indicate the same elements. DETAILED DESCRIPTION

[0039] The specific embodiments described below in conjunction with the accompanying drawings are intended as descriptions of various configurations and are not intended to limit the scope of the present disclosure. Instead, the specific embodiments include specific details for providing a thorough understanding of the subject matter of the present invention. It will be apparent to those skilled in the art that these specific details are not required in every case, and in some instances, well-known structures and components are shown in block diagram form for clarity of presentation.

[0040] The present disclosure as a whole relates to providing or participating in authorized shared access between two or more wireless devices in one or more wireless communication systems (also referred to as wireless communication networks). In various specific implementations, the techniques and apparatuses may be used in wireless communication networks, such as code division multiple access (CDMA) networks, time division multiple access (TDMA) networks, frequency division multiple access (FDMA) networks, orthogonal FDMA (OFDMA) networks, single carrier FDMA (SC-FDMA) networks, LTE networks, GSM networks, fifth generation (5G) or new radio (NR) networks (sometimes referred to as “5G NR” network, system or device) and other communication networks. As described in this document, the terms “network” and “system” can be used interchangeably.

[0041] A CDMA network may implement, for example, a radio technology such as Universal Terrestrial Radio Access (UTRA), cdma2000, etc. UTRA includes Wideband CDMA (W-CDMA) and Low Chip Rate (LCR). CDMA2000 covers IS-2000, IS-95, and IS-856 standards.

[0042] For example, a TDMA network may implement a radio technology such as the Global System for Mobile Communications (GSM). The Third Generation Partnership Project (3GPP) defines the standard for the GSM EDGE (Enhanced Data Rates for GSM Evolution) Radio Access Network (RAN), also known as GERAN. GERAN is the radio component of GSM / EDGE together with a network connecting base stations (e.g., Ater and Abis interfaces) and base station controllers (A interfaces, etc.). The radio access network represents the component of the GSM network through which phone calls and packet data are routed from the public switched telephone network (PSTN) and the Internet to subscriber phones (also known as user terminals or user equipment (UE)) and from subscriber phones to the PSTN and the Internet. The network of a mobile phone operator may include one or more GERANs, which may be coupled with the UTRAN in the case of a UMTS / GSM network. Additionally, the operator network may also include one or more LTE networks, or one or more other networks. Various different network types may use different radio access technologies (RATs) and RANs.

[0043] OFDMA networks may implement radio technologies such as Evolved UTRA (E-UTRA), Institute of Electrical and Electronics Engineers (IEEE) 802.11, IEEE 802.16, IEEE 802.20, flash-OFDM, etc. UTRA, E-UTRA, and GSM are part of the Universal Mobile Telecommunications System (UMTS). Specifically, Long Term Evolution (LTE) is a version of UMTS that uses E-UTRA. UTRA, E-UTRA, GSM, UMTS, and LTE are described in documents provided by an organization named "3rd Generation Partnership Project" (3GPP), and cdma2000 is described in documents from an organization named "3rd Generation Partnership Project 2" (3GPP2). These various radio technologies and standards are known or under development. For example, 3GPP is a collaboration between telecommunications association groups with the goal of defining globally applicable third generation (3G) mobile phone specifications. 3GPP LTE is a 3GPP project with the goal of improving the UMTS mobile phone standard. 3GPP can define specifications for next generation mobile networks, mobile systems, and mobile devices. The present disclosure may describe certain aspects with reference to LTE, 4G or 5G NR technology; however, the description is not intended to be limited to a particular technology or application, and one or more aspects described with reference to one technology may be understood to be applicable to another technology. Additionally, one or more aspects of the present disclosure may relate to shared access to wireless spectrum between networks using different radio access technologies or radio air interfaces.

[0044] 5G networks are expected to have diverse deployments, diverse spectrum, and diverse services and devices that can be implemented using a unified air interface based on OFDM. To achieve these goals, in addition to developing new radio technologies for 5G NR networks, further enhancements to LTE and LTE-A are also being considered. 5G NR will be able to scale to provide coverage (1) to massive Internet of Things (IoT), with ultra-high density (e.g., about 1M nodes / km 2 ), ultra-low complexity (e.g., about 10s of bits / second), ultra-low power consumption (e.g., about 10+ years of battery life), and deep coverage with the ability to reach challenging locations; (2) including mission-critical control with strong security to protect sensitive personal, financial, or classified information, ultra-high reliability (e.g., about 99.9999% reliability), ultra-low latency (e.g., about 1 millisecond (ms)), and users with a wide range of mobility or lack of mobility; and (3) providing services with enhanced mobile broadband (including extremely high capacity (e.g., about 10Tbps / km 2 ), coverage of very high data rates (e.g., multi-Gbps rates, 100+Mbps user experienced rates), and deep awareness with advanced discovery and optimization).

[0045] Devices, networks, and systems may be configured to communicate via one or more portions of the electromagnetic spectrum. The electromagnetic spectrum is typically subdivided into various categories, bands, channels, etc. based on frequency / wavelength. In 5G NR, two initial operating bands have been identified as frequency ranges designated FR1 (410MHz-7.125GHz) and FR2 (24.25GHz-52.6GHz). Frequencies between FR1 and FR2 are often referred to as mid-band frequencies. Although a portion of FR1 is greater than 6GHz, FR1 is often (interchangeably) referred to as the "sub-6GHz" band in various documents and articles. A similar naming issue sometimes arises for FR2, which is often (interchangeably) referred to as the "millimeter wave" (mmWave) band in documents and articles, although it is different from the extremely high frequency (EHF) band (30GHz-300GHz) identified as the "mmWave" band by the International Telecommunication Union (ITU).

[0046] In view of the above aspects, unless otherwise specifically stated, it should be understood that if used in this article, the term "below 6 GHz" and the like can broadly represent frequencies that can be less than 6 GHz, can be within FR1, or can include mid-band frequencies. In addition, unless otherwise specifically stated, it should be understood that if used in this article, the term "mmWave" and the like can broadly represent frequencies that can include mid-band frequencies, can be within FR2, or can be within the EHF band.

[0047] 5G NR devices, networks, and systems can be implemented using optimized OFDM-based waveform features. These features may include scalable parameter sets and transmit time intervals (TTIs); a common flexible framework that efficiently multiplexes services and features using dynamic, low-latency time division duplex (TDD) designs or frequency division duplex (FDD) designs; and advanced wireless technologies such as massive multiple-input multiple-output (MIMO), robust mmWave transmission, advanced channel decoding, and device-centric mobility. The scalability of parameter sets in 5G NR and the scaling of subcarrier spacing can effectively address the operation of various services across different spectrums and different deployments. For example, in various outdoor and macro coverage deployments of less than 3GHz FDD or TDD implementations, the subcarrier spacing may appear at 15kHz, such as over bandwidths of 1MHz, 5MHz, 10MHz, 20MHz, etc. For various other outdoor and small cell coverage deployments with TDD greater than 3GHz, the subcarrier spacing may appear at 30kHz over 80MHz / 100MHz bandwidth. For various other indoor broadband implementations, using TDD on the unlicensed portion of the 5GHz band, the subcarrier spacing may occur at 60kHz over a 160MHz bandwidth. Finally, for various deployments transmitting over mmWave components at 28GHz TDD, the subcarrier spacing may occur at 120kHz over a 500MHz bandwidth.

[0048] 5G NR's scalable parameter sets facilitate scalable TTIs for diverse latency and quality of service (QoS) requirements. For example, shorter TTIs can be used for low latency and high reliability, while longer TTIs can be used for higher spectral efficiency. Efficient multiplexing of long and short TTIs allows transmission to start on symbol boundaries. 5G NR also anticipates a self-contained integrated subframe design where uplink or downlink scheduling information, data, and acknowledgments are in the same subframe. The self-contained integrated subframe supports communications in unlicensed or contention-based shared spectrum, and adaptive uplink or downlink can be flexibly configured on a per-cell basis to dynamically switch between uplink and downlink to meet current business needs.

[0049] For clarity, certain aspects of the devices and techniques may be described below with reference to example 5G NR implementations or in a 5G-centric manner, and 5G terminology may be used as illustrative examples in portions of the description below; however, the description is not intended to be limited to 5G applications.

[0050] Furthermore, it should be understood that in operation, a wireless communication network adapted according to the concepts herein can operate using any combination of licensed spectrum or unlicensed spectrum depending on load and availability. Therefore, it will be apparent to one of ordinary skill in the art that the systems, devices, and methods described herein can be applied to other communication systems and applications beyond the specific examples provided.

[0051] Although various aspects and specific implementations are described in this application by way of illustration of some examples, it will be understood by those skilled in the art that additional specific implementations and use cases may be generated in many different arrangements and scenarios. The innovations described herein may be implemented across many different platform types, devices, systems, shapes, sizes, and packaging arrangements. For example, specific implementations or uses may be implemented via integrated chip specific implementations or other devices based on non-module components (e.g., end-user devices, vehicles, communication devices, computing devices, industrial equipment, retail devices or purchasing devices, medical devices, AI-enabled devices, etc.). Although some examples may or may not specifically point to use cases or applications, the applicability of a wide range of described innovations may occur. The scope of specific implementations may range from chip-level or modular components to non-modular, non-chip-level specific implementations, and further to aggregated, distributed, or original equipment manufacturer (OEM) devices or systems containing one or more of the described aspects. In some practical environments, the devices in combination with the various aspects and features described may also necessarily include additional components and features for implementing and practicing the various aspects claimed and described. It is intended that the innovations described herein may be implemented in a wide variety of implementations of different sizes, shapes, and configurations, including both large and small devices, chip-level components, multi-component systems (e.g., radio frequency (RF) chains, communication interfaces, processors), distributed arrangements, end-user devices, etc.

[0052] Figure 1 1 is a block diagram illustrating details of an example wireless communication system according to one or more aspects. The wireless communication system may include a wireless network 100. The wireless network 100 may, for example, include a 5G wireless network. As will be appreciated by those skilled in the art, Figure 1 Components appearing in are likely to have related corresponding components in other network arrangements, including, for example, cellular-style network arrangements and non-cellular-style network arrangements (e.g., device-to-device or peer-to-peer or ad hoc network arrangements, etc.).

[0053] Figure 1The illustrated wireless network 100 includes many base stations 105 and other network entities. A base station may be a station that communicates with a UE and may also be referred to as an evolved Node B (eNB), a next generation eNB (gNB), an access point, etc. Each base station 105 may provide communication coverage for a specific geographic area. In 3GPP, the term "cell" may refer to a specific geographic coverage area of ​​a base station or a base station subsystem serving the coverage area, depending on the context in which the term is used. In a specific implementation of the wireless network 100 herein, the base station 105 may be associated with the same operator or different operators (e.g., the wireless network 100 may include multiple operator wireless networks). Additionally, in a specific implementation of the wireless network 100 herein, the base station 105 may provide wireless communication using one or more frequencies (e.g., one or more frequency bands in a licensed spectrum, an unlicensed spectrum, or a combination thereof) in the same frequency as a neighboring cell. In some examples, a separate base station 105 or UE 115 may be operated by more than one network operating entity. In other examples, each base station 105 and UE 115 may be operated by a single network operating entity.

[0054] A base station may provide communication coverage for a macro cell or a small cell (e.g., a pico cell or a femto cell), or other type of cell. A macro cell generally covers a relatively large geographic area (e.g., several kilometers in radius) and may allow unrestricted access by UEs with a service subscription with a network provider. A small cell (such as a pico cell) will generally cover a relatively small geographic area and may allow unrestricted access by UEs with a service subscription with a network provider. A small cell (such as a femto cell) will generally also cover a relatively small geographic area (e.g., a home), and in addition to unrestricted access, may also provide restricted access by UEs associated with the femto cell (e.g., UEs in a closed subscriber group (CSG), UEs of users in a home, etc.). A base station for a macro cell may be referred to as a macro base station. A base station for a small cell may be referred to as a small cell base station, a pico base station, a femto base station, or a home base station. In Figure 1 In the example shown, base stations 105d and 105e are conventional macro base stations, while base stations 105a-105c are macro base stations implemented using one of 3-dimensional (3D), full-dimensional (FD), or massive MIMO. Base stations 105a-105c use their higher-dimensional MIMO capabilities to employ 3D beamforming in elevation and azimuth beamforming to increase coverage and capacity. Base station 105f is a small cell base station, which can be a home node or a portable access point. The base station can support one or more (e.g., two, three, four, etc.) cells.

[0055] Wireless network 100 may support synchronous or asynchronous operation. For synchronous operation, base stations may have similar frame timings, and transmissions from different base stations may be approximately aligned in time. For asynchronous operation, base stations may have different frame timings, and transmissions from different base stations may not be aligned in time. In some cases, the network may be enabled or configured to handle dynamic switching between synchronous or asynchronous operation.

[0056] UE 115 is dispersed throughout the wireless network 100, and each UE may be stationary or mobile. It should be understood that, although in the standards and specifications promulgated by 3GPP, mobile devices are generally referred to as UEs, such devices may be additionally or otherwise referred to as mobile stations (MS), subscriber stations, mobile units, subscriber units, wireless units, remote units, mobile devices, wireless devices, wireless communication devices, remote devices, mobile subscriber stations, access terminals (AT), mobile terminals, wireless terminals, remote terminals, mobile phones, terminals, user agents, mobile clients, clients, gaming devices, augmented reality devices, vehicle components, vehicle devices or vehicle modules or some other suitable terminology by those skilled in the art. In this document, a "mobile" device or UE does not necessarily have the ability to move and may be fixed. Some non-limiting examples of mobile devices (for example) may include one or more specific implementations of UE 115, including mobile phones, cellular phones, smart phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, laptops, personal computers (PCs), notebooks, netbooks, smart books, tablet computers, and personal digital assistants (PDAs). The mobile device may additionally be an IoT or "Internet of Everything" (IoE) device, such as a car or other transportation vehicle, a satellite radio, a global positioning system (GPS) device, a global navigation satellite system (GNSS) device, a logistics controller, a drone, a multi-rotor helicopter, a quad-rotor helicopter, a smart energy or security device, a solar panel or solar array, city lighting, tap water or other infrastructure; industrial automation and enterprise equipment; consumer and wearable devices, such as glasses, wearable cameras, smart watches, health or fitness trackers, mammal implantable devices, gesture tracking devices, medical devices, digital audio players (e.g., MP3 players), cameras, game consoles, etc.; and digital home or smart home devices, such as home audio, video and multimedia devices, appliances, sensors, vending machines, smart lighting, home security systems, smart meters, etc. In one aspect, the UE may be a device including a universal integrated circuit card (UICC). In another aspect, the UE may be a device that does not include a UICC. In some aspects, a UE that does not include a UICC may also be referred to as an IoE device. Figure 1The UEs 115a-115d of the specific implementation illustrated in the figure are examples of mobile smart phone type devices accessing the wireless network 100. The UE may also be a machine specially configured for connected communications (including machine type communications (MTC), enhanced MTC (eMTC), narrowband IoT (NB-IoT), etc.). Figure 1 The illustrated UEs 115e - 115k are examples of various machines that access the wireless network 100 and are configured for communication.

[0057] A mobile device, such as UE 115, may be able to communicate with any type of base station, whether macro, pico, femto, relay, etc. Figure 1 In the figure, the communication link (represented as a lightning ball) indicates the wireless transmission between the UE and the serving base station (which is a base station designated to serve the UE on the downlink or uplink) or the expected transmission between the base stations and the backhaul transmission between the base stations. The UE can operate as a base station or other network node in some scenarios. The backhaul communication between the base stations of the wireless network 100 can be carried out using wired or wireless communication links.

[0058] In operation, at the wireless network 100, base stations 105a-105c use 3D beamforming and coordinated spatial techniques (e.g., coordinated multipoint (CoMP) or multi-connectivity) to serve UEs 115a and 115b. Macro base station 105d performs backhaul communications with base stations 105a-105c and small cells (base station 105f). Macro base station 105d also transmits multicast services that are subscribed to and received by UEs 115c and 115d. Such multicast services may include mobile TV or streaming video, or may include other services for providing community information, such as weather emergencies or alerts, such as Amber alerts or gray alerts.

[0059] The wireless network 100 of the specific implementation supports mission critical communications with ultra-reliable and redundant links for mission critical devices such as UE 115e as a drone. The redundant communication links with UE 115e include links from macro base stations 105d and 105e and small cell base station 105f. Other machine type devices such as UE 115f (thermometer), UE 115g (smart meter) and UE 115h (wearable device) can communicate directly with base stations such as small cell base station 105f and macro base station 105e through the wireless network 100, or in a multi-hop configuration by communicating with another user device that relays its information to the network, such as UE 115f communicating temperature measurement information to smart meter UE 115g, which then reports it to the network through small cell base station 105f. The wireless network 100 may also provide additional network efficiency through dynamic, low-latency TDD communications or low-latency FDD communications (eg, in a vehicle-to-vehicle (V2V) mesh network between UEs 115i - 115k communicating with a macro base station 105e ).

[0060] Figure 2 1 is a block diagram illustrating an example of a base station 105 and a UE 115 according to one or more aspects. The base station 105 and the UE 115 may be Figure 1 For a restricted association scenario (as described above), base station 105 may be any one of the base stations in the UE and one of the UEs. Figure 1 The small cell base station 105f in the example of FIG. 105 and the UE 115 may be a UE 115c or 115d operating in the service area of ​​the base station 105f, which will be included in the list of accessible UEs of the small cell base station 105f in order to access the small cell base station 105f. The base station 105 may also be some other type of base station. Figure 2 As shown in FIG. 1 , the base station 105 may be equipped with antennas 234a through 234t and the UE 115 may be equipped with antennas 252a through 252r for facilitating wireless communication.

[0061] At the base station 105, the transmit processor 220 may receive data from the data source 212 and control information from the controller 240 (such as a processor). The control information may be for a physical broadcast channel (PBCH), a physical control format indicator channel (PCFICH), a physical hybrid ARQ (automatic repeat request) indicator channel (PHICH), a physical downlink control channel (PDCCH), an enhanced physical downlink control channel (EPDCCH), an MTC physical downlink control channel (MPDCCH), etc. The data may be for a physical downlink shared channel (PDSCH), etc. Additionally, the transmit processor 220 may process (e.g., encode and symbol map) the data and the control information, respectively, to obtain data symbols and control symbols. The transmit processor 220 may also generate, for example, reference symbols for a primary synchronization signal (PSS) and a secondary synchronization signal (SSS) and a cell-specific reference signal. The transmit (TX) MIMO processor 230 may perform spatial processing (e.g., pre-coding) on ​​data symbols, control symbols, or reference symbols, if applicable, and may provide output symbol streams to modulators (MODs) 232a to 232t. For example, the spatial processing performed on data symbols, control symbols, or reference symbols may include pre-coding. Each modulator 232 may process a corresponding output symbol stream (e.g., for OFDM, etc.) to obtain an output sample stream. Additionally or alternatively, each modulator 232 may process the output sample stream (e.g., analog-convert, amplify, filter, and up-convert) to obtain a downlink signal. The downlink signals from modulators 232a to 232t may be transmitted via antennas 234a to 234t, respectively.

[0062] At the UE 115, antennas 252a to 252r may receive downlink signals from the base station 105 and may provide received signals to demodulators (DEMODs) 254a to 254r, respectively. Each demodulator 254 may condition (e.g., filter, amplify, downconvert, and digitize) a corresponding received signal to obtain input samples. Each demodulator 254 may further process the input samples (e.g., for OFDM, etc.) to obtain received symbols. The MIMO detector 256 may obtain received symbols from the demodulators 254a to 254r, perform MIMO detection on the received symbols when necessary, and provide detected symbols. The receive processor 258 may process (e.g., demodulate, deinterleave, and decode) the detected symbols, provide decoded data for the UE 115 to the data sink 260, and provide decoded control information to the controller 280 (e.g., a processor).

[0063] On the uplink, at the UE 115, the transmit processor 264 may receive and process data from a data source 262 (e.g., for a physical uplink shared channel (PUSCH)) and control information from the controller 280 (e.g., for a physical uplink control channel (PUCCH)). Additionally, the transmit processor 264 may also generate reference symbols for reference signals. The symbols from the transmit processor 264 may be pre-decoded by the TX MIMO processor 266, if necessary, further processed by the modulators 254a to 254r (e.g., for SC-FDM, etc.), and transmitted to the base station 105. At the base station 105, the uplink signal from the UE 115 may be received by the antenna 234, processed by the demodulator 232, detected by the MIMO detector 236, if necessary, and further processed by the receive processor 238 to obtain decoded data and control information transmitted by the UE 115. The receive processor 238 may provide the decoded data to the data sink 239 and the decoded control information to the controller 240.

[0064] The controllers 240 and 280 may direct the operation at the base station 105 and the UE 115, respectively. The controller 240 or other processors and modules at the base station 105 or the controller 280 or other processors and modules at the UE 115 may perform or direct the execution of various processes for the techniques described herein, such as performing or directing Figures 4 to 15 , or other processes for the techniques described herein. Memories 242 and 282 may store data and program codes for base station 105 and UE 115, respectively. Scheduler 244 may schedule UEs for data transmission on a downlink or uplink.

[0065] In some cases, the UE 115 and the base station 105 may operate in a shared radio spectrum band, which may include licensed or unlicensed (e.g., contention-based) spectrum. In the unlicensed frequency portion of the shared radio spectrum band, the UE 115 or the base station 105 may traditionally perform a medium sensing procedure to compete for access to the spectrum. For example, the UE 115 or the base station 105 may perform a listen-before-talk or listen-before-transmit (LBT) procedure (such as a clear channel assessment (CCA)) before communicating to determine whether the shared channel is available. In some specific implementations, the CCA may include an energy detection procedure to determine whether there are any other active transmissions. For example, the device may infer that a change in the received signal strength indicator (RSSI) of the power meter indicates that the channel is occupied. Specifically, a signal power concentrated in a certain bandwidth and exceeding a predetermined noise floor may indicate another wireless transmitter. CCA may also include detection of a specific sequence indicating the use of the channel. For example, another device may send a specific preamble before sending a data sequence. In some cases, the LBT procedure may include the wireless node adjusting its own backoff window based on the amount of energy detected on the channel or acknowledgement / negative acknowledgement (ACK / NACK) feedback for its own transmitted packets (as an indication of a collision).

[0066] The deployment of a communication system (such as a 5G New Radio (NR) system) can be arranged with various components or constituent parts in a variety of ways. In a 5G NR system or network, a network node, a network entity, a mobility element of a network, a radio access network (RAN) node, a core network node, a network element or a network equipment (such as a base station (BS), or one or more units (or one or more components) performing base station functionality) can be implemented in an aggregated architecture or a decomposed architecture. For example, a BS (such as a Node B (NB), an evolved NB (eNB), an NR BS, a 5G NB, an access point (AP), a transmit receive point (TRP) or a cell, etc.) can be implemented as an aggregated base station (also known as an independent BS or a monolithic BS) or a decomposed base station.

[0067] A converged base station may be configured to utilize a radio protocol stack that is physically or logically integrated within a single RAN node. A decomposed base station may be configured to utilize a protocol stack that is physically or logically distributed between two or more units, such as one or more central or centralized units (CUs), one or more distributed units (DUs), or one or more radio units (RUs). In some aspects, a CU may be implemented within a RAN node, and one or more DUs may be co-located with the CU, or alternatively, may be geographically or virtually distributed in one or more other RAN nodes. A DU may be implemented to communicate with one or more RUs. Each of the CU, DU, and RU may also be implemented as a virtual unit, namely a virtual central unit (VCU), a virtual distributed unit (VDU), or a virtual radio unit (VRU).

[0068] Base station type operations or network designs may take into account the aggregated nature of base station functionality. For example, a disaggregated base station may be used in an integrated access backhaul (IAB) network, an open radio access network (O-RAN (network configurations such as those initiated by the O-RAN Alliance)), or a virtualized radio access network (vRAN, also known as a cloud radio access network (C-RAN)). Disaggregation may include distributing functionality across two or more units at various physical locations, as well as virtually distributing functionality of at least one unit, which may enable flexibility in network design. Individual units of a disaggregated base station or disaggregated RAN architecture may be configured for wired or wireless communication with at least one other unit.

[0069] See also Figure 3A , depicts a block diagram illustrating an example of encoder-decoder operation for channel state feedback according to one or more aspects. Figure 3A In the example of , the encoder of the UE receives Vin and generates Z. The UE sends Z to the base station (gNB), and the decoder of the base station generates Vout based on decoding Z. Vin may include or correspond to uncompressed or original channel state feedback (CSF). Z may include or correspond to compressed CSF, and Vout may include or correspond to reconstructed or decompressed CSF (e.g., CSI and / or precoding vectors).

[0070] In some implementations, the channel state information data and / or Vin includes or corresponds to a pre-decoder vector or a channel vector. Additionally or alternatively, the channel state information data and / or Vin includes an original channel.

[0071] To improve the performance, it has been proposed to perform cross-node (X-node) machine learning (ML) training of the encoder and decoder for CSF. In cross-node ML, the neural network (NN) is split into two parts, the encoder on the UE side and the decoder on the network side.

[0072] In "multi-vendor training", each vendor (e.g., UE vendor, gNB vendor) has its own server that participates in offline training. The UE vendor server communicates with the gNB vendor server using a server-to-server connection during training. However, doing so involves sharing the vendor's models. Since the model is associated with the architecture of the encoder / decoder and its model, providing vendor-specific models can lead to reverse engineering of proprietary information, such as hardware architecture (e.g., encoder / decoder architecture). Since this is generally disadvantageous, new schemes are needed to train multi-vendor encoders and / or decoders that can work with more devices.

[0073] Without multi-vendor training, each UE-gNB pair needs to maintain a different encoder-decoder pair. For example, in the first scenario (Scenario A) with multiple UE vendors from one gNB vendor, a common network decoder is trained to work with multiple UE encoders. The benefit here is that the base station does not need a separate decoder for each UE in the cell.

[0074] In the second scenario (Scenario B) for one UE vendor with multiple gNB vendors, a common encoder is trained to work with multiple gNB decoders. The benefit here is that when the UE moves from one cell to another, it does not require a separate encoder for each gNB.

[0075] In the third scenario (Scenario C) of multiple UE vendors with multiple gNB vendors, a common encoder-decoder pair is trained, and the UE encoder is trained to work with multiple gNB vendors and the gNB decoder is trained to work with multiple UE vendors. Such a framework enables increased compatibility and flexibility, and enables devices to have reduced requirements for network operation (e.g., fewer encoders / decoders).

[0076] See also Figure 3B, depicts a block diagram illustrating an example of concurrent training according to one or more aspects. Concurrent training includes joint training of an encoder and decoder at a single device, such as a UE or a base station server. For example, a UE vendor (e.g., Qualcomm) may train encoder and decoder models using its own dataset and share the trained decoder model with a gNB vendor (e.g., Ericsson or Nokia). As another example, a network vendor (e.g., Ericsson or Nokia) may train encoder and decoder models using its own dataset and share the trained encoder model with a UE vendor (e.g., Qualcomm). Such training may be performed “offline” rather than while connected to a network or involving interaction with a network. As described above, due to the symmetry that typically exists between the encoder and decoder, the decoder shared with the network vendor may reveal or imply specific implementation details of the UE modem.

[0077] To overcome these challenges, enhanced sequential training techniques (such as UE-driven sequential training) may be used to protect proprietary designs while still enabling true multi-vendor encoding and decoding. In various aspects described herein, a UE-side device or server may share training information with one or more network devices that enables the network to train its decoder. The training information (e.g., a sequential training data set or a UE-driven sequential training data set) may include or correspond to Z and Vin or Z and Vout. The sequential training data set may be generated based on a standard CSI input data set or an aggregated CSI input data set. The CSI information of the CSI data set may include or correspond to the same type of CSI information that would typically be used in joint / concurrent encoder and decoder training. Thus, multiple devices may be used to generate CSI or input information for generating training data sets, and multiple training data sets may be used to train a single universal decoder for use in a network with multiple UEs, or at least train a specific universal decoder for a network vendor or an entire class of network devices. The universal decoder may work / pair with encoders of multiple different UE vendors, different UE categories or types, etc. An example of such UE-driven sequential training is provided in Figure 3C Shown in.

[0078] See also Figure 3C , depicts a block diagram illustrating an example of sequential training according to one or more aspects. Figure 3C In the example, the UE side or UE vendor equipment trains its encoder and decoder concurrently (such as by referring to Figure 3B Concurrent joint training of the UE encoder and decoder) to generate a training data set of Z and Vin or Z and Vout. The decoder may include or correspond to the reference Fig. 8A and Figure 8BA reference decoder is described further. During training, an input dataset is used, such as the Vin dataset.

[0079] As reference Figure 3A As explained, the UE encoder encodes Vin to produce Z, and the UE decoder decodes Z to produce Vout. Vin and Vout may be provided to a loss function or another comparison device or logic component to determine the difference or gradient. The difference or gradient between the encoder input and the decoder output may represent an error. AI and / or ML methods (such as CNN or TF) may be used to adjust the model weights of the encoder and / or decoder to better match Vin and Vout, commonly referred to as training.

[0080] When training the encoder and decoder, the UE device can track and store data to generate a training data set. Alternatively, once the model is complete, the UE can feed a standard input set or raw input information to generate the corresponding Z and Vout. The UE can then generate a training data set based on two of Vin, Z, and Vout for training the encoder and / or decoder. The combination of Z and Vin and Z and Vout can be used to train the decoder on the network side.

[0081] The UE device then provides the training information to the network side, where the network side can train its decoder based on the received training information to train it or pair it with the UE side encoder. For example, as shown in the network size, Z from the training information is provided to the network decoder. The network decoder generates Vout,gnb based on Z from the training information. The network device can provide the generated Vout,gnb and the received Vout to the loss function for comparison. The difference between the two Vout is the error or gradient. The network can then use the gradient to adjust the decoder model weights (decoder model).

[0082] Alternatively, when Vin is provided, the network may provide Vin to a loss function to compare it to Vout,gnb to generate a difference or gradient. Similarly, the network may then use the gradient to adjust the decoder model weights (e.g., decoder model). When Vout is used, the network is allowed to make the same errors as the UE (have the same error). Using Vin may allow the original error or error to be corrected, but may introduce new errors. The network may select the most favorable training set based on results such as in real-world performance.

[0083] Figure 4An example of a wireless communication system 400 supporting UE-driven sequential training according to aspects of the present disclosure is illustrated. In some examples, the wireless communication system 400 can implement aspects of the wireless communication system 100. For example, the wireless communication system 400 may include a network such as one or more network entities (e.g., base station 105 and base station server 401) and one or more UE-side devices such as UE 115 and UE server 403. Figure 4 As illustrated in the example of , the network entity includes an entity corresponding to a base station such as base station 105. Alternatively, the network entity may include or correspond to a different network device (e.g., not a base station). UE-driven sequential training can reduce latency and increase throughput. For example, avoiding switching NN models reduces latency and increases throughput by avoiding inducing delays in switching NN models. Therefore, network and device performance can be improved.

[0084] The base station 105, the UE 115, the base station server 401, and the UE server 403 may be configured to communicate via one or more portions of the electromagnetic spectrum. The electromagnetic spectrum is typically subdivided into various categories, bands, channels, etc. based on frequency / wavelength. In 5GNR, two initial operating bands have been identified as frequency ranges designated FR1 (410 MHz–7.125 GHz) and FR2 (24.25 GHz–52.6 GHz). The frequencies between FR1 and FR2 are typically referred to as mid-band frequencies. Although a portion of FR1 is greater than 6 GHz, FR1 is typically (interchangeably) referred to as the “below 6 GHz” band in various documents and articles. Similar naming issues sometimes arise with respect to FR2, although unlike the extremely high frequency (EHF) band (30 GHz–300 GHz) identified as the “millimeter wave” (mmWave) band by the International Telecommunication Union (ITU), FR2 is typically (interchangeably) referred to as the “mmWave” band in various documents and articles.

[0085] In view of the above aspects, unless otherwise specifically stated, it should be understood that if used in this article, the term "below 6 GHz" and the like can broadly represent frequencies that can be less than 6 GHz, can be within FR1, or can include mid-band frequencies. In addition, unless otherwise specifically stated, it should be understood that if used in this article, the term "mmWave" and the like can broadly represent frequencies that can include mid-band frequencies, can be within FR2, or can be within the EHF band.

[0086] It should be noted that for some data channels, the SCS may be equal to 15kHz, 30kHz, 60kHz, or 120kHz. The base station 105 and the UE 115 may be configured to communicate via one or more component carriers (CCs), such as a representative first CC 481, a second CC 482, a third CC 483, and a fourth CC 484. Although four CCs are shown, this is for illustration only, and more or less than four CCs may be used. One or more CCs may be used to convey control channel transmissions, data channel transmissions, and / or sidelink channel transmissions.

[0087] Such transmissions may include a physical downlink control channel (PDCCH), a physical downlink shared channel (PDSCH), a physical uplink control channel (PUCCH), a physical uplink shared channel (PUSCH), a physical sidelink control channel (PSCCH), a physical sidelink shared channel (PSSCH), or a physical sidelink feedback channel (PSFCH). Such transmissions may be scheduled by aperiodic grants and / or periodic grants.

[0088] Each periodic grant may have a corresponding configuration, such as configuration parameters / settings. A periodic grant configuration may include a configured grant (CG) configuration and settings. Additionally or alternatively, one or more periodic grants (eg, their CGs) may have or be assigned to a CC ID, such as an expected CC ID.

[0089] Each CC may have a corresponding configuration, such as configuration parameters / settings. The configuration may include bandwidth, bandwidth part, HARQ process, TCI state, RS, control channel resources, data channel resources, or a combination thereof. Additionally or alternatively, one or more CCs may have or be assigned to a cell ID, a bandwidth part (BWP) ID, or both. The cell ID may include a unique cell ID for the CC, a virtual cell ID, or a specific cell ID for a specific CC in multiple CCs. Additionally or alternatively, one or more CCs may have or be assigned to a HARQ ID. Each CC may also have corresponding management functions, such as beam management, BWP switching functions, or both. In some implementations, two or more CCs are quasi-co-located so that the CCs have the same beam and / or the same symbols.

[0090] In some implementations, the control information may be communicated via the base station 105, the base station server 401, the UE 115, and the UE server 403. For example, the control information may be communicated using a MAC-CE transmission, an RRC transmission, a DCI (downlink control information) transmission, an UCI (uplink control information) transmission, an SCI (sidelink control information) transmission, another transmission, or a combination thereof.

[0091] The UE 115 may include various components (e.g., structures, hardware components) for performing one or more functions described herein. For example, these components may include a processor 402, a memory 404, a transmitter 410, a receiver 412, an encoder 413, a decoder 414, an input manager 415, a training module 416, and antennas 252a-r. The processor 402 may be configured to execute instructions stored at the memory 404 to perform the operations described herein. In some implementations, the processor 402 includes or corresponds to the controller / processor 280, and the memory 404 includes or corresponds to the memory 282. The memory 404 may also be configured to store input information 406, training information 408, encoder information 442, decoder information 444, setting data, or a combination thereof, as further described herein.

[0092] The input information 406 may, for example, include channel state feedback information. The channel state feedback information may, for example, include one or more measurements performed by the UE on one or more signals sent by a base station, such as one or more channel state information reference signals (CSI-RS) sent by the base station. In some embodiments, the channel state feedback information may include sensed raw channel information or singular vector information of one or more beamforming vectors. The input information 406 may be classified by a vendor of the base station associated with the input information or by a model of the base station associated with the input information. For example, input information 406 including sensing or measurement information of one or more signals from a base station identified with a first vendor may be classified as input information associated with a base station identified with the first vendor.

[0093] The training information 408 may include one or more training data sets for a decoder of the network device. Such training information may be received, for example, from the UE server 403. The base station server 401 or the base station 105 may use the training information to train the decoder (e.g., perform sequential training of the decoder) and / or adjust the operation of the decoder. In some embodiments, the training information may include instructions or codes for the decoder received from the UE server 403. The training information may include or correspond to a sequential training data set driven by the UE. The training information may include a tuple of encoder or decoder inputs and outputs, such as Z and Vin or Z and Vout.

[0094] The encoder information 442 may include one or more encoder parameters for an encoder of a UE device. Such parameters may be generated, for example, by the UE or received at the UE from the UE server 403. The UE 115 may use the encoder parameter information to adjust the operation of the encoder so as to encode information to be sent to the base station, such as encoding channel state feedback information for sending to the base station. In some embodiments, the encoder parameter information may include instructions or codes for the encoder received from the UE server 403.

[0095] Decoder information 444 may include one or more decoder parameters for a decoder of the network device. Such parameters may be generated, for example, by the base station or received at the base station from the base station server 401. The base station 105 may use the decoder parameter information to adjust the operation of the decoder for the decoder information received at the base station, such as for decoding the channel / channel state feedback information from the UE. In some embodiments, the decoder parameter information may include instructions or codes for the decoder of the base station 105 received from the base station server 401.

[0096] The setting data includes or corresponds to data associated with UE-driven sequential training operations. The setting data may include one or more types of UE-driven sequential training operation modes and / or thresholds or conditions for switching between UE-driven sequential training modes and / or configurations thereof. For example, the setting data may have data indicating different thresholds and / or conditions for different UE-driven sequential training modes (such as network-assisted mode, increased complexity mode, same complexity mode, etc. or a combination thereof).

[0097] The transmitter 410 is configured to send data to one or more other devices, and the receiver 412 is configured to receive data from one or more other devices. For example, the transmitter 410 can send data via a network (such as a wired network, a wireless network, or a combination thereof), and the receiver 412 can receive data via the network. For example, the UE 115 can be configured to send and / or receive data via the following: a direct device-to-device connection, a local area network (LAN), a wide area network (WAN), a modem-to-modem connection, the Internet, an intranet, an extranet, a cable transmission system, a cellular communication network, any combination of the above, or any other communication network now known or later developed that allows two or more electronic devices to communicate therein. In some specific implementations, the transmitter 410 and the receiver 412 can be replaced by a transceiver. Additionally or alternatively, the transmitter 410, the receiver 412, or both may include or correspond to the reference Figure 2 One or more components of UE 115 are described.

[0098] The encoder 413 and the decoder 414 may be configured to encode and decode data for transmission.The input information generation manager may generate input information, such as by sensing or measuring one or more signals from one or more base stations or processing the sensed or measured information.

[0099] The training module may be configured to train the encoder, the decoder, or both. For example, the encoder 413 and the reference decoder may be trained by the training module of the UE. For example, the concurrent training module may employ one or more ML algorithms to jointly train the encoder and the decoder using the input information 406. For example, if the input information 406 includes CSI information, the training module may adjust the parameters of the encoder and the decoder so that when the input information is input to the encoder 413, the encoder generates output information similar to the output information of the decoder. The training information 408 may be generated during or after the training of the encoder and the decoder by the UE 115.

[0100] UE server 403 may include one or more elements similar to UE 115. In some implementations, UE 115 and UE server 403 are different types of UEs. For example, either UE may be of higher quality or have different operating constraints. For illustration, one of the UEs may have a larger form factor or may be a current generation device and therefore have more advanced capabilities and / or reduced battery constraints, higher processing constraints, etc.

[0101] The base station 105 includes a processor 430, a memory 432, a transmitter 434, a receiver 436, an encoder 437, a decoder 438, an input manager 439, a training module 440, and antennas 234a-234t. The processor 430 may be configured to execute instructions stored at the memory 432 to perform the operations described herein. In some implementations, the processor 430 includes or corresponds to the controller / processor 240, and the memory 432 includes or corresponds to the memory 242. The memory 432 may be configured to store the input information 406, the training information 408, the encoder information 442, the decoder information 444, the setting data, or a combination thereof, similar to the UE 115 and as further described herein.

[0102] The transmitter 434 is configured to send data to one or more other devices, and the receiver 436 is configured to receive data from one or more other devices. For example, the transmitter 434 can send data via a network (such as a wired network, a wireless network, or a combination thereof), and the receiver 436 can receive data via the network. For example, the UE and / or the base station 105 can be configured to send and / or receive data via a direct device-to-device connection, a local area network (LAN), a wide area network (WAN), a modem-to-modem connection, the Internet, an intranet, an extranet, a cable transmission system, a cellular communication network, any combination of the above, or any other communication network now known or later developed that allows two or more electronic devices to communicate. In some specific implementations, the transmitter 434 and the receiver 436 can be replaced by a transceiver. Additionally or alternatively, the transmitter 434, the receiver 436, or both may include or correspond to reference Figure 2 One or more components of UE 115 are described.

[0103] The encoder 437 and the decoder 438 may include the same functionality as described with reference to the encoder 413 and the decoder 414, respectively. The input manager 439 may include similar functionality as described with reference to the input manager 415. The training module 440 may include similar functionality as described with reference to the training module 416. For example, the training module may be configured to train (sequentially train) the decoder 438 of the base station 105. For illustration, the training module may employ one or more ML algorithms to train the decoder 438 using the training information 408. The training module may adjust the parameters of the decoder 438 to enhance the operation of the decoder in continuously decoding the training information 408. For illustration, the decoder 438 may be fed with Z from the training information 408 and output a base station decoded output that is compared with the UE-side encoder input or decoder output of the training information 408. The difference with the comparison may be used to adjust the decoder (e.g., its weights) and generate decoder information 444.

[0104] Base station server 401 may include one or more elements similar to base station 105. In some implementations, base station 105 and base station server 401 are different types of base stations. For example, either base station device may be of higher quality or have different operating constraints. For illustration, one of the base station devices may have a larger form factor or may be a current generation device and therefore have more advanced capabilities and / or reduced power constraints, higher processing constraints, etc.

[0105] During operation of the wireless communication system 400, the network (e.g., the base station 105) may determine that the UE 115 has UE-driven sequential training capability. For example, the UE 115 may send a message 448 including a UE-driven sequential training indicator 490 (e.g., a UE-driven sequential training capability indicator). The indicator 490 may indicate UE-driven sequential training capability for one or more communication modes (such as downlink, uplink, etc.). In some embodiments, the network entity (e.g., the base station 105) transmits control information to indicate to the UE 115 that a UE-driven sequential training operation and / or a specific type of UE-driven sequential training operation is to be used. For example, in some embodiments, a configuration transmission 450 is sent to the UE 115. The configuration transmission 450 may include or indicate settings for using a UE-driven sequential training operation or adjusting or implementing a specific type of UE-driven sequential training operation. For example, the configuration transmission 450 may include decoder information 444 (e.g., Figure 4 ), input information 406, training information 408, encoder information 442, setting data, or any combination thereof.

[0106] During operation, the devices of the wireless communication system 400 perform UE-driven sequential training operations. For example, the network and the UE 115 may exchange transmissions via uplink and / or downlink communications and generate channel state information or feedback.

[0107] exist Figure 4 In the example of , the base station 105 optionally sends CSI-RS 452 to the UE 115 via a downlink channel. The CSI-RS includes a reference signal for the UE 115 to measure and generate or estimate channel conditions, channel state information. The estimated channel conditions may include uplink conditions, downlink conditions, sidelink conditions, etc. The UE 115 can report or feedback the CSI to the base station 105 as a CSF.

[0108] For example, UE 115 receives CSI-RS 452, and UE 115 measures CSI-RS 452 to generate CSI. UE 115 may generate input information 406 based on the CSI. Additionally or alternatively, UE 115 may generate input information 406 based on historical CSI information (such as from previous communications or communications from other devices). UE 115 may participate in concurrent training of its encoder and decoder by training module 416, such as reference Figure 3C During training of the encoder and decoder, the UE 115 may generate encoder information 442 and / or decoder information 444, such as encoder and decoder model weights.

[0109] UE 115 may also generate training information 408 based on or after training the encoder and decoder. In addition, UE 115 may generate training information 454 (eg, aggregated training information) based on training information 408 and second received training information, such as reference information. Figures 5 to 7 described.

[0110] UE 115 sends training information 454 to base station 105. For example, UE 115 may send training information 454 in an uplink message / uplink channel. Alternatively, UE 115 may send training information 454 to base station server 401 or UE server 403. When providing training information 454 to UE server 403, UE server 403 may aggregate training information 454 and relay the aggregated training information to base station 105 or base station server 401.

[0111] In other specific implementations, the UE 115 sends input information 406 (eg, CSI information) to the UE server 403, and the UE server 403 generates the training information 454. The UE server 403 may aggregate the additional input information and generate the training information 454, as described in reference to Figures 5 to 7 described.

[0112] The base station 105 receives the training information 454 and trains its decoder 438 based on the training information 454. For example, the base station 105 may train its decoder 438 as shown in FIG. Figure 3C During or after training of the decoder 438, the base station 105 generates decoder information 444 (eg, decoder model weights).

[0113] Alternatively, the training information may be provided to the base station server 401, and the base station server 401 may train the encoder based on the training information 454 to generate the decoder information 444. The base station server may provide the decoder information 444 to the base station for use, as shown in FIG. Figures 5 to 7 described.

[0114] After the shared or common encoder of the UE 115 is trained and the shared or common decoder of the base station 105 is trained, the UE 115 and the base station 105 can communicate one or more transmissions 456 using the respective encoders and decoders. For example, the UE 115 transmits the transmission of the one or more transmissions 456 by encoding the data (e.g., the second CSI data) based on the encoder information 442 (the encoder model trained based on the encoder information 442). The base station 105 receives the transmission of the one or more transmissions 456 by decoding the encoded data (e.g., the second CSI data) based on the decoder information 444 (the decoder model trained based on the decoder information 444).

[0115] As a result, the network (e.g., base station 105, base station server 401, UE 115, and UE server 403) may be able to more effectively and efficiently train multiple vendor encoders and decoders. Improved encoding and decoding operations may be achieved, such as improved compression and reconstruction of CSI information, resulting in lower overhead and errors. As a result, network performance and experience may be increased due to increased throughput and reduced failures.

[0116] Figure 5 is a timing diagram of a system 500 that supports UE-driven sequential training. The system 500 may include a UE 115, a UE-side server 120a, a base station-side server 120b, and a base station 105. The UE 115 and the UE-side server 120a may be associated with the same UE vendor (such as the same manufacturer or designer), the same UE category (such as Advanced, RedCap, or both). For example, the UE-side server 120a may generate training data to enable a specific vendor to train one or more decoders for implementation by a base station associated with the specific vendor. For illustration, the base station-side server 120b may train one or more decoders based on the training information received from the UE-side server 120a, and for implementation by a network device associated with the specific vendor and operable by the same vendor.

[0117] As another example, the UE-side server 120a may train one or more encoders for implementation by UEs associated with a particular vendor and operable by the same vendor. For illustration, the UE-side server 120a may train a shared or universal UE encoder based on multiple sets of input information, and may share or distribute encoder model information to multiple UEs to train the UEs.

[0118] The UE-side server 120a can control the training of one or more encoders by the UE-side server 120a by generating training information for use by the UE-side server 120a when training one or more encoders. Such control can achieve interoperability between the encoder implemented by the UE and the decoder implemented by the base station and enhanced encoding and decoding efficiency and reliability without requiring the supplier operating the UE-side server 120a to disclose details of the decoder or encoder trained by the UE-side server 120a and implemented on one or more base stations, such as not disclosing decoder output information or encoder parameters. In some embodiments, the system 500 may include multiple UE-side servers associated with multiple corresponding UE suppliers or multiple base station side servers associated with multiple base station side servers. For example, a single UE-side server 120a in a single training session can train an encoder using input information received from multiple UE-side servers associated with different corresponding UE suppliers, and can generate the same set of training information and send it to multiple UE-side servers associated with different corresponding UE suppliers.

[0119] At 510, the UE 115 may generate input information. The input information may, for example, include channel state feedback information. Generating the input information at 402 may include performing one or more measurements of one or more signals (such as CSI-RS) transmitted by one or more base stations (such as the base station 105). Such base stations may, for example, include base stations associated with the same vendor as the base station side server 120b. In some embodiments, the UE 115 may generate the input information using signals received from multiple base stations associated with multiple vendors.

[0120] At 515, UE 115 may send input information to UE-side server 120a. UE-side server 120a may be, for example, a UE-side server operated by the same vendor as the vendor associated with UE 115. The input information sent may, for example, include input information generated based on signals from multiple base stations associated with multiple different corresponding vendors, multiple base stations associated with a single vendor, or a single base station associated with a single vendor. In some embodiments, input information received from base stations associated with different vendors may be identified for separate processing by UE-side server 120a. UE-side server 120a may receive input information from multiple UEs associated with the same vendor as the UE-side server. Therefore, in some embodiments, the UE-side server may store multiple sets of input information associated with multiple corresponding base station vendors. Additionally or alternatively, UE-side server 120a may aggregate sets of input information from multiple UEs and for a specific base station vendor.

[0121] At 520, the UE-side server 120a may train the encoder and decoder. For example, the UE-side server 120a may train the encoder to generate training information for sending to one or more base stations. The encoder may include or correspond to a UE-side encoder, such as a shared or universal encoder. In addition, the UE-side server 120a may determine one or more encoder parameters for sending to one or more UEs. In some specific implementations, when training the encoder, the UE-side server 120a may also train the corresponding decoder (e.g., a reference decoder) to generate training information, such as performing single-sided concurrent training, which may be done offline. The UE-side server 120a may use the input information received at 515 to train the encoder and decoder. In some embodiments, the UE-side server 120a may use one or more ML algorithms to train the encoder and decoder. In some embodiments, the UE-side server 120a may use input information received from one or more UEs, input information received from one or more other UE-side servers, or other input information generated by the UE-side server 120a to train the encoder and decoder.

[0122] As described above, the UE-side server 120a can train multiple encoders, such as one encoder per vendor, one encoder per category, one encoder for each combination of vendor and category, etc. The UE-side server 120a can then send corresponding training data for each category to the corresponding base station or base station-side server.

[0123] In some implementations, when the encoder is trained by the UE-side server 120a, the base station-side device (e.g., gNB or server) can provide input information to the encoder, such as reference decoder information or model weights, such as reference Figure 6 As further described, the trained encoder can encode and output encoded information, such as encoded input information for a decoder.

[0124] At 525, the UE-side server 120a may generate training information. For example, the UE-side server 120a may generate training information using the trained encoder and decoder. The training information may be generated during training or after the training is completed. In some implementations, the input information may be provided (re-provided) after the training is completed to generate a sequential training data set for training the network-side decoder. The UE-side server 120a may generate a training data set based on Z and Vin or Z and Vout.

[0125] In some embodiments, the UE-side server 120a may train an encoder and a decoder at 520 and generate training information for each of a plurality of UE-side servers associated with different UE vendors at 525. For example, the UE-side server 120a may train an encoder specific to a particular UE vendor using one or more actual network decoders or reference decoders. In some embodiments, the UE-side server 120a may train a single encoder and decoder and generate a single training information for a plurality of UE vendors using input information from a plurality of UE vendors in a single training session. In some embodiments, the UE-side server 120a may train a plurality of encoders. The UE-side server 120a may generate a plurality of sets of training information, such as a plurality of sets of training information for a plurality of base station-side servers. For example, a set of training information for a particular UE-side server may be generated by passing the input information received from each respective UE or UE-side server through an encoder and a decoder trained using input information from each respective UE vendor and / or category, or by passing the input information set received from each respective UE-side device through a single encoder and a decoder pair to generate a set of training information.

[0126] At 530, the UE-side server 120a may send the training information generated by the trained encoder and decoder to the base station-side server 120b. For example, the UE-side server 120a may send the training information via a wired or wireless connection. For example, the training information may include or correspond to a training data set including Z and Vin or including Z and Vout. These components may be arranged as a tuple, a corresponding pair of information, or multiple items stored as a single variable or input for training the decoder.

[0127] At 535, the base station side server 120b may train a decoder using training information received from multiple items stored in a single variable. In some embodiments, the base station side server 120b may train multiple decoders using multiple sets of training information received from multiple corresponding UE side servers or devices and associated with different vendors, different categories of devices, or both. Training a decoder (or multiple decoders) may include applying one or more ML algorithms to generate decoder parameters. The decoder parameters may, for example, include computer code, instructions, weights, vectors, or other decoder parameters for use by the base station 105. In some embodiments, in order to train the decoder, the base station side server 120b may pass the input information (e.g., Z) of the tuple of training information through the decoder, and may adjust the parameters of the decoder (e.g., model weights) until the output (Vout, gnb) of the decoder approaches or matches the output (e.g., Vin or Vout) of the corresponding tuple. When the decoder is trained, the base station side server 120b may send the decoder parameters to the base station 105 at 540.

[0128] At 545, the base station 105 may use the received decoder parameters to decode the information. Thus, the base station side server 120b may provide training information for training a decoder to be used by one or more base stations. Such training may be remote, in that the remote base station side server and the UE side server may collaborate to train the decoder for use by one or more base stations, the training may be offline, in that the UE side server and the base station side server may train the encoder and decoder when the encoder and decoder are not being used to encode or decode information for transmission, and the training may be sequential, in that the UE side server may train the encoder to generate training information for use by the base station side server in training the decoder.

[0129] At 550, the UE-side server 120a may send encoder parameters to the UE-side server 120a. Although the sending of encoder parameters (encoder parameter information or encoder model information) is illustrated as being sent after the model information and before the sending of decoder parameters, the sending of encoder parameters may occur at any time after the generation or adjustment of the encoder parameters (such as at any time after 520 or 525).

[0130] At 555, the UE 115 may use the received encoder parameters to encode the information. Thus, the UE-side server 120a may provide encoder parameters for use in training an encoder to be used by one or more UEs. Such training may be remote, in that a remote UE-side server and a remote base station-side server may collaborate to train an encoder for use by one or more UEs, the training may be offline, in that the UE-side server and the base station-side server may train the encoder when the encoder is not being used to encode information for transmission, and the training may be sequential, in that the base station-side server may train the encoder to generate training information for use by the UE-side server in training the encoder.

[0131] Alternatively, the UE may train its own encoder based on input and similar to how the UE server sends information.Although described as a UE server, the UE server may include or correspond to an advanced UE or a master UE.

[0132] Figure 6 6 is a timing diagram of a system 600 that supports UE-driven sequential training. The system 600 may include one or more UEs (eg, UE 115), a UE-side server 120a, and one or more network devices such as a base station-side server 120b or a base station 105. Figure 6 In the example illustrated in , two UEs, namely a first UE 115a and a second UE 115b, are illustrated, with a single UE-side server 120a and a single base station 105. The UEs 115a and 115b and the UE-side server 120a may be associated with the same UE vendor (such as the same manufacturer or designer), the same UE category (such as Advanced, RedCap, or both). For example, the UE-side server 120a may aggregate input information from multiple UEs, generate aggregated training information, and provide the aggregated training information for a network (such as a specific network vendor) to train a network-side decoder. As another example, the UE-side server 120a may train one or more encoders for implementation by UEs associated with a specific vendor and operable by the same vendor.

[0133] At 605, the first UE 115a may generate first input information. The first input information may, for example, include first channel state feedback information. Generating the first input information at 605 may include performing one or more measurements of one or more signals (such as CSI-RS) transmitted by one or more base stations (such as base station 105). In some embodiments, the first UE 115a may generate the first input information using signals received from multiple base stations associated with multiple vendors. Additionally or alternatively, the first UE 115a may generate the first input information using signals from communications with a UE-side server 120a. In other specific implementations, the first UE 115a retrieves historical input information.

[0134] At 610, the first UE 115a may send first input information to the UE-side server 120a. The UE-side server 120a may, for example, be a UE-side server operated by the same vendor as the vendor associated with the first UE 115a. In some embodiments, first input information received from base stations associated with different vendors may be identified for separate processing by the UE-side server 120a. The UE-side server 120a may receive different sets of input information from multiple UEs associated with the same vendor as the UE-side server. Therefore, in some embodiments, the UE-side server 120a may store multiple sets of input information associated with multiple corresponding base station vendors. Additionally or alternatively, the UE-side server 120a may aggregate sets of input information from multiple UEs and for a specific base station vendor.

[0135] At 615, the second UE 115b may send second input information to the UE-side server 120a. The UE-side server 120a may, for example, be a UE-side server operated by the same vendor as the vendor associated with the second UE 115b. The second input information sent may, for example, include second input information generated by the second UE 115b and based on signals from multiple base stations associated with multiple different corresponding vendors, multiple base stations associated with a single vendor, or a single base station associated with a single vendor.

[0136] In some implementations, the first UE 115a and the second UE 115b are the same type of UE, such as both are advanced UEs, both are Red Cap UEs, etc. In other implementations, the first UE 115a and the second UE 115b are different types of UEs. When the UEs are of different types, the UE-side server 120a may not aggregate their input data or may aggregate their input data in different ways. Similarly, when the UEs are from different vendors and / or generate their corresponding input information from one or more base stations of different types operated by different vendors, the UE-side server 120a may not aggregate their input data or may aggregate their input data in different ways. Although in Figure 6 Not shown, but the second UE 115b may generate second input information.

[0137] At 620, the UE-side server 120a may generate (e.g., aggregate) aggregate input information. For example, the UE-side server 120 may combine the first input information and the second input information to generate the aggregate input information. As another example, the UE-side server 120 may modify the first input information based on the second input information to generate the aggregate input information, or may modify the second input information based on the first input information to generate the aggregate input information. In some specific implementations, the UE-side server 120 may generate (e.g., aggregate) the aggregate input information based on the first input information and the second input information based on determining that the first UE 115a and the second UE 115b are similar devices (such as having the same encoder architecture, having similar encoding complexity, being similar or the same type of device, etc.).

[0138] At 625, the UE-side server 120 may train the encoder and decoder. For example, the UE-side server 120a may train the encoder to generate (aggregate) training information for sending to one or more base stations. The encoder may include or correspond to a UE-side encoder, such as a shared or universal encoder. In addition, the UE-side server 120a may determine one or more encoder parameters for sending to one or more UEs. In some specific implementations, when training the encoder, the UE-side server 120a may also train the corresponding decoder (e.g., a reference decoder) to generate training information, such as performing single-sided concurrent training, which may be done offline. The UE-side server 120a may use the aggregated input information generated at 620 to train the encoder and decoder. In some embodiments, the UE-side server 120a may use one or more ML algorithms to train the encoder and decoder. In some embodiments, the UE-side server 120a may use aggregated input information received from one or more UEs, input information received from one or more other UE-side servers, or other input information generated by the UE-side server 120a to train the encoder and decoder.

[0139] As reference Figure 5 As described, the UE-side server 120a can train multiple encoders and decoders, such as encoder-decoder pairs for different categories of UEs and / or for different vendors. In addition, as described in reference 645, in some specific implementations, the UE-side server 120a can train a specific encoder-decoder pair based on network-side information (such as decoder information).

[0140] At 630, the UE-side server 120a generates training information, such as reference Figure 5 525. Since the training information may be based on aggregate input information, the training information may be referred to as aggregate or aggregated training information. The aggregated training information may be used to train a network-side decoder that is paired with and capable of operating with multiple UEs (ie, encoders thereof).

[0141] At 635, the UE-side server 120a may send the training information to the base station 105. Additionally or alternatively, the UE-side server 120a may send the training information to the base station-side server and / or one or more other base stations. The sending of the training information may be wired or wireless.

[0142] After receiving the training information, the base station 105 may Figure 5 The decoder is trained as described in 535. Additionally or alternatively, the base station 105 may send training information, decoder information, or both to one or more other network devices such as a base station and / or a base station-side server.

[0143] At 640, the UE-side server 120a may send encoder parameters to one or more UE-side devices (such as one or more UEs and / or one or more other UE-side servers). Figure 6 As illustrated in the example of , the UE-side server 120a sends encoder parameters to the second UE 115b and optionally to the first UE 115a. In some specific implementations, the UE (such as the second UE 115b) can send encoder parameters to one or more other UE-side devices (such as one or more UEs and / or one or more other UE-side servers).

[0144] Although the sending of encoder parameters (encoder parameter information or encoder model information) is illustrated as being sent after the training information, the sending of encoder parameters may occur at any time after the training of the encoder and the generation or adjustment of the encoder parameters (such as at any time after 625 or 630).

[0145] After receiving the encoder parameters, UE 115 may encode the information using the received encoder parameters, as shown in FIG. Figure 5 555. Therefore, the UE-side server 120a can provide encoder parameters for training an encoder to be used by one or more UEs. This training can be remote, because the remote UE-side server and the remote base station-side server can collaborate to train the encoder for use by one or more UEs, the training can be offline, because the UE-side server and the base station-side server can train the encoder when the encoder is not used to encode information for transmission, and the training can be sequential, because the base station-side server can train the encoder to generate training information for use by the UE-side server when training the encoder.

[0146] In addition, after receiving training information based on the aggregated input information from multiple UE side devices, the base station side device can train one or more network decoders based on the training information. One or more base station devices can then use the decoder (trained on the training information) to decode the encoded information, as shown in reference Figure 5 545. Therefore, the UE-side server 120a may provide training information for training a decoder to be used by one or more network devices that may be used with multiple UE devices and UE vendors.

[0147] In some implementations, the base station 105 (or the base station-side server) may generate input information for the UE-side server 120a (or the UE) to use for training the encoder and decoder. For example, the base station 105 may generate information about an actual decoder that it uses or will use in communications with the UE or a reference decoder used in training the UE-side encoder. The reference decoder may be different from the actual decoder (e.g., more complex than the actual decoder) and further refer to Fig. 8A and Figure 8B As an example, the base station 105 may generate decoder information, reference decoder information, initial weight information, final weight information, or a combination thereof. At 645, the base station 105 may send the generated input information to the UE-side server 120a for training the encoder and decoder at 625. Thus, the base station-side device may enable or help the UE-side to train its encoder and generate training information for training the decoder. This additional information may enable increased accuracy and reduced bottlenecks.

[0148] Figure 7 is a timing diagram of a system 700 that supports UE-driven sequential training. The system 700 may include one or more UE-side devices (e.g., UE 115 or a UE-side server) and one or more network devices such as a base station-side server 120b or a base station 105. Figure 7In the illustrated example, two UE-side servers, namely the first UE-side server 120a and the second UE-side server 120c, are illustrated, and two base stations, namely the first base station 105 and the second base station 105, are illustrated. The UE-side servers 120a and 120c may be associated with different UE vendors, such as different manufacturers or designers, different UE categories, or both. For example, the first UE-side server 120a may aggregate first input information of multiple first UEs from a specific vendor and / or type, and the second UE-side server 120c may aggregate second input information of multiple second UEs from different vendors and / or types. Each UE-side server may generate corresponding aggregated training information based on its own aggregated input information, and may provide corresponding aggregated training information for the network (such as a specific network vendor or type) to train the network-side decoder. As an example relative to the previous figure, the UE-side server may optionally train one or more encoders for implementation by a UE associated with the UE-side server.

[0149] At 710, the first UE-side server 120a may send first training information to the first base station 105a; at 715, the second UE-side server 120c may send second training information to the first base station 105a. Figure 7 , but the first UE side server 120a and the second UE side server 120c may generate their respective training information, as shown in the above reference Figure 5 525 and Figure 6 of 630.

[0150] At 720, the first base station 105a may generate (e.g., aggregate) aggregated training information. For example, the first base station 105a may combine the first training information and the second training information to generate the aggregated training information. As another example, the first base station 105a may modify the first training information based on the second training information to generate the aggregated training information, or may modify the second training information based on the first training information to generate the aggregated training information. In some implementations, the first base station 105a may generate (e.g., aggregate) aggregated training information based on the first training information and the second training information based on determining that the first UE-side server 120a and the second UE-side server 120c correspond to or are associated with similar devices, such as their training data is generated based on input information from devices having the same encoder architecture, having similar encoding complexity, being similar or the same type of devices, etc. In some implementations, the first base station 105a may aggregate the training information even if the first UE-side server 120a and the second UE-side server 120c are from different vendors or their training information relates to different vendors / UEs.

[0151] In other specific implementations, the aggregated data may also be based on network data. For example, one or more network devices (e.g., the second base station 105b) may provide training information to the first base station 105a, such as received UE-driven or UE-side training information. As another example, one or more network devices may provide input information data (such as decoder information, reference decoder information, or decoder weight information) to one or more of the UE-side devices for generating the underlying training information provided to the first base station 105a. Examples of such input information refer to Figure 6 describe.

[0152] At 725, the first base station 105a may use the training information to train the decoder. In some embodiments, the first base station 105a may use multiple sets of training information associated with different vendors, different categories of equipment, or both received from multiple corresponding UE-side servers to train multiple decoders. Training the decoder may include applying one or more ML algorithms to generate decoder parameters. The decoder parameters may, for example, include computer code, instructions, weights, vectors, or other decoder parameters for use by the first base station 105a, the second base station 105b, and / or one or more other base stations. In some embodiments, in order to train the decoder, the first base station 105a may pass the input information of the tuple of training information through the decoder, and may adjust the parameters of the decoder until the output of the decoder approaches or matches the output of the corresponding tuple. When training the decoder, at 730, the first base station 105a may send the decoder parameters to the second base station 105b.

[0153] Despite Figure 7 In the example of the first base station 105a, the first base station 105a trains its own decoder and generates decoder model data for sending / sharing with other network-side devices, but in other specific implementations, the first base station 105a may share the aggregated training data with one or more other network devices (such as the second base station 105b and / or one or more base station-side servers).

[0154] At 735, the second base station 105b may use the received decoder parameters to train or update the decoder. The second base station 105b may train the decoder as at 535 and with reference to Figure 5 The second base station 105b may update the trained decoder based on retraining the trained decoder based on the decoder parameters or by generating or training a second decoder to be used with an additional device class.

[0155] After 735, the base station 105 may use the received decoder parameters to decode the information. For example, the base station 105 may use a decoder trained based on the received decoder parameters to decode the encoded information received from one or more different UEs. Thus, the UE-side server may provide training information for training a decoder to be used by one or more base stations.

[0156] Additionally or alternatively, in other specific implementations, Figures 4 to 7 One or more operations of may be added, removed, or replaced. For example, in some implementations, Figure 6 and Figure 7 The example steps can be used together with and / or with Figure 4 or Figure 5 For illustration, Figure 6 The generation of aggregate input information can be done with Figure 4 , Figure 5 and Figure 7 As another example, Figure 7 The generation of aggregated training information can be done with Figures 4 to 6 Use with examples.

[0157] Despite Figures 4 to 7 Specific types of devices (BS or BS server and UE or UE server) are described in the examples, but in other specific implementations, other types of devices and combinations of devices may be used. In particular, all types of additional devices may be used to further increase the applicability and versatility of the encoder and decoder. As another example, UE-side devices (UE and UE server) and base station-side devices (BS and BS server) may be interchangeable with each other. For illustration, Figure 7 One or more of the UE-side servers in the example of may be UEs in other examples. As another example, Figure 7 One or more of the base stations in the example may be a base station server.

[0158] The complexity of the encoder and decoder affects the overall encoding and decoding capabilities of the network. For example, the amount or degree of compression and the degree of lossless reconstruction depend on the complexity and accuracy of the encoder and decoder. This includes the ML model used to train the encoder and decoder. For example, some ML models have higher complexity than other models (e.g., TF is higher than CNN). In addition, adding layers to NNs generally increases complexity. Therefore, complexity (e.g., complexity scores) can be assigned to encoders and decoders based on the type (ML architecture) and number of layers. In addition, complexity can be assigned to encoders and / or decoders based on one or more other factors and / or without the type or number of layers.

[0159] In operation, there may be different scenarios for the complexity of the encoders and decoders used in the network. For example, the network decoder may have the same complexity as the best UE encoder. In such implementations, operation of the network may result in performance degradation compared to the same encoder-decoder pair in 1-to-1 concurrent training. This is because sequential training may impart additional variance that may result in additional errors in encoding / compression and decoding / decompression without increasing complexity.

[0160] As another example, the network decoder may be more complex than the optimal UE encoder. In such implementations, the operation of the network may be improved compared to the above examples. To illustrate, in such examples, the decoder will no longer be a limiting factor and it may be able to compensate for sequential training and training with a training data set rather than concurrently with actual input and output.

[0161] However, as indicated above, it is not possible for vendors to share specific implementation details of their respective encoders and decoders. In various aspects described herein, it is proposed to share basic encoder and / or decoder information to assist in the training process. For example, basic complexity scores or class indications can be used to avoid unnecessary bottlenecks. As another example, ML architecture and layer information can be used for better encoder-decoder pairing. Additionally, initial weights can be provided to assist in the training process without providing a complete encoder or decoder model.

[0162] As described above, a UE-side device such as a UE or a server can jointly train a UE encoder and a decoder to generate training data for the network. The decoder can be referred to as a training decoder or a reference decoder because it is a pseudo-representative of the decoder of the network decoder to be used in actual operation. In different specific implementations, the UE-side device may have different levels of knowledge of the decoder used by the network. This knowledge can range from complete lack of knowledge to nearly complete knowledge. In some specific implementations, we can assign a level to the amount of knowledge that the UE has about the decoder used by the network.

[0163] For example, the first level of understanding (e.g., level 0) may correspond to a decoder that is unaware of the network. In some such implementations, the UE may not know the architecture, number of layers, complexity level, etc. of the decoder's neural network. In such implementations, the UE may select a reference decoder based on the architecture or type of its own encoder to provide a better match or symmetry. Alternatively, the UE may select a more complex reference decoder. For example, the UE may determine to increase the complexity score from its own complexity score, such as by increasing the layers in NN complexity (e.g., CNN to Transformer NN) or layer complexity (e.g., adding layers).

[0164] As another example, a second level of understanding (e.g., level 1) may include a basic understanding of the network's decoders. In such an implementation, the UE selects a decoder based on such understanding. Information about the network decoder may be obtained from a network device or a public database. As an illustrative example, the decoder information corresponds to a reference decoder NN architecture, such as the type and number of layers of the NN (e.g., a CNN with 2 layers). Similar to the first level, the UE may select a more complex reference decoder than indicated. For example, the UE may determine to increase the complexity score from the complexity score of the reference decoder, such as by increasing the layers in NN complexity (e.g., CNN to transformer NN) or layer complexity (e.g., adding layers). When training the UE encoder and reference decoder concurrently, the UE adjusts the weights of both the UE encoder and the reference decoder to optimize both simultaneously.

[0165] Training the UE encoder with a reference decoder can impose some structure on the latent space (representation of z). If Ref-Dec has less complexity compared to the actual gNB decoder, the reference decoder may actually cause a performance bottleneck during operation due to the performance limitations imposed during training. When the reference decoder has more complexity compared to the actual network decoder, the actual network decoder will be the bottleneck. Since this is a practical limitation, performance can increase when using a more complex reference decoder compared to the actual network decoder.

[0166] As another example, the third level of understanding (e.g., level 2) may include working understanding of the decoder of the network. For example, the UE may have a second level of understanding (e.g., level 1) and have model information, such as initial weights or final weights. The weight information may include or correspond to an actual network decoder or a reference network decoder provided by the network. In such implementations, the UE may select a reference decoder based on the first level information, and may then use the received decoder weight information to train the encoder and decoder. In some such implementations, the UE may fix the decoder weights and only adjust the encoder weights. In some other such implementations, the UE may adjust the received decoder weights and encoder weights.

[0167] See also Fig. 8A , depicts a block diagram illustrating an example of sequential training with a reference decoder according to one or more aspects. Fig. 8A In an example, two UE vendors or categories use the same reference decoder to train corresponding different encoder implementations, such as described above. As shown, a first UE trains a first encoder type (encoder type 1) using a reference decoder having a first type, and a second UE trains a second encoder type (encoder type 2) using a reference decoder having a first type (decoder type 1).

[0168] After the encoder is trained and the training data is generated, the decoder can be trained by the network, as shown in reference Figures 3C to 7 For example, the training information generated from the training of the encoder is used to train Figure 8B The shared base station / generic base station decoder (gNB decoder or "real" decoder).

[0169] See also Figure 8B , depicts a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects. Figure 8B In the example, Fig. 8A Both UEs are connected to Figure 8B Specifically, two UE encoders (Type 1 and Type 2) operate with an actual decoder, which may be of the same type (e.g., Type 1) or a different type (Type 2) as the reference decoder. Alternatively, the actual decoder may be of the same type, but may have a higher complexity, such as by having one or more additional layers. Each of the encoders is paired with a decoder, in that corresponding training data from the first encoder and the second encoder is used to train the actual decoder.

[0170] If the reference decoder has the same complexity as the actual decoder, then the reference decoder may be the actual bottleneck. For example, if both the reference decoder and the actual decoder are type 1 decoders with similar number of layers, their decoders may be classified as the same complexity type.

[0171] If the reference decoder has a higher complexity than the actual decoder, then the actual decoder will be the bottleneck. For example, if the reference decoder is a type 1 decoder and the actual decoder is a type 2 decoder, then the actual decoder can be said to have a lower complexity score. Additionally or alternatively, the actual decoder may have fewer layers than the reference decoder. The complexity can be based on the decoder's architecture type, training model, layers, etc. Examples of different decoder architectures are given in 9A to 9E Example in.

[0172] See also Fig. 9A , depicts a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects. Fig. 9AThe decoder shown in corresponds to a decoder with preprocessing (e.g., general preprocessing). The preprocessing is configured to condition the input data for input to the ML, such as for input into the NN. The NN may be a simple NN such as having one to two layers. Alternatively, the NN may be a complex NN having 3 or more layers. As examples of preprocessing, the processing may include changing dimensions, connecting data to the input data for identification, conditioning, reordering data, aligning subspaces of different inputs, such as rotating data by multiplication, addition, or subtraction. These actions may be configured to address UE-specific, encoder type-specific, or vendor-specific aspects that result in differences in the input (z).

[0173] See also Fig. 9B , depicts a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects. Fig. 9B The decoder shown in corresponds to a decoder with 1-hot encoding preprocessing and UE-specific / UE-specific preprocessing layers. Fig. 9B In an example of , an encoder output, such as Z1 or Z2, is received at a one-hot encoder. The one-hot encoder appends, adds, or connects one or more additional bits to the input (Z1 and Z2). In a simple single-layer example, vector Z has an original dimension of 64 bits, and two bits are added to the front or left end of the vector as bits 0 and 1. For example, bits of 1,0 are added to Z1 and bits of 0,1 are added to Z2, and these added bits enable identification and routing of the modified (one-hot encoded) vector corresponding to each UE (e.g., each UE type or each UE vendor) to the corresponding processing layer. Each corresponding processing layer is configured to restore or reduce the dimension of the UE-specific input back to the original dimension (z-tilde).

[0174] In addition, each corresponding processing layer performs its own actions to condition the data for ML processing. For example, if the subspaces of Z1 and Z2 are not aligned, one or more of Z1 or Z2 can be rotated to align the subspaces. For illustration, the first subspace of Z1 can be rotated to the second subspace of Z2, or the two subspaces of Z1 and Z2 can be rotated by different amounts to a standard or reference subspace. Although the output of each UE-specific layer is shown as being combined before being received by the decoder, the Z tilde for encoder 1 and encoder 2 is not combined, but this acts more like a switch to direct the conditioned / pre-processed encoder output to the decoder for decoding.

[0175] See also Fig. 9C , depicts a block diagram illustrating an example of a decoder with preprocessing according to one or more aspects. Fig. 9C The decoder shown in corresponds to a decoder with 1-hot encoding preprocessing and common preprocessing layers. Fig. 9B Compared with the decoder in Fig. 9C The decoder in uses a series of common layers or a common set of layers to process / condition each encoder output for input to the decoder. An exemplary arrangement of common pre-processing layers is illustrated in Figure C, and the example arrangement is referenced Fig.9D Shown and further described in detail.

[0176] Fig. 9C The common preprocessing layer of can be configured to reduce or restore the input vector back to its original dimension. For example, one-hot encoding can add 1 or more bits to the length of a vector or the dimension of a matrix. The common preprocessing layer can be configured to restore (e.g., reduce) the original dimension of Z. As illustrated in the previous example, one-hot encoding adds two bits and changes the dimension from 64 to 66. The common preprocessing layer is configured to output a Z tilde with a length of 64 bits. In some specific implementations, the common preprocessing layer can also be configured to convert each input into a reference orientation or subspace.

[0177] See also Fig.9D , depicts a block diagram illustrating an example of a common pre-processing layer for a decoder according to one or more aspects. Fig.9D In more detail, Fig. 9C An example of a common preprocessing layer. As shown in the figure, Fig.9D Three common preprocessing layers are included. The three layers include a Gaussian layer sandwiched between a first linear layer and a second linear layer. As an illustrative example, the first linear layer may be configured to receive the output of a one-hot encoder, and the Gaussian layer (e.g., a Gaussian error linear unit (GELU) activation layer) may be configured to receive the output of the first linear layer. The second linear layer may be configured to receive the output of the Gaussian layer and provide (output) input to a shared common decoder.

[0178] In certain implementations, the first layer may adjust or align the input vector that has been one-hot encoded. In some implementations, the first linear layer and / or the Gaussian layer are configured to further increase the dimension of the input vector, such as from 66 to 128. Additionally or alternatively, the Gaussian activation layer and / or the second linear layer are configured to reduce the dimension of the modified vector, such as from 128 to 64.

[0179] See also Fig.9E , depicts a block diagram illustrating an example of a split architecture decoder in accordance with one or more aspects. Fig.9EThe decoder shown in corresponds to a decoder without dedicated separate preprocessing. Instead, the decoder has a split architecture and includes one or more per-UE processing layers and one or more common processing layers (e.g., generic or non-UE-specific layers). The decoder may also store per-UE parameters in a memory (e.g., a decoder memory) for use with the one or more per-UE processing layers. In a specific implementation, the decoder has corresponding UE-specific parameters for each UE-specific layer. Processing the input (Z) at one or more per-UE processing layers with UE-specific parameters enables the decoder to resolve deviations in input or format from encoder to encoder (UE to UE).

[0180] Compared with existing decoders, Fig.9E The split architecture decoder maintains the same computational complexity, increasing the memory requirement for storing the per-UE parameters for each UE layer. For example, the split architecture decoder may have a name number for the operation. The decoder or base station may need to store multiple weights, such as the product of the number of UEs multiplied by the number of parameters.

[0181] Fig.10 is a flow chart illustrating example blocks executed by a wireless communication device (eg, a UE or a base station) configured according to an aspect of the present disclosure. These example blocks will also be described with respect to Fig.14 The description will be made using the UE 115 illustrated in FIG. Fig.14 1 is a block diagram illustrating a UE 115 configured according to one aspect of the present disclosure. The UE 115 includes Figure 2 and / or Figure 4 1401a-1401r and antennas 252a-252r. Figure 2 As illustrated in relation to UE 115, modulators / demodulators 254a-254r, a MIMO detector 256, a receive processor 258, a transmit processor 264, and a TX MIMO processor 266 are included. Fig.14 As illustrated in the example of , the memory 282 stores CSI logic 1402, training logic 1403, encoding logic 1404, encoder information 1405, training information 1406, input information 1407, and setting data 1408. The data (1402 to 1408) stored in the memory 282 may include or correspond to data stored in Figure 4The data (406, 408, 442 and / or 444) in the memory 404.

[0182] At block 1000, a wireless communication device (such as a UE-side device) obtains channel state information data associated with a second network node. For example, UE 115 or UE-side server 120a may generate and / or receive CSI information, as described in reference to Figures 4 to 7 described.

[0183] At block 1001, the UE trains a shared UE encoder based on channel state information data and based on a decoder to generate a sequential training data set. For example, the UE 115 or the UE-side server 120a may train a shared UE encoder based on channel state information and based on a decoder selected by or for the UE to generate a sequential training data set, as described in reference to Figures 4 to 7 described.

[0184] At block 1002, the UE sends a sequential training data set to a third network node. For example, the UE 115 or the UE-side server 120a may send a sequential training data set to a base station-side node, as shown in FIG. Figures 4 to 7 Alternatively, the UE server may send a sequential training data set to another UE server for aggregation before sending to the base station side node.

[0185] In other specific implementations, a wireless communication device (e.g., a UE or a base station) may perform additional blocks (or the wireless communication device may be configured to further perform additional operations). For example, a wireless communication device (e.g., a UE 115) may perform one or more of the above operations. As another example, a wireless communication device (e.g., a UE 115) may perform one or more aspects given below.

[0186] In a first aspect, a sequential training dataset comprises a UE-driven sequential training dataset configured to implement sequential training of a decoder based on concurrent training of a UE encoder and a decoder.

[0187] In a second aspect, alone or in combination with the first aspect, a sequential training data set comprises (z, Vin), wherein Vin comprises an input vector for an encoder, and wherein z comprises an output from the encoder based on Vin.

[0188] In a third aspect, alone or in combination with one or more of the above aspects, a sequential training data set comprises (z, Vout), wherein z comprises a decoder input of a vector, and wherein Vout comprises a decoder output of a vector.

[0189] In a fourth aspect, alone or in combination with one or more of the above aspects, the channel state information data (and / or Vin?) comprises or corresponds to a pre-decoder vector or a channel vector.

[0190] In a fifth aspect, alone or in combination with one or more of the above aspects, the channel state information data (and / or Vin?) comprises an original channel or a singular vector (eg, a perturbation vector).

[0191] In a sixth aspect, either alone or in combination with one or more of the above aspects, Vin comprises uncompressed / original channel state feedback (CSF), wherein z comprises compressed CSF, and wherein Vout comprises reconstructed / decompressed CSF, and the first network node further: encodes Vin using a shared UE encoder to generate Z; decodes Z using a reference decoder to generate Vout; compares Vout with Vin; and adjusts the encoder, the decoder, or both based on the comparison.

[0192] In a seventh aspect, alone or in combination with one or more of the above aspects, in order to adjust the encoder, the decoder or both based on the loss function comparison, it includes: calculating the gradient based on the comparison; and adjusting the encoder model weights, the decoder model weights or both based on the gradient.

[0193] In an eighth aspect, alone or in combination with one or more of the above aspects, in order to obtain channel state information data associated with a second network node, it includes: receiving channel state information data from the second network node; or generating channel state information data based on communication with the second network node.

[0194] In a ninth aspect, alone or in combination with one or more of the above aspects, the first network node comprises a UE server.

[0195] In a tenth aspect, alone or in combination with one or more of the above aspects, the second network node comprises a UE, and wherein the third network node comprises a base station server.

[0196] In an eleventh aspect, alone or in combination with one or more of the above aspects, the network node further: receives second channel state information data associated with a fourth network node (e.g., a second UE); and generates aggregated channel state information based on the channel state information and the second channel state information, wherein in order to train a shared UE encoder based on the channel state information, it includes: training the shared UE encoder based on the aggregated channel state information.

[0197] In a twelfth aspect, alone or in combination with one or more of the above aspects, the first network node further: receives second channel state information data associated with a fourth network node (e.g., a second UE); trains a shared UE encoder based on the second channel state information to update a sequential training data set and generates an updated sequential training data set; and sends the updated sequential training data set.

[0198] In a thirteenth aspect, alone or in combination with one or more of the above aspects, in order to train a shared UE encoder, it includes: performing training of a shared UE encoder and decoder (e.g., concurrent training) to generate a sequential training data set and encoder model weights; and sending the encoder model weights to a second network node.

[0199] In a fourteenth aspect, alone or in combination with one or more of the above aspects, the decoder corresponds to a reference decoder of the base station, wherein the network node further determines the type of the reference decoder based on the type of the UE encoder.

[0200] In a fifteenth aspect, alone or in combination with one or more of the above aspects, the network node further determines the type of reference decoder based on the type or architecture of the shared UE encoder.

[0201] In a sixteenth aspect, either alone or in combination with one or more of the above aspects, the decoder corresponds to a reference decoder of a base station, and the network node further: obtains reference decoder information of the base station; and determines a reference decoder based on the reference decoder information of the base station.

[0202] In a seventeenth aspect, either alone or in combination with one or more of the above aspects, the decoder corresponds to a reference decoder for a base station, wherein the network node further: obtains reference decoder information and decoder model weights of the decoder of the base station, wherein the decoder model weights include initial weights or final weights; and determines a reference decoder based on the reference decoder information of the base station, wherein the shared UE encoder is further trained based on the decoder model weights.

[0203] In some such aspects, the initial weights enable the UE server to also fine tune and update the weights of the decoder. The final weights are not updated by the UE server.

[0204] In an eighteenth aspect, alone or in combination with one or more of the above aspects, the network node further sends the sequential training data set to a fourth network node.

[0205] In a nineteenth aspect, either alone or in combination with one or more of the above aspects, the first network node includes a UE, and in order for the network node to further send data to a fourth network node (e.g., a BS, such as a third node or another node) by encoding the data based on encoder model information, the encoder model information being generated based on training a shared UE encoder.

[0206] In a twentieth aspect, either alone or in combination with one or more of the above aspects, the encoder is a CSI encoder, and wherein encoding the data comprises encoding the CSI data to generate compressed CSI data.

[0207] In a twenty-first aspect, either alone or in combination with one or more of the above aspects, the encoder is a pre-coded information encoder, and wherein encoding the data comprises encoding the pre-coded information to generate compressed pre-coded information.

[0208] In a twenty-second aspect, alone or in combination with one or more of the above aspects, the network node further sends the second data to a fifth network node (eg, a second BS) by encoding the second data based on the encoder model information.

[0209] In a twenty-third aspect, alone or in combination with one or more of the above aspects, the network node further: receives a CSI-RS from a fourth network node; measures the CSI-RS to generate measurement data; generates (e.g., estimates) CSI based on the measurement data; and encodes the CSI based on the encoder.

[0210] In a twenty-fourth aspect, alone or in combination with one or more of the above aspects, the network node further: receives a CSI-RS from a fourth network node; measures the CSI-RS to generate measurement data; generates (e.g., estimates) CSI based on the measurement data; generates precoding information based on the CSI; and encodes the precoding information based on the encoder.

[0211] Therefore, the wireless communication device can perform UE-driven sequential training operations of the wireless communication device.By performing UE-driven sequential training, encoding and decoding operations can be improved, which increases throughput and reduces latency, errors, and overhead.

[0212] Fig.11 1 is a flow chart illustrating example blocks executed by a wireless communication device (eg, a UE or a network entity such as a base station) configured according to an aspect of the present disclosure. These example blocks will be described with respect to Fig.15 The description is made with reference to the illustrated base station 105. Fig.15 is a block diagram illustrating a base station 105 configured according to one aspect of the present disclosure. The base station 105 includes Figure 2 and / or Figure 4 105. For example, the base station 105 includes a controller / processor 240 that operates to execute logic or computer instructions stored in a memory 242 and to control components of the base station 105 that provide the features and functions of the base station 105. The base station 105 transmits and receives signals via wireless radios 1501a-1501t and antennas 234a-234t under the control of the controller / processor 240. The wireless radios 1501a-1501t include, for example, Figure 2 Various components and hardware for base station 105 are illustrated, including modulators / demodulators 232a-232r, MIMO detector 236, receive processor 238, transmit processor 220, and TX MIMO processor 230. Fig.15 As illustrated in the example of , the memory 242 stores logic 1502, training logic 1503, decoding logic 1504, decoder information 1505, training information 1506, input information 1507, and setting data 1508. The data (1502 to 1508) stored in the memory 242 may include or correspond to data stored in Figure 4 The data (406, 408, 442 and / or 444) in the memory 432.

[0213] At block 1100, a wireless communication device such as a network device (eg, base station 105) receives a sequential training data set from a second network node. For example, base station 105 or base station-side server 120b receives training information, such as reference Figures 4 to 7 described.

[0214] At block 1101, the wireless communication device trains a base station decoder based on a sequential training data set to generate decoder model information. For example, the base station 105 or the base station side server 120b trains a shared or common base station decoder based on a sequential training data set to generate decoder model parameters, as shown in FIG. Figures 4 to 7 described.

[0215] At block 1102, the wireless communication device sends decoder model information for a base station decoder to a third network node. For example, the base station 105 or the base station side server 120b sends the decoder model information for the base station decoder to another BS side device, as shown in FIG. Figures 4 to 7 The other BS-side device may be a base station (eg, another base station) or a base station-side server (eg, another base station-side server).

[0216] In other specific implementations, a wireless communication device (e.g., such as a UE or a base station) may perform additional blocks (or the wireless communication device may be configured to further perform additional operations). For example, the wireless communication device may perform one or more of the above operations. As another example, the wireless communication device may perform as described in reference Figure 4 One or more aspects described to Figure 8 and as presented below.

[0217] In a first aspect, the decoder model information enables other network nodes to train a shared base station decoder for decoding encoded data from multiple different types of UEs.

[0218] In a second aspect, alone or in combination with the first aspect, the network node further: receives a second sequential training data set from a fourth network node (e.g., a second UE / UE server); and generates an aggregated sequential training data set based on the sequential training data set and the second sequential training data set, and wherein in order to train the base station decoder based on the sequential training data set, it includes: training the base station decoder based on the aggregated sequential training data set to generate decoder model information.

[0219] In a third aspect, alone or in combination with one or more of the above aspects, the network node further sends reference decoder information to the UE or UE server, wherein the reference decoder information enables the UE or UE server to use the reference decoder information as a reference decoder when training the UE encoder.

[0220] In a fourth aspect, alone or in combination with one or more of the above aspects, the reference decoder information includes decoder architecture information, decoder layer information, decoder category information, or a combination thereof.

[0221] In a fifth aspect, alone or in combination with one or more of the above aspects, the decoder class information indicates decoder architecture complexity information, decoder layer complexity information or a combined complexity level.

[0222] In a sixth aspect, either alone or in combination with one or more of the above aspects, the network node further: sends reference decoder information and decoder model weights to the UE or UE server, wherein the reference decoder information and decoder model weights enable the UE or UE server to use the reference decoder information and decoder model weights as a reference decoder when training the UE encoder, wherein the decoder model weights include initial weights or final weights.

[0223] In a seventh aspect, either alone or in combination with one or more of the above aspects, the base station decoder is more complex than any UE encoder.

[0224] In an eighth aspect, either alone or in combination with one or more of the above aspects, the architecture of the base station decoder is the same type of architecture as the most complex UE encoder, and wherein the base station decoder has more layers than the most complex UE encoder.

[0225] In a ninth aspect, alone or in combination with one or more of the above aspects, the base station decoder has substantially similar complexity as the most complex UE encoder.

[0226] In a tenth aspect, either alone or in combination with one or more of the above aspects, the base station decoder is of the same type of architecture as the most complex UE encoder, and wherein the base station decoder has the same number of layers as the most complex UE encoder.

[0227] In an eleventh aspect, either alone or in combination with one or more of the above aspects, the network node further: receives a first compressed CSI from a first UE; receives a second compressed CSI from a second UE; decodes the first compressed CSI to generate a first decoded CSI; and decodes the second compressed CSI to generate a second decoded CSI.

[0228] In a twelfth aspect, alone or in combination with one or more of the above aspects, a base station decoder comprises: a pre-processor; and a shared common decoder (eg, a universal decoder for multiple types of UEs and / or multiple vendor UEs).

[0229] In the thirteenth aspect, alone or in combination with one or more of the above aspects, the network node further: receives first UE compressed CSI from the first UE; receives second UE compressed CSI from the second UE; preprocesses the first UE compressed CSI by a preprocessor to generate first preprocessed CSI; preprocesses the second UE compressed CSI by the preprocessor to generate second preprocessed CSI; decodes the first preprocessed CSI by a shared common decoder to generate first decoded CSI; and decodes the second preprocessed CSI by a shared common decoder to generate second decoded CSI.

[0230] In a fourteenth aspect, either alone or in combination with one or more of the above aspects, the preprocessor is configured to perform one-hot encoding.

[0231] In a fifteenth aspect, alone or in combination with one or more of the above aspects, the pre-processor includes multiple UE-specific layers.

[0232] In the sixteenth aspect, alone or in combination with one or more of the above aspects, the network node further: receives a first compressed CSI from a first UE; receives a second compressed CSI from a second UE; one-hot encodes the first compressed CSI by a one-hot encoder to generate a first one-hot encoded compressed CSI; one-hot encodes the second compressed CSI by a one-hot encoder to generate a second one-hot encoded compressed CSI; pre-processes the first one-hot encoded compressed CSI by a first layer among multiple UE-specific layers to generate a first pre-processed CSI; pre-processes the first one-hot encoded compressed CSI by a second layer among multiple UE-specific layers to generate a second pre-processed CSI; decodes the first pre-processed CSI by a shared common decoder to generate a first decoded CSI; and decodes the second pre-processed CSI by a shared common decoder to generate a second decoded CSI.

[0233] In a seventeenth aspect, either alone or in combination with one or more of the above aspects, a pre-processor comprises a set of common processing layers.

[0234] In an eighteenth aspect, either alone or in combination with one or more of the above aspects, the set of common processing layers comprises: a first linear layer configured to receive the output of a one-hot encoder; a Gaussian layer (e.g., GELU); a Gaussian error linear unit (GELU) configured to receive the output of the first linear layer; and a second linear layer configured to receive the output of the Gaussian layer and provide input to a shared common decoder.

[0235] In the nineteenth aspect, alone or in combination with one or more of the above aspects, the network node further: receives a first UE compressed CSI from a first UE; receives a second UE compressed CSI from a second UE; one-hot encodes the first UE compressed CSI by a one-hot encoder to generate a first one-hot encoded compressed CSI; one-hot encodes the second UE compressed CSI by a one-hot encoder to generate a second one-hot encoded compressed CSI; preprocesses the first one-hot encoded compressed CSI by a common processing layer set to generate a first preprocessed CSI; preprocesses the first one-hot encoded compressed CSI by a common processing layer set to generate a second preprocessed CSI; decodes the first preprocessed CSI by a shared common decoder to generate a first decoded CSI; and decodes the second preprocessed CSI by a shared common decoder to generate a second decoded CSI.

[0236] In the twentieth aspect, either alone or in combination with one or more of the above aspects, the decoder comprises a general decoder comprising: one or more UE-specific layers configured to pre-process compressed CSI based on stored per-UE parameters; one or more common layers configured to decode the pre-processed CSI; and stored per-UE parameters.

[0237] In aspect twenty-first, alone or in combination with one or more of the above aspects, the network node further: receives a first compressed CSI from a first UE; receives a second compressed CSI from a second UE; processes the first compressed CSI by one or more UE-specific layers based on a first stored UE parameter in the stored per-UE parameters to generate a first adjusted CSI; processes the second compressed CSI by one or more UE-specific layers based on a second stored UE parameter in the stored per-UE parameters to generate a second adjusted CSI; decodes the first adjusted CSI by one or more common layers to generate a first decoded CSI; and decodes the second adjusted CSI by one or more common layers to generate a second decoded CSI.

[0238] Therefore, the wireless communication device can perform UE-driven sequential training operations of the wireless communication device.By performing UE-driven sequential training, encoding and decoding operations can be improved, which increases throughput and reduces latency, errors, and overhead.

[0239] Fig.12 is a flow chart illustrating example blocks executed by a wireless communication device (eg, a UE or a network entity such as a base station) configured according to an aspect of the present disclosure. These example blocks will also be described with respect to Fig.14 The description is based on the UE 115 illustrated and described above.

[0240] At block 1200, a wireless communication device (such as a UE side device) sends channel state information data to a second network node. For example, UE 115 sends CSI data to another node, as shown in FIG. Figures 4 to 7 The other nodes may include or correspond to a UE side device or a BS side device. When it is a UE side device, the device may include or correspond to a UE or a UE server. When it is a BS side device, the device may include or correspond to a BS or a BS server.

[0241] At block 1201, a wireless communication device receives encoder model information from a second network node, the encoder model information being based on channel state information. For example, a UE or a UE receives encoder model parameters from another UE-side device, such as reference Figures 4 to 7 described.

[0242] At block 1202, the wireless communication device sends data to a third network node by encoding the data based on encoder model information. For example, the UE sends the encoded CSI information to the base station, as shown in FIG. Figures 4 to 7 described.

[0243] In other specific implementations, a wireless communication device (e.g., such as a UE or a base station) may perform additional blocks (or the wireless communication device may be configured to further perform additional operations). For example, the wireless communication device may perform one or more of the above operations. As another example, the wireless communication device may perform as described in reference Figures 4 to 7 One or more aspects as described and presented below.

[0244] In a first aspect, the encoder is a CSI encoder, and wherein encoding the data comprises encoding CSI data to generate compressed CSI data.

[0245] In a second aspect, alone or in combination with the first aspect, the encoder is a pre-coded information encoder, and wherein encoding the data comprises encoding the pre-coded information data to generate compressed pre-coded information data.

[0246] In a third aspect, either alone or in combination with one or more of the above aspects, the network node further: sends the second encoded data to a fourth network node (e.g., a second BS) by encoding the second data based on the encoder model information, the fourth network node being different from the third network node (e.g., a different type of BS or vendor).

[0247] In a fourth aspect, alone or in combination with one or more of the above aspects, the first network node comprises a UE.

[0248] In a fifth aspect, alone or in combination with one or more of the above aspects, the second network node comprises a UE server, and wherein the third network node comprises a base station.

[0249] In a sixth aspect, alone or in combination with one or more of the above aspects, the channel state information data includes or corresponds to historical CSI data from a first network node in communication with one or more other nodes.

[0250] In a seventh aspect, alone or in combination with one or more of the above aspects, the channel state information data comprises or corresponds to CSI data from a first network node in communication with a second network node.

[0251] In an eighth aspect, either alone or in combination with one or more of the above aspects, the first network node is connected to the second network node via a non-cellular communication link (e.g., WiFi, Bluetooth, etc.), and wherein the channel state information data or the encoder model information is sent via the non-cellular communication link.

[0252] In a ninth aspect, alone or in combination with one or more of the above aspects, the network node further: receives a CSI-RS from a third network node; measures the CSI-RS to generate measurement data; generates (e.g., estimates) CSI based on the measurement data; and encodes the CSI based on the encoder.

[0253] In a tenth aspect, alone or in combination with one or more of the above aspects, the network node further: receives a CSI-RS from a third network node; measures the CSI-RS to generate measurement data; generates (e.g., estimates) CSI based on the measurement data; generates precoding information based on the CSI; and encodes the precoding information based on the encoder.

[0254] Therefore, the wireless communication device can perform UE-driven sequential training operations of the wireless communication device.By performing UE-driven sequential training, encoding and decoding operations can be improved, which increases throughput and reduces latency, errors, and overhead.

[0255] Fig.13 is a flow chart illustrating example blocks executed by a wireless communication device (eg, a UE or a network entity such as a base station) configured according to an aspect of the present disclosure. These example blocks will also be described with respect to Fig.15 The description is made with reference to the illustrated base station 105.

[0256] At block 1300, a wireless communication device such as a network device (e.g., base station 105) receives decoder model information of a shared base station decoder from a second network node. For example, base station 105 receives decoder information, such as decoder model parameters, from another base station or from a base station server, such as reference Figures 4 to 7 described.

[0257] At block 1301, the wireless communication device receives encoded data from a third network node by decoding the encoded data based on a shared base station decoder. For example, the base station 105 receives the encoded data and decodes the data using a decoder that is adjusted or trained using decoder information, such as reference Figures 4 to 7 The decoder information may be generated based on the UE driven sequence information.

[0258] In other specific implementations, a wireless communication device (e.g., such as a UE or a base station) may perform additional blocks (or the wireless communication device may be configured to further perform additional operations). For example, the wireless communication device may perform one or more of the above operations. As another example, the wireless communication device may perform as described in reference Figures 4 to 7 One or more aspects as described and presented below.

[0259] In the first aspect, the network node further trains the shared base station decoder based on the decoder model information.

[0260] In a second aspect, alone or in combination with the first aspect, the network node further receives the second coded data from a fourth network node (eg, a second UE) by decoding the second coded data based on a shared base station decoder, wherein the fourth network node is a different type of node from the third network node.

[0261] In a third aspect, alone or in combination with one or more of the above aspects, the first network node comprises a base station.

[0262] In a fourth aspect, alone or in combination with one or more of the above aspects, the second network node comprises a base station server, and wherein the third network node comprises a UE.

[0263] Therefore, the wireless communication device can perform UE-driven sequential training operations of the wireless communication device.By performing UE-driven sequential training, encoding and decoding operations can be improved, which increases throughput and reduces latency, errors, and overhead.

[0264] As described herein, a node (which may be referred to as a node, a network node, a network entity, or a wireless node) may include, may be, or may be included in (e.g., as a component of) a base station (e.g., any base station described herein), a UE (e.g., any UE described herein), a network controller, an apparatus, a device, a computing system, an integrated access and backhaul (IAB) node, a distributed unit (DU), a central unit (CU), a remote / radio unit (RU) (which may also be referred to as a remote radio unit (RRU)), and / or another processing entity configured to perform any of the techniques described herein. For example, the network node may be a UE. For another example, the network node may be a base station or a network entity. For another example, a first network node may be configured to communicate with a second network node or a third network node. In one aspect of this example, the first network node may be a UE, the second network node may be a base station, and the third network node may be a UE. In another aspect of this example, the first network node may be a UE, the second network node may be a base station, and the third network node may be a base station. In yet other aspects of this example, the first network node, the second network node, and the third network node may be different relative to these examples. Similarly, references to UE, base station, device, equipment, computing system, etc. may include disclosure of UE, base station, device, equipment, computing system, etc. as a network node. For example, a disclosure that a UE is configured to receive information from a base station also discloses that a first network node is configured to receive information from a second network node. Consistent with the present disclosure, once a specific example is expanded according to the present disclosure (e.g., a UE is configured to receive information from a base station also discloses that a first network node is configured to receive information from a second network node), a broader example of a narrower example may be interpreted in reverse, but in a broad, open-ended manner. In the above example that a UE is configured to receive information from a base station also discloses that a first network node is configured to receive information from a second network node, the first network node may refer to a first UE configured to receive information, a first base station, a first device, a first device, a first computing system, a first set of one or more components, a first processing entity, etc.; and the second network node may refer to a second UE, a second base station, a second device, a second device, a second computing system, a second set of one or more components, a second processing entity, etc.

[0265] As described herein, different terms may be used in various aspects to describe the communication of information (e.g., any information, signal, etc.). Disclosure of one communication term includes disclosure of other communication terms. For example, a first network node may be described as being configured to send information to a second network node. In this example and consistent with the present disclosure, disclosure that the first network node is configured to send information to the second network node includes disclosure that the first network node is configured to provide, transmit, output, communicate, or send information to the second network node. Similarly, in this example and consistent with the present disclosure, disclosure that the first network node is configured to send information to the second network node includes disclosure that the second network node is configured to receive, obtain, or decode information provided, transmitted, output, communicated, or sent by the first network node.

[0266] It will be appreciated by those skilled in the art that any of a variety of different technologies and techniques may be used to represent information and signals. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be mentioned throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or optical particles, or any combination thereof.

[0267] This article is about Figures 1 to 15 The components, functional blocks and modules include processors, electronic devices, hardware devices, electronic components, logic circuits, memories, software codes, firmware codes, etc., or any combination thereof. Software should be broadly interpreted as meaning instructions, instruction sets, codes, code segments, program codes, programs, subroutines, software modules, applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures, and / or functions, etc., whether they are referred to as software, firmware, middleware, microcode, hardware description language or other terms. In addition, the features discussed herein may be implemented via dedicated processor circuits, via executable instructions, or a combination thereof.

[0268] The technician will further understand that the various exemplary logic boxes, modules, circuits and algorithm steps described in conjunction with the disclosure of this article can be implemented as electronic hardware, computer software or a combination of the two. In order to clearly illustrate this interchangeability of hardware and software, various exemplary components, boxes, modules, circuits and steps have been generally described in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints proposed to the entire system. Those skilled in the art can implement the described functionality in different ways for each specific application, but such specific implementation decisions should not be interpreted as causing departure from the scope of the present disclosure. The technician will also easily recognize that the order or combination of components, methods or interactions described herein are only examples, and the components, methods or interactions of various aspects of the present disclosure can be combined or performed in a manner other than the manner illustrated and described herein.

[0269] The various illustrative logic components, logic blocks, modules, circuits, and algorithmic processes described in conjunction with the specific implementations disclosed herein may be implemented as electronic hardware, computer software, or a combination of both. The interchangeability of hardware and software has been generally described in terms of functionality and is exemplified in the various illustrative components, blocks, modules, circuits, and processes described above. Whether such functionality is implemented in hardware or software depends on the specific application and the design constraints imposed on the overall system.

[0270] The hardware and data processing devices for implementing the various exemplary logic components, logic blocks, modules and circuits described in conjunction with the various aspects disclosed herein may be implemented or executed using a general purpose single-chip or multi-chip processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic component, a discrete hardware component or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, or any conventional processor, controller, microcontroller or state machine. In some specific implementations, the processor may be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in combination with a DSP core, or any other such configuration. In some specific implementations, specific processes and methods may be performed by circuits specific to a given function.

[0271] In one or more aspects, the functions described may be implemented in hardware, digital electronic circuitry, computer software, firmware, including the structures disclosed in this specification and their structural equivalents, or any combination thereof. Specific implementations of the subject matter described in this specification may also be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on computer storage media for execution by data processing apparatus or for controlling the operation of data processing apparatus.

[0272] If implemented in software, the function may be stored as one or more instructions or codes on a computer-readable medium or sent through a computer-readable medium. The process of the method or algorithm disclosed herein may be implemented in a processor-executable software module that may reside on a computer-readable medium. Computer-readable media include both computer storage media and communication media, which include any media that can be implemented to transfer a computer program from one place to another. The storage medium may be any available medium that a computer can access. As an example and not limitation, such computer-readable media may include random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), CD-ROM or other optical disk storage devices, magnetic disk storage devices or other magnetic storage devices, or any other medium that can be used to store the required program code in the form of an instruction or data structure and can be accessed by a computer. In addition, any connection may be appropriately referred to as a computer-readable medium. Disks and optical disks as used herein include compact disks (CDs), laser disks, optical disks, digital versatile disks (DVDs), floppy disks, and blue-ray disks, wherein disks generally reproduce data magnetically, and optical disks reproduce data optically with lasers. The above combination should also be included in the scope of computer-readable media. Additionally, the operations of a method or algorithm may reside as a code and instruction set, or any combination of code and instruction sets, on a machine-readable medium or computer-readable medium, which may be incorporated into a computer program product.

[0273] Various modifications to the specific implementations described in this disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to some other specific implementations without departing from the spirit or scope of the disclosure. Therefore, the claims are not intended to be limited to the specific implementations shown herein, but are to be consistent with the broadest scope consistent with the disclosure, the principles and novel features disclosed herein.

[0274] Additionally, one of ordinary skill in the art will readily recognize that the terms "upper" and "lower" are sometimes used for ease of describing the drawings and indicate relative positions corresponding to the orientation of the drawings on a correctly oriented page, and may not reflect the correct orientation of any device as implemented.

[0275] Certain features described in this specification in the context of separate implementations may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented in multiple implementations, either individually or in any suitable subcombination. Furthermore, while features may be described above as functioning in certain combinations and even initially claimed as such, one or more features from a claimed combination may in some cases be removed from the combination, and a claimed combination may be directed to subcombinations or variations of subcombinations.

[0276] Similarly, although the operations are depicted in a specific order in the accompanying drawings, this should not be understood as requiring such operations to be performed in the specific order shown or in a sequential order, or to perform all illustrated operations to achieve the desired result. In addition, the accompanying drawings can schematically depict one or more example processes in the form of a flow chart. However, other operations that are not depicted can be combined in the example processes schematically illustrated. For example, one or more additional operations can be performed before, after, simultaneously or between any illustrated operations. In some environments, multitasking and parallel processing are advantageous. In addition, the separation of various system components in the specific implementation described above should not be understood as requiring such separation in all specific implementations, and it should be understood that the program components and systems can generally be integrated together in a single software product or packaged into multiple software products. Additionally, some other specific implementations also fall within the scope of the appended claims. In some cases, the actions described in the claims can be performed in different orders and still achieve the desired result.

[0277] As used herein, including in the claims, the term "or" used in a list of two or more items means that any of the listed items may be taken alone, or any combination of two or more of the listed items may be taken. For example, if a composition is described as containing components A, B, or C, the composition may contain A alone; B alone; C alone; A and B in combination; A and C in combination; B and C in combination; or A, B, and C in combination. In addition, as used herein, including in the claims, "or" as used in a list of items beginning with "at least one" indicates a separate list, so that, for example, a list of "at least one of A, B, or C" means A or B or C or AB or AC or BC or ABC (i.e., A and B and C) or any combination of any of these. The term "substantially" is defined as to a large extent, but not necessarily entirely, what is specified (and includes what is specified; for example, substantially 90 degrees includes 90 degrees, and substantially parallel includes parallel), as understood by one of ordinary skill in the art. In any disclosed implementation, the term "substantially" may be replaced with "within [percent] of" a specified content, where the percentage includes 0.1%, 1%, 5%, or 10%.

[0278] The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Therefore, the disclosure is not intended to be limited to the examples and designs described herein, but to be consistent with the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A first network node for wireless communication, comprising: at least one processor; and a memory coupled to the at least one processor, wherein the at least one processor is configured to: obtaining channel state information data associated with a second network node; training a shared UE encoder based on the channel state information data and based on a decoder to generate a sequential training data set; as well as The sequential training data set is sent to a third network node.

2. The first network node according to claim 1, wherein the sequential training data set comprises a UE-driven sequential training data set, and the UE-driven sequential training data set is configured to implement sequential training of the decoder of the third network node based on concurrent training of the shared UE encoder and the decoder at the first network node.

3. The first network node according to claim 1, wherein the sequential training data set comprises: (z, Vin), wherein Vin comprises an input vector for the shared UE encoder and z comprises an output from the shared UE encoder based on Vin; or (z, Vout), where z comprises the decoder input of the vector and Vout comprises the decoder output of the vector.

4. The first network node according to claim 1, wherein the channel state information data comprises or corresponds to a pre-decoder vector or a channel matrix.

5. The first network node of claim 1 , wherein the at least one processor is configured to: encoding an uncompressed or raw channel state feedback (CSF) using the shared UE encoder to generate a compressed CSF; as well as decoding the compressed CSF using the decoder to generate a reconstructed or decompressed CSF; as well as comparing the reconstructed or decompressed CSF with the uncompressed or original CSF; as well as The shared UE encoder, the decoder, or both are adjusted based on the comparison.

6. The first network node of claim 1, wherein the first network node comprises a UE server, wherein the second network node comprises a UE, and wherein the third network node comprises a base station server.

7. The first network node of claim 1 , wherein the at least one processor is configured to: receiving second channel state information data associated with a fourth network node; and Generating aggregate channel state information based on the channel state information data and the second channel state information data, and wherein the at least one processor is configured to train the shared UE encoder based on the channel state information data, comprises: The shared UE encoder is trained based on the aggregated channel state information.

8. The first network node of claim 1 , wherein the at least one processor is configured to: receiving second channel state information data associated with a fourth network node; training the shared UE encoder based on the second channel state information data to update the sequential training data set and generate an updated sequential training data set; and The updated sequential training data set is sent.

9. The first network node of claim 1 , wherein the at least one processor is configured to train the shared UE encoder, comprising: performing training of the shared UE encoder and the decoder to generate the sequential training data set and encoder model weights, and wherein the at least one processor is further configured to: The encoder model weights are sent to the second network node.

10. The first network node of claim 1, wherein the decoder corresponds to a reference decoder of a base station, wherein the at least one processor is configured to: The type of the reference decoder is determined based on the type of the shared UE encoder.

11. The first network node according to claim 10, wherein the at least one processor is configured to: The type of the reference decoder is determined based on a type or architecture of the shared UE encoder.

12. The first network node of claim 1 , wherein the decoder corresponds to a reference decoder of a base station, wherein the at least one processor is configured to: obtaining reference decoder information of a base station; and A reference decoder is determined based on the reference decoder information of the base station.

13. The first network node of claim 1 , wherein the decoder corresponds to a reference decoder of a base station, wherein the at least one processor is configured to: obtaining reference decoder information and decoder model weights of a decoder of a base station, wherein the decoder model weights include initial weights or final weights; and The reference decoder is determined based on the reference decoder information of the base station, wherein the shared UE encoder is further trained based on the decoder model weights.

14. The first network node of claim 1 , wherein the at least one processor is configured to: The sequential training data set is sent to a fourth network node.

15. The first network node of claim 1, wherein the first network node comprises a UE, and wherein The data is sent to a fourth network node by encoding the data based on encoder model information generated based on training the shared UE encoder.

16. The first network node according to claim 15, wherein: The shared UE encoder is a CSI encoder, and encoding the data comprises encoding the CSI data to generate compressed CSI data; or The shared UE encoder is a precoding information encoder, and encoding the data includes encoding precoding information to generate compressed precoding information.

17. A first network node for wireless communication, comprising: at least one processor; and a memory coupled to the at least one processor, wherein the at least one processor is configured to: receiving a sequential training data set from a second network node; Training a base station decoder based on the sequential training data set to generate decoder model information; as well as The decoder model information for the base station decoder is sent to a third network node.

18. The first network node of claim 17, wherein the decoder model information enables other network nodes to train a shared base station decoder for decoding encoded data from a plurality of different types of UEs.

19. The first network node of claim 17, wherein the at least one processor is configured to: receiving a second sequential training data set from a fourth network node; and generating an aggregated sequential training data set based on the sequential training data set and the second sequential training data set, and wherein the at least one processor is configured to train the base station decoder based on the sequential training data set, comprising: The base station decoder is trained based on the aggregated sequential training data set to generate the decoder model information.

20. The first network node of claim 17, wherein the at least one processor is configured to: Reference decoder information is sent to a UE or a UE server, wherein the reference decoder information enables the UE or the UE server to use the reference decoder information as a reference decoder when training a UE encoder.

21. The first network node according to claim 20, wherein the reference decoder information comprises decoder architecture information, decoder layer information, decoder category information, or a combination thereof.

22. The first network node according to claim 21, wherein the decoder class information indicates decoder architecture complexity information, decoder layer complexity information or a combined complexity level.

23. The first network node of claim 17, wherein the at least one processor is configured to: Sending reference decoder information and decoder model weights to a UE or a UE server, wherein the reference decoder information and the decoder model weights enable the UE or the UE server to use the reference decoder information and the decoder model weights as a reference decoder when training a UE encoder, wherein the decoder model weights include initial weights or final weights.

24. The first network node of claim 17, wherein the base station decoder comprises: Preprocessor; and Share a common decoder.

25. The first network node of claim 24, wherein the pre-processor is configured to perform one-hot encoding.

26. The first network node of claim 24, wherein the pre-processor comprises a plurality of UE-specific layers.

27. The first network node of claim 24, wherein the pre-processor comprises a set of common processing layers, and wherein the set of common processing layers comprises: a first linear layer configured to receive an output of a one-hot encoder; a Gaussian layer configured to receive an output of the first linear layer; and A second linear layer is configured to receive the output of the Gaussian layer and provide input to the shared common decoder.

28. The first network node of claim 17, wherein the base station decoder comprises a universal base station decoder, the universal base station decoder comprising: one or more UE-specific layers, the one or more UE-specific layers configured to pre-process the compressed CSI based on the stored per-UE parameters; one or more common layers configured to decode the pre-processed CSI; and The stored per-UE parameters.

29. A first network node for wireless communication, comprising: at least one processor; and a memory coupled to the at least one processor, wherein the at least one processor is configured to: sending channel state information data to a second network node; receiving encoder model information from the second network node, the encoder model information being based on the channel state information data; as well as The data is transmitted to a third network node by encoding the data based on the encoder model information.

30. A first network node for wireless communication, comprising: at least one processor; and a memory coupled to the at least one processor, wherein the at least one processor is configured to: receiving decoder model information for a shared base station decoder from a second network node; and The encoded data is received from a third network node by decoding the encoded data based on the shared base station decoder.

Images