Offline multi-vendor training for cross-node machine learning

Abstract

Aspects of the present disclosure generally relate to wireless communications. In some aspects, a first server may send encoder training data regarding an encoder to a set of second servers associated with training a set of decoders and associated with a set of vendors, where the first server is associated with a particular vendor of the set of vendors. The first server may receive decoder training data associated with a configuration of the encoder from the set of second servers and based at least in part on transmitting the encoder training data. Numerous other aspects are described.

Description

Technical Field

[0001] Aspects of the present disclosure relate generally to wireless communications, and to techniques and apparatus for offline multi-vendor training for cross-node machine learning. Background Art

[0002] Wireless communication systems are widely deployed to provide a variety of telecommunication services, such as telephony, video, data, messaging, and broadcasting. Typical wireless communication systems may employ multiple access technologies that can support communication with multiple users by sharing available system resources (e.g., bandwidth, transmit power, etc.). Examples of such multiple access technologies include code division multiple access (CDMA) systems, time division multiple access (TDMA) systems, frequency division multiple access (FDMA) systems, orthogonal frequency division multiple access (OFDMA) systems, single carrier frequency division multiple access (SC-FDMA) systems, time division synchronous code division multiple access (TD-SCDMA) systems, and long term evolution (LTE). LTE / LTE-Advanced is a set of enhancements to the Universal Mobile Telecommunications System (UMTS) mobile standard promulgated by the Third Generation Partnership Project (3GPP).

[0003] A wireless network may include one or more network nodes that support communications for wireless communication devices, such as user equipment (UE) or multiple UEs. A UE may communicate with a network node via downlink communications and uplink communications. A "downlink" (or "DL") refers to a communication link from a network node to a UE, and an "uplink" (or "UL") refers to a communication link from a UE to a network node. Some wireless networks may support device-to-device communications, such as via a local link (e.g., a side link (SL), a wireless local area network (WLAN) link, and / or a wireless personal area network (WPAN) link, etc.).

[0004] The above-mentioned multiple access technologies have been adopted in various telecommunication standards to provide a common protocol that enables different UEs to communicate at a city, country, region and / or global level. New Radio (NR) (which may be referred to as 5G) is an enhancement set to the LTE mobile standard promulgated by 3GPP. NR is designed to better support mobile broadband Internet access by using orthogonal frequency division multiplexing (OFDM) (CP-OFDM) with a cyclic prefix (CP) on the downlink, using CP-OFDM and / or single carrier frequency division multiplexing (SC-FDM) (also known as discrete Fourier transform spread OFDM (DFT-s-OFDM)) on the uplink, and supporting beamforming, multiple input multiple output (MIMO) antenna technology and carrier aggregation to improve spectrum efficiency, reduce costs, improve services, utilize new spectrum, and better integrate with other open standards. As the demand for mobile broadband access continues to increase, further improvements in LTE, NR and other radio access technologies remain useful. Summary of the invention

[0005] Some aspects described herein relate to a method performed by a first server. The method may include sending encoder training data about an encoder to a set of second servers associated with training a set of decoders and associated with a set of vendors, wherein the first server is associated with a particular vendor in the set of vendors. The method may include receiving decoder training data associated with a configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data.

[0006] Some aspects described herein relate to a method performed by a second server. The method may include receiving encoder training data for a plurality of encoders from a plurality of first servers associated with the plurality of encoders and a set of vendors, wherein the second server is associated with a particular vendor in the set of vendors. The method may include sending decoder training data for a decoder associated with a configuration of the plurality of encoders to the plurality of first servers and based at least in part on receiving the encoder training data.

[0007] Some aspects described herein relate to a first server. The first server may include a memory and one or more processors coupled to the memory. The one or more processors may be configured to send encoder training data about an encoder to a set of second servers associated with training a set of decoders and associated with a set of suppliers, wherein the first server is associated with a particular supplier in the set of suppliers. The one or more processors may be configured to receive decoder training data associated with a configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data.

[0008] Some aspects described herein relate to a second server. The second server may include a memory and one or more processors coupled to the memory. The one or more processors may be configured to receive encoder training data about the multiple encoders from multiple first servers associated with the multiple encoders and a set of suppliers, wherein the second server is associated with a specific supplier in the set of suppliers. The one or more processors may be configured to send decoder training data associated with the configuration of the multiple encoders to the multiple first servers and based at least in part on receiving the encoder training data.

[0009] Some aspects described herein relate to a non-transitory computer-readable medium storing a set of instructions by a first server. The set of instructions, when executed by one or more processors of the first server, may cause the first server to send encoder training data about an encoder to a set of second servers associated with training a set of decoders and associated with a set of suppliers, wherein the first server is associated with a particular supplier in the set of suppliers. The set of instructions, when executed by one or more processors of the first server, may cause the first server to receive decoder training data associated with a configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data.

[0010] Some aspects described herein relate to a non-transitory computer-readable medium storing a set of instructions by a second server. The set of instructions, when executed by one or more processors of the second server, may cause the second server to receive encoder training data about the plurality of encoders from a plurality of first servers associated with the plurality of encoders and a set of suppliers, wherein the second server is associated with a particular supplier in the set of suppliers. The set of instructions, when executed by one or more processors of the second server, may cause the second server to send decoder training data of a decoder associated with a configuration of the plurality of encoders to the plurality of first servers and based at least in part on receiving the encoder training data.

[0011] Some aspects described herein are directed to a first device. The first device may include means for sending encoder training data for an encoder to a set of second devices associated with training a set of decoders and associated with a set of vendors, wherein the first device is associated with a particular vendor in the set of vendors. The first device may include means for receiving decoder training data associated with a configuration of the encoder from the set of second devices and based at least in part on sending the encoder training data.

[0012] Some aspects described herein relate to a second device. The second device may include means for receiving encoder training data for a plurality of encoders from a plurality of first devices associated with a plurality of encoders and a set of vendors, wherein the second device is associated with a particular vendor in the set of vendors. The second device may include means for sending decoder training data for a decoder associated with a configuration of the plurality of encoders to the plurality of first devices and based at least in part on receiving the encoder training data.

[0013] The various aspects generally include methods, apparatus, systems, computer program products, non-transitory computer-readable media, user equipment, base stations, network entities, network nodes, wireless communication devices and / or processing systems as fully described herein with reference to the drawings and description and as illustrated in the drawings and description.

[0014] The features and technical advantages of examples according to the present disclosure have been outlined quite extensively 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. The characteristics of the concepts disclosed herein, both in terms of their organization and method of operation, and the associated advantages will be better understood by considering the following description in conjunction with the accompanying drawings. Each of the figures in the accompanying drawings is provided for the purpose of illustration and description and not as a definition of limitations to the claims.

[0015] Although various aspects are described in the present disclosure by illustrating some examples, it will be understood by those skilled in the art that such aspects can be implemented in many different arrangements and scenarios. The technology described herein can be implemented using different platform types, devices, systems, shapes, sizes and / or packaging arrangements. For example, some aspects can be implemented via integrated chip implementations or other devices based on non-module components (e.g., end-user devices, vehicles, communication equipment, computing equipment, industrial equipment, retail / shopping equipment, medical equipment and / or artificial intelligence devices). Various aspects can be implemented in chip-level components, modular components, non-modular components, non-chip-level components, device-level components and / or system-level components. The equipment incorporating the various aspects and features described may include additional components and features for implementing and practicing the various aspects claimed and described. For example, the transmission and reception of wireless signals may include one or more components (e.g., hardware components, including antennas, radio frequency (RF) chains, power amplifiers, modulators, buffers, processors, interleavers, adders and / or summers) for analog and digital purposes. The various aspects described herein are intended to be practiced in various devices, components, systems, distributed arrangements and / or end-user devices of various sizes, shapes and configurations. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order that the above-mentioned features of the present disclosure may be fully understood, a more specific description of the invention briefly summarized above may be obtained by reference to various aspects, some of which are illustrated in the accompanying drawings. However, it should be noted that the accompanying drawings illustrate only certain typical aspects of the present disclosure and are therefore not to be considered as limiting the scope thereof, as the specification may admit of other equally effective aspects. The same reference numerals in different drawings may identify the same or similar elements.

[0017] Figure 1 is a diagram illustrating an example of a wireless network according to the present disclosure.

[0018] Figure 2 is a diagram illustrating an example of communication between a network node and a user equipment (UE) in a wireless network according to the present disclosure.

[0019] Figure 3 is a diagram illustrating an example decomposed base station architecture according to the present disclosure.

[0020] Figure 4 is a diagram illustrating an example of physical channels and reference signals in a wireless network according to the present disclosure.

[0021] FIG. 5A to FIG. 5D is a diagram illustrating an example associated with offline multi-vendor training for cross-node machine learning according to the present disclosure.

[0022] Figure 6 is a diagram illustrating an example process performed, for example, by a first server according to the present disclosure.

[0023] Figure 7 is a diagram illustrating an example process performed, for example, by a second server according to the present disclosure.

[0024] Figure 8 is a diagram of an example apparatus for wireless communications according to the present disclosure.

[0025] Fig. 9 is a diagram of an example apparatus for wireless communications according to the present disclosure.

[0026] Fig.10 is a diagram of an example apparatus for wireless communications according to the present disclosure.

[0027] Fig.11 is a diagram illustrating an example of training and using a machine learning model in conjunction with offline multi-vendor training for cross-node machine learning according to the present disclosure. DETAILED DESCRIPTION

[0028] The various aspects of the present disclosure are described more fully below with reference to the accompanying drawings. However, the present disclosure can be embodied in many different forms, and should not be interpreted as being limited to any specific structure or function presented throughout the present disclosure. On the contrary, these aspects are provided so that the present disclosure will be thorough and complete, and the scope of the present disclosure will be fully conveyed to those skilled in the art. It should be understood by those skilled in the art that the scope of the present disclosure is intended to cover any aspect of the disclosure disclosed herein, whether it is implemented independently or in combination with any other aspect of the present disclosure. For example, any number of aspects set forth herein may be used to implement a device or practice method. In addition, the scope of the present disclosure is intended to cover such a device or method that is practiced using other structures, functions, or structures and functions in addition to or different from the various aspects of the disclosure set forth herein. It should be understood that any aspect of the disclosure disclosed herein may be embodied by one or more elements of the claims.

[0029] Several aspects of telecommunication systems will now be presented with reference to various devices and techniques. These devices and techniques will be described in the following detailed description and illustrated in the accompanying drawings by various blocks, modules, components, circuits, steps, processes, algorithms, etc. (collectively referred to as "elements"). These elements can be implemented using hardware, software, or a combination thereof. Whether these elements are implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system.

[0030] Although various aspects may be described herein using terms generally associated with 5G or new radio (NR) radio access technologies (RATs), various aspects of the present disclosure may be applicable to other RATs, such as 3G RATs, 4G RATs, and / or post-5G RATs (e.g., 6G).

[0031] Figure 1 1 is a diagram illustrating an example of a wireless network 100 according to the present disclosure. The wireless network 100 may be a 5G (e.g., NR) network and / or a 4G (e.g., Long Term Evolution (LTE)) network, or may include elements of a 5G (e.g., NR) network and / or elements of a 4G (e.g., Long Term Evolution (LTE)) network, as well as other examples. The wireless network 100 may include one or more network nodes 110 (shown as network node 110a, network node 110b, network node 110c, and network node 110d), user equipment (UE) 120 or multiple UEs 120 (shown as UE 120a, UE 120b, UE 120c, UE 120d, and UE 120e), and / or other entities. The network node 110 is a network node that communicates with the UE 120. As shown in the figure, the network node 110 may include one or more network nodes. For example, the network node 110 may be a converged network node, which means that the converged network node is configured to utilize a radio protocol stack that is physically or logically integrated within a single radio access network (RAN) node (e.g., within a single device or unit). As another example, the network node 110 may be a decomposed network node (sometimes referred to as a decomposed base station), which means that the network node 110 is configured to utilize a protocol stack that is physically or logically distributed between two or more nodes (such as one or more central units (CUs), one or more distributed units (DUs), or one or more radio units (RUs)).

[0032] In some examples, the network node 110 is or includes a network node that communicates with the UE 120 via a radio access link, such as an RU. In some examples, the network node 110 is or includes a network node that communicates with other network nodes 110 via a fronthaul link or a midhaul link, such as a DU. In some examples, the network node 110 is or includes a network node that communicates with other network nodes 110 via a midhaul link or communicates with the core network via a backhaul link, such as a CU. In some examples, the network node 110 (such as an aggregated network node 110 or a decomposed network node 110) may include multiple network nodes, such as one or more RUs, one or more CUs, and / or one or more DUs. The network node 110 may include, for example, an NR base station, an LTE base station, a Node B, an eNB (e.g., in 4G), a gNB (e.g., in 5G), an access point, a transmit receive point (TRP), a DU, a RU, a CU, a mobility element of a network, a core network node, a network element, a network equipment, a RAN node, or a combination thereof. In some examples, network nodes 110 may be interconnected to each other or to one or more other network nodes 110 in wireless network 100 via various types of fronthaul, midhaul, and / or backhaul interfaces, such as direct physical connections, air interfaces, or virtual networks, using any suitable transport network.

[0033] In some examples, the network node 110 may provide communication coverage for a particular geographic area. In the Third Generation Partnership Project (3GPP), the term "cell" may refer to the coverage area of ​​the network node 110 and / or the network node subsystem serving the coverage area, depending on the context in which the term is used. The network node 110 may provide communication coverage for a macro cell, a pico cell, a femto cell, and / or another type of cell. A macro cell may cover a relatively large geographic area (e.g., a radius of several kilometers) and may allow unrestricted access by a UE 120 with a service subscription. A pico cell may cover a relatively small geographic area and may allow unrestricted access by a UE 120 with a service subscription. A femto cell may cover a relatively small geographic area (e.g., a home) and may allow restricted access by a UE 120 associated with the femto cell (e.g., a UE 120 in a closed subscriber group (CSG)). A network node 110 for a macro cell may be referred to as a macro network node. A network node 110 for a pico cell may be referred to as a pico network node. The network node 110 for a femto cell may be referred to as a femto network node or a home network node. Figure 1In the example shown in , network node 110a may be a macro network node for macro cell 102a, network node 110b may be a pico network node for pico cell 102b, and network node 110c may be a femto network node for femto cell 102c. A network node may support one or more (e.g., three) cells. In some examples, a cell may not necessarily be stationary, and the geographic area of ​​a cell may move depending on the location of a mobile network node 110 (e.g., a mobile network node).

[0034] In some aspects, the term "base station" or "network node" may refer to a clustered base station, a decomposed base station, an integrated access and backhaul (IAB) node, a relay node, or one or more components thereof. For example, in some aspects, a "base station" or "network node" may refer to a CU, a DU, a RU, a near real-time (near RT) RAN intelligent controller (RIC) or a non-real-time (non-RT) RIC, or a combination thereof. In some aspects, the term "base station" or "network node" may refer to a device configured to perform one or more functions (such as those described herein in conjunction with network node 110). In some aspects, the term "base station" or "network node" may refer to multiple devices configured to perform one or more functions. For example, in some distributed systems, each of a number of different devices (which may be located in the same geographic location or in different geographic locations) may be configured to perform at least a portion of a function, or to repeat the execution of at least a portion of the function, and the term "base station" or "network node" may refer to any one or more of these different devices. In some aspects, the term "base station" or "network node" may refer to one or more virtual base stations or one or more virtual base station functions. For example, in some aspects, two or more base station functions may be instantiated on a single device. In some aspects, the term "base station" or "network node" may refer to one of the base station functions but not the other. In this way, a single device may include more than one base station.

[0035] The wireless network 100 may include one or more relay stations. A relay station is a network node that can receive transmissions of data from an upstream node (e.g., a network node 110 or a UE 120) and transmit transmissions of data to a downstream node (e.g., a UE 120 or a network node 110). A relay station may be a UE 120 that is capable of relaying transmissions for other UEs 120. Figure 1 In the example shown in , a network node 110d (e.g., a relay network node) may communicate with a network node 110a (e.g., a macro network node) and a UE 120d to facilitate communications between the network node 110a and the UE 120d. A network node 110 that relays communications may be referred to as a relay station, a relay base station, a relay network node, a relay node, a relay, etc.

[0036] The wireless network 100 may be a heterogeneous network that includes different types of network nodes 110, such as macro network nodes, pico network nodes, femto network nodes, relay network nodes, etc. These different types of network nodes 110 may have different transmit power levels, different coverage areas, and / or different impacts on interference in the wireless network 100. For example, a macro network node may have a high transmit power level (e.g., 5 watts to 40 watts), while a pico network node, a femto network node, and a relay network node may have a lower transmit power level (e.g., 0.1 watt to 2 watts).

[0037] The network controller 130 may be coupled to or in communication with a set of network nodes 110 and may provide coordination and control for the network nodes 110. The network controller 130 may communicate with the network nodes 110 via a backhaul communication link or a midhaul communication link. The network nodes 110 may also communicate directly with each other or indirectly via a wireless backhaul communication link or a wired backhaul communication link. In some aspects, the network controller 130 may be or may include a CU or a core network device.

[0038] UE 120 can be dispersed throughout the wireless network 100, and each UE 120 can be stationary or mobile. UE 120 can include, for example, an access terminal, a terminal, a mobile station and / or a subscriber unit. UE 120 can be a cellular phone (e.g., a smart phone), a personal digital assistant (PDA), a wireless modem, a wireless communication device, a handheld device, a laptop computer, a cordless phone, a wireless local loop (WLL) station, a tablet computer, a camera, a gaming device, a netbook, a smartbook, an ultrabook, a medical device, a biometric device, a wearable device (e.g., a smart watch, smart clothing, smart glasses, a smart wristband, smart jewelry (e.g., a smart ring or a smart bracelet)), an entertainment device (e.g., a music device, a video device and / or a satellite radio), a vehicle component or sensor, a smart meter / sensor, an industrial manufacturing equipment, a global positioning system device, a UE function of a network node and / or any other suitable device configured to communicate via a wireless or wired medium.

[0039] Some UEs 120 may be considered as machine type communication (MTC) or evolved or enhanced machine type communication (eMTC) UEs. MTC UEs and / or eMTC UEs may include, for example, robots, drones, remote devices, sensors, meters, monitors, and / or location tags, which may communicate with a network node, another device (e.g., a remote device), or some other entity. Some UEs 120 may be considered as Internet of Things (IoT) devices and / or may be implemented as NB-IoT (narrowband IoT) devices. Some UEs 120 may be considered as customer premises equipment. UE 120 may be included inside a housing that houses components of UE 120, such as a processor component and / or a memory component. In some examples, the processor component and the memory component may be coupled together. For example, a processor component (e.g., one or more processors) and a memory component (e.g., a memory) may be operably coupled, communicatively coupled, electronically coupled, and / or electrically coupled.

[0040] In general, any number of wireless networks 100 may be deployed in a given geographic area. Each wireless network 100 may support a specific RAT and may operate on one or more frequencies. RAT may be referred to as a radio technology, air interface, etc. Frequency may be referred to as a carrier, frequency channel, etc. In a given geographic area, each frequency may support a single RAT to avoid interference between wireless networks of different RATs. In some cases, NR or 5G RAT networks may be deployed.

[0041] In some examples, two or more UEs 120 (e.g., shown as UE 120a and UE 120e) may communicate directly (e.g., without using network node 110 as an intermediary to communicate with each other) using one or more side link channels. For example, UE 120 may communicate using peer-to-peer (P2P) communication, device-to-device (D2D) communication, vehicle-to-everything (V2X) protocols (e.g., which may include vehicle-to-vehicle (V2V) protocols, vehicle-to-infrastructure (V2I) protocols, or vehicle-to-pedestrian (V2P) protocols), and / or mesh networks. In such examples, UE 120 may perform scheduling operations, resource selection operations, and / or other operations described elsewhere herein as being performed by network node 110.

[0042] The devices of the wireless network 100 may communicate using an electromagnetic spectrum that may be subdivided into various categories, bands, channels, etc., based on frequency or wavelength. For example, the devices of the wireless network 100 may communicate using one or more operating bands. In 5G NR, two initial operating bands have been identified as frequency range designations FR1 (410 MHz-7.125 GHz) and FR2 (24.25 GHz-52.6 GHz). It should be understood that, although a portion of FR1 is greater than 6 GHz, FR1 is often (interchangeably) referred to as the "below 6 GHz" band in various documents and articles. A similar naming issue sometimes occurs with respect to FR2, which is often (interchangeably) referred to as the "millimeter wave" band in documents and articles, although it is different from the extremely high frequency (EHF) band (30 GHz-300 GHz) identified as the "millimeter wave" band by the International Telecommunication Union (ITU).

[0043] Frequencies between FR1 and FR2 are generally referred to as mid-band frequencies. Recent 5G NR research has identified the operating bands for these mid-band frequencies as frequency range designation FR3 (7.125GHz-24.25GHz). The bands falling within FR3 can inherit FR1 characteristics and / or FR2 characteristics, so the features of FR1 and / or FR2 can be effectively extended to mid-band frequencies. In addition, higher frequency bands are currently being explored to extend 5G NR operations to more than 52.6GHz. For example, three higher operating bands have been identified as frequency range designations FR4a or FR4-1 (52.6GHz-71GHz), FR4 (52.6GHz-114.25GHz) and FR5 (114.25GHz-300GHz). Each of these higher frequency bands falls within the EHF band.

[0044] In view of the above examples, unless otherwise specifically stated, it should be understood that if the term "below 6 GHz" or the like is used herein, the term may broadly refer to frequencies that may be lower than 6 GHz, may be within FR1, or may include mid-band frequencies. In addition, unless otherwise specifically stated, it should be understood that if the term "millimeter wave" or the like is used herein, the term may broadly refer to frequencies that may include mid-band frequencies, may be within FR2, FR4, FR4-a or FR4-1, and / or FR5, or may be within the EHF band. It is contemplated that the frequencies included in these operating bands (e.g., FR1, FR2, FR3, FR4, FR4-a, FR4-1, and / or FR5) may be modified, and the techniques described herein are applicable to those modified frequency ranges.

[0045] In some aspects, UE 120 may include a communication manager 140. As described in greater detail elsewhere herein, communication manager 140 may receive an encoder trained by UE server 180; may use the encoder to encode and / or compress channel state feedback (CSF); and may send the encoded and / or compressed CSF to network node 110 for processing. Additionally or alternatively, communication manager 140 may perform one or more other operations described herein.

[0046] In some aspects, network node 110 may include communication manager 150. As described in greater detail elsewhere herein, communication manager 150 may receive a decoder from access point (AP) server 190; may receive an encoded and / or compressed CSF from UE 120; may process the encoded and / or compressed CSF to decode and / or decompress and recover the CSF; and may configure communications on a channel based at least in part on recovering the CSF. Additionally or alternatively, communication manager 150 may perform one or more other operations described herein.

[0047] In some aspects, the first server (e.g., UE server 180) may include a communication manager 182. As described in more detail elsewhere herein, the communication manager 182 may send encoder training data about the encoder to a set of second servers associated with training a set of decoders and associated with a set of vendors, wherein the first server is associated with a particular vendor in the set of vendors; and receive decoder training data associated with the configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data. Additionally or alternatively, the communication manager 182 may perform one or more other operations described herein.

[0048] In some aspects, the second server (e.g., AP server 190) may include a communication manager 192. As described in more detail elsewhere herein, the communication manager 192 may receive encoder training data about the plurality of encoders from a plurality of first servers associated with the plurality of encoders and a set of vendors, wherein the second server is associated with a particular vendor in the set of vendors; and send decoder training data of the decoders associated with the configuration of the plurality of encoders to the plurality of first servers and based at least in part on receiving the encoder training data. Additionally or alternatively, the communication manager 192 may perform one or more other operations described herein.

[0049] As indicated above, Figure 1 are provided as examples. Other examples can be found in the Figure 1 Different than described.

[0050] Figure 22 is a diagram illustrating an example 200 of a network node 110 communicating with a UE 120 in a wireless network 100 according to the present disclosure. The network node 110 may be equipped with a set of antennas 234a to 234t, such as T antennas (T≥1). The UE 120 may be equipped with a set of antennas 252a to 252r, such as R antennas (R≥1). The network node 110 of example 200 includes one or more radio frequency components, such as an antenna 234 and a modem 254. In some examples, the network node 110 may include an interface, a communication component, or another component that facilitates communication with the UE 120 or another network node. Some network nodes 110 may not include a radio frequency component that facilitates direct communication with the UE 120, such as one or more CUs or one or more DUs.

[0051] At the network node 110, the transmit processor 220 may receive data intended for the UE 120 (or a set of UEs 120) from the data source 212. The transmit processor 220 may select one or more modulation and coding schemes (MCS) for the UE 120 based at least in part on one or more channel quality indicators (CQI) received from the UE 120. The network node 110 may process (e.g., encode and modulate) the data for the UE 120 based at least in part on the MCS selected for the UE 120, and may provide data symbols for the UE 120. The transmit processor 220 may process system information (e.g., for semi-static resource allocation information (SRPI)) and control information (e.g., CQI requests, grants, and / or upper layer signaling), and provide overhead symbols and control symbols. The transmit processor 220 may generate reference symbols for reference signals (e.g., cell-specific reference signals (CRS) or demodulation reference signals (DMRS)) and synchronization signals (e.g., primary synchronization signals (PSS) or secondary synchronization signals (SSS)). The transmit (TX) multiple-input multiple-output (MIMO) processor 230 may perform spatial processing (e.g., pre-coding) on ​​data symbols, control symbols, overhead symbols, and / or reference symbols, where applicable, and may provide a set of output symbol streams (e.g., T output symbol streams) to a corresponding set of modems 232 (e.g., T modulators) (shown as modems 232a to 232t). For example, each output symbol stream may be provided to a modulator component (shown as MOD) of the modem 232. Each modem 232 may process a corresponding output symbol stream (e.g., for OFDM) using a corresponding modulator component to obtain an output sample stream. Each modem 232 may also process (e.g., convert to analog, amplify, filter, and / or up-convert) the output sample stream using a corresponding modulator component to obtain a downlink signal. The modems 232a to 232t may transmit a set of downlink signals (e.g., T downlink signals) via a set of corresponding antennas 234 (e.g., T antennas) (shown as antennas 234a to 234t).

[0052] At the UE 120, a set of antennas 252 (shown as antennas 252a to 252r) may receive downlink signals from the network node 110 and / or other network nodes 110 and may provide a set of received signals (e.g., R received signals) to a set of modems 254 (e.g., R modems) (shown as modems 254a to 254r). For example, each received signal may be provided to a demodulator component (shown as DEMOD) of the modem 254. Each modem 254 may use a corresponding demodulator component to condition (e.g., filter, amplify, downconvert, and / or digitize) the received signal to obtain input samples. Each modem 254 may use a demodulator component to further process the input samples (e.g., for OFDM) to obtain received symbols. A MIMO detector 256 may obtain received symbols from the modem 254, may perform MIMO detection on the received symbols where applicable, and may provide detected symbols. The receive processor 258 may process (e.g., demodulate and decode) the detected symbols, may provide decoded data for the UE 120 to the data sink 260, and may provide decoded control information and system information to the controller / processor 280. The term "controller / processor" may refer to one or more controllers, one or more processors, or a combination thereof. The channel processor may determine a reference signal received power (RSRP) parameter, a received signal strength indicator (RSSI) parameter, a reference signal received quality (RSRQ) parameter, and / or a CQI parameter, among other things. In some examples, one or more components of the UE 120 may be included in the housing 284.

[0053] The network controller 130 may include a communication unit 294, a controller / processor 290, and a memory 292. The network controller 130 may include, for example, one or more devices in a core network. The network controller 130 may communicate with the network node 110 via the communication unit 294.

[0054] One or more antennas (e.g., antennas 234a to 234t and / or antennas 252a to 252r) may include, or may be included within, one or more antenna panels, one or more antenna groups, one or more groups of antenna elements, and / or one or more antenna arrays, etc. An antenna panel, an antenna group, a group of antenna elements, and / or an antenna array may include one or more antenna elements (in a single housing or multiple housings), a group of coplanar antenna elements, a group of non-coplanar antenna elements, and / or may be coupled to one or more transmit and / or receive components (such as, Figure 2 One or more antenna elements of one or more components).

[0055] On the uplink, at the UE 120, the transmit processor 264 may receive and process data from the data source 262 and control information from the controller / processor 280 (e.g., for reports including RSRP, RSSI, RSRQ, and / or CQI). The transmit processor 264 may generate reference symbols for one or more reference signals. The symbols from the transmit processor 264 may be pre-decoded by the TX MIMO processor 266, where applicable, further processed by the modem 254 (e.g., for DFT-s-OFDM or CP-OFDM), and sent to the network node 110. In some examples, the modem 254 of the UE 120 may include a modulator and a demodulator. In some examples, the UE 120 includes a transceiver. The transceiver may include any combination of an antenna 252, a modem 254, a MIMO detector 256, a receive processor 258, a transmit processor 264, and / or a TX MIMO processor 266. The transceiver may be used by a processor (eg, controller / processor 280) and memory 282 to perform aspects of any of the methods described herein (eg, with reference to FIG. 5A to FIG. 11 ).

[0056] At the network node 110, uplink signals from the UE 120 and / or other UEs may be received by the antenna 234, processed by the modem 232 (e.g., a demodulator component (shown as DEMOD) of the modem 232), detected by the MIMO detector 236 (where applicable), and further processed by the receive processor 238 to obtain decoded data and control information transmitted by the UE 120. The receive processor 238 may provide the decoded data to the data sink 239 and the decoded control information to the controller / processor 240. The network node 110 may include a communication unit 244 and may communicate with the network controller 130 via the communication unit 244. The network node 110 may include a scheduler 246 to schedule one or more UEs 120 for downlink communication and / or uplink communication. In some examples, the modem 232 of the network node 110 may include a modulator and a demodulator. In some examples, the network node 110 includes a transceiver. The transceiver may include any combination of antenna 234, modem 232, MIMO detector 236, receive processor 238, transmit processor 220, and / or TX MIMO processor 230. The transceiver may be used by a processor (e.g., controller / processor 240) and memory 242 to perform aspects of any of the methods described herein (e.g., reference 200 to FIG. 1 ). FIG. 5A to FIG. 11 ).

[0057] The controller / processor 240 of the network node 110, the controller / processor 280 of the UE 120, and / or Figure 2 Any other component in the network node 110 may perform one or more techniques associated with offline multi-vendor training for cross-node machine learning, as described in more detail elsewhere herein. For example, the controller / processor 240 of the network node 110, the controller / processor 280 of the UE 120, and / or Figure 2 Any other component of the Figure 6 The process of 600 Figure 7 700 and / or operations of other processes as described herein. Memory 242 and memory 282 may store data and program codes for network node 110 and UE 120, respectively. In some examples, memory 242 and / or memory 282 may include a non-transitory computer-readable medium storing one or more instructions (e.g., code and / or program code) for wireless communication. For example, one or more instructions, when executed by one or more processors of network node 110 and / or UE 120 (e.g., directly, or after compilation, conversion, and / or interpretation), may cause the one or more processors, UE 120, and / or network node 110 to perform or direct, for example, Figure 6 The process of 600 Figure 7 The process 700 and / or operations of other processes as described herein. In some examples, executing instructions may include running instructions, converting instructions, compiling instructions, and / or interpreting instructions, etc.

[0058] In some aspects, the UE 120 includes: means for receiving an encoder; means for encoding a CSF using the encoder; and / or means for sending the encoded CSF to the network node 110. In some aspects, the means for the UE 120 to perform operations described herein may include, for example, one or more of the communications manager 140, antenna 252, modem 254, MIMO detector 256, receive processor 258, transmit processor 264, TX MIMO processor 266, controller / processor 280, or memory 282.

[0059] In some aspects, the network node 110 includes: a means for receiving a decoder; a means for receiving the encoded CSF; and / or a means for decoding the CSF using the decoder. In some aspects, the means for the network node 110 to perform operations described herein may include, for example, one or more of the following: the communication manager 150, the transmit processor 220, the TX MIMO processor 230, the modem 232, the antenna 234, the MIMO detector 236, the receive processor 238, the controller / processor 240, the memory 242, or the scheduler 246.

[0060] In some aspects, a first server (e.g., UE server 180) includes: a means for sending encoder training data about an encoder to a set of second servers associated with training a set of decoders and associated with a set of suppliers, wherein the first server is associated with a particular supplier in the set of suppliers; and / or a means for receiving decoder training data associated with the configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data. In some aspects, the means for the first server to perform operations described herein may include, for example, one or more of a communication manager 182, an input component, an output component, a processor, or a memory.

[0061] In some aspects, the second server (e.g., AP server 190) includes: a component for receiving encoder training data about the plurality of encoders from a plurality of first servers associated with the plurality of encoders and a set of suppliers, wherein the second server is associated with a particular supplier in the set of suppliers; and / or a component for sending decoder training data associated with the configuration of the plurality of encoders to the plurality of first servers and at least in part based on receiving the encoder training data. In some aspects, the components for the second server to perform the operations described herein may include, for example, one or more of the communication manager 192, an input component, an output component, a processor, or a memory.

[0062] Although Figure 2 The blocks in the 2000 and 2010 are illustrated as distinct components, but the functionality described above for these blocks may be implemented in a single hardware, software, or combined component or in various combinations of components. For example, the functionality described with respect to the transmit processor 264, the receive processor 258, and / or the TX MIMO processor 266 may be performed by or under the control of the controller / processor 280.

[0063] As indicated above, Figure 2 are provided as examples. Other examples can be found in the Figure 2 Different than described.

[0064] The deployment of a communication system (such as a 5G NR system) can be arranged with various components or components 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 RAN node, a core network node, a network element, a base station or network equipment can be implemented in an aggregated or decomposed architecture. For example, a base station (such as a node B (NB), an evolved NB (eNB), an NR BS, a 5G NB, an access point (AP), a TRP or a cell, etc.) or one or more units (or one or more components) that perform base station functionality can be implemented as an aggregated base station (also referred to as an independent base station or a monolithic base station) or a decomposed base station. "Network entity" or "network node" may refer to a decomposed base station or one or more units of a decomposed base station (such as one or more CUs, one or more DUs, one or more RUs, or a combination thereof).

[0065] An aggregated base station (e.g., an aggregated network node) may be configured to utilize a radio protocol stack physically or logically integrated within a single RAN node (e.g., within a single device or unit). A decomposed base station (e.g., a decomposed network node) may be configured to utilize a protocol stack physically or logically distributed between two or more units (such as one or more CUs, one or more DUs, or one or more RUs). In some examples, a CU may be implemented within a network node, and one or more DUs may be co-located with the CU, or alternatively, may be geographically or virtually distributed across one or more other network nodes. A DU may be implemented to communicate with one or more RUs. Each of a CU, a DU, and a RU may also be implemented as a virtual unit, such as a virtual central unit (VCU), a virtual distributed unit (VDU), or a virtual radio unit (VRU), and the like.

[0066] Base station type operations or network designs may take into account the aggregated nature of base station functionality. For example, a decomposed base station may be utilized in an IAB network, an open radio access network (O-RAN (such as a network configuration initiated by the O-RAN Alliance)), or a virtualized radio access network (vRAN, also known as a cloud radio access network (C-RAN)) to facilitate the scaling of a communication system by separating base station functionality into one or more units that can be deployed separately. A decomposed base station may include functionality implemented across two or more units at various physical locations, as well as functionality implemented virtually for at least one unit, which may enable flexibility in network design. Individual units of a decomposed base station may be configured for wired or wireless communication with at least one other unit of the decomposed base station.

[0067] Figure 33 is a diagram illustrating an example disaggregated base station architecture 300 according to the present disclosure. The disaggregated base station architecture 300 may include a CU 310 that may communicate directly with a core network 320 via a backhaul link, or indirectly with the core network 320 through one or more disaggregated control units, such as a near-RT RIC 325 via an E2 link, or a non-RT RIC 315 associated with a service management and orchestration (SMO) framework 305, or both. The CU 310 may communicate with one or more DUs 330 via respective midhaul links, such as via an F1 interface. Each of the DUs 330 may communicate with one or more RUs 340 via respective fronthaul links. Each of the RUs 340 may communicate with one or more UEs 120 via respective radio frequency (RF) access links. In some implementations, a UE 120 may be served simultaneously by multiple RUs 340.

[0068] Each of the units (including CU 310, DU 330, RU 340) and the near-RT RIC 325, non-RT RIC 315, and SMO framework 305 may include or be coupled to one or more interfaces, the one or more interfaces being configured to receive or send signals, data, or information (collectively referred to as signals) via a wired or wireless transmission medium. Each of the units or an associated processor or controller that provides instructions to one or more communication interfaces of the corresponding unit may be configured to communicate with one or more of the other units via a transmission medium. In some examples, each of the units may include a wired interface and a wireless interface, the wired interface being configured to receive signals or send signals to one or more of the other units via a wired transmission medium, the wireless interface being configured to receive signals or send signals to one or more of the other units via a wired transmission medium, or both.

[0069] In some aspects, CU 310 may host one or more high-level control functions. Such control functions may include radio resource control (RRC) functions, packet data convergence protocol (PDCP) functions, or service data adaptation protocol (SDAP) functions, etc. Each control function may be implemented using an interface that is configured to communicate signals with other control functions hosted by CU 310. CU 310 may be configured to handle user plane functionality (e.g., central unit-user plane (CU-UP) functionality), control plane functionality (e.g., central unit-control plane (CU-CP) functionality), or a combination thereof. In some specific implementations, CU310 may be logically split into one or more CU-UP units and one or more CU-CP units. When implemented in an O-RAN configuration, the CU-UP unit may communicate bidirectionally with the CU-CP unit via an interface (such as an E1 interface). As needed, CU 310 may be implemented to communicate with DU 330 for network control and signaling.

[0070] Each DU 330 may correspond to a logical unit including one or more base station functions for controlling the operation of one or more RUs 340. In some aspects, the DU 330 may host one or more of a radio link control (RLC) layer, a MAC layer, and one or more high physical (PHY) layers, at least in part, according to a functional partition such as that defined by 3GPP. In some aspects, one or more high PHY layers may be implemented by one or more modules for forward error correction (FEC) encoding and decoding, scrambling, modulation and demodulation, etc. In some aspects, the DU 330 may also host one or more low PHY layers, such as one or more modules for fast Fourier transform (FFT), inverse FFT (iFFT), digital beamforming or physical random access channel (PRACH) extraction and filtering, etc. Each layer (which may also be referred to as a module) may be implemented using an interface that is configured to communicate signals with other layers (and modules) hosted by the DU 330 or with control functions hosted by the CU 310.

[0071] Each RU 340 may implement lower layer functionality. In some deployments, the RU 340 controlled by the DU 330 may correspond to a logical node that hosts RF processing functions or low PHY layer functions based on functional split (e.g., functional split defined by 3GPP) (such as lower layer functional split), such as performing FFT, performing iFFT, digital beamforming, or PRACH extraction and filtering, etc. In this architecture, each RU 340 may be operated to handle over-the-air (OTA) communications with one or more UEs 120. In some specific implementations, real-time and non-real-time aspects of control plane and user plane communications with the RU 340 may be controlled by the corresponding DU 330. In some scenarios, this configuration may enable each DU 330 and CU 310 to be implemented in a cloud-based RAN architecture (such as a vRAN architecture).

[0072] The SMO framework 305 may be configured to support RAN deployment and provisioning of non-virtualized network elements and virtualized network elements. For non-virtualized network elements, the SMO framework 305 may be configured to support the deployment of dedicated physical resources for RAN coverage requirements, which may be managed via an operation and maintenance interface (such as an O1 interface). For virtualized network elements, the SMO framework 305 may be configured to interact with a cloud computing platform (such as an open cloud (O-Cloud) platform 390) to perform network element lifecycle management (such as instantiating virtualized network elements) via a cloud computing platform interface (such as an O2 interface). Such virtualized network elements may include, but are not limited to, CU 310, DU 330, RU 340, non-RT RIC 315, and near-RTRIC 325. In some specific implementations, the SMO framework 305 may communicate with hardware aspects of the 4G RAN, such as an open eNB (O-eNB) 311, via the O1 interface. Additionally, in some implementations, the SMO framework 305 can communicate directly with each of the one or more RUs 340 via a corresponding O1 interface. The SMO framework 305 can also include a non-RT RIC 315 configured to support the functionality of the SMO framework 305.

[0073] The non-RT RIC 315 may be configured to include logic functions that enable non-real-time control and optimization of RAN elements and resources, artificial intelligence / machine learning (AI / ML) workflows including model training and updating, or policy-based guidance of applications / features in the near-RT RIC 325. The non-RT RIC 315 may be coupled to or in communication with the near-RT RIC 325 (such as via an A1 interface). The near-RT RIC 325 may be configured to include logic functions that enable near-real-time control and optimization of RAN elements and resources via data collection and actions through an interface (such as via an E2 interface) that connects one or more CUs 310, one or more DUs 330, or both, and the O-eNB with the near-RT RIC 325.

[0074] In some implementations, in order to generate an AI / ML model to be deployed in the near-RT RIC 325, the non-RT RIC 315 may receive parameters or external enrichment information from an external server. Such information may be utilized by the near-RT RIC 325 and may be received from a non-network data source or from a network function at the SMO framework 305 or the non-RT RIC 315. In some examples, the non-RT RIC 315 or the near-RT RIC 325 may be configured to tune RAN behavior or performance. For example, the non-RT RIC 315 may monitor long-term trends and patterns of performance and employ AI / ML models to perform corrective actions through the SMO framework 305 (such as via reconfiguration of the O1 interface) or via the creation of RAN management policies (such as A1 interface policies).

[0075] As indicated above, Figure 3 are provided as examples. Other examples can be found in the Figure 3 Different than described.

[0076] Figure 4 4 is a diagram illustrating an example 400 of physical channels and reference signals in a wireless network according to the present disclosure. Figure 4 As shown, the downlink channel and the downlink reference signal may carry information from the network node 110 to the UE 120 , and the uplink channel and the uplink reference signal may carry information from the UE 120 to the network node 110 .

[0077] As shown, the downlink channel may include a physical downlink control channel (PDCCH) carrying downlink control information (DCI), a physical downlink shared channel (PDSCH) carrying downlink data, or a physical broadcast channel (PBCH) carrying system information, etc. In some aspects, PDSCH communication may be scheduled by PDCCH communication. As further shown, the uplink channel may include a physical uplink control channel (PUCCH) carrying uplink control information (UCI), a physical uplink shared channel (PUSCH) carrying uplink data, or a physical random access channel (PRACH) for initial network access, etc. In some aspects, UE 120 may send confirmation (ACK) or negative confirmation (NACK) feedback (e.g., ACK / NACK feedback or ACK / NACK information) in the UCI on PUCCH and / or PUSCH.

[0078] As further shown in the figure, the downlink reference signal may include a synchronization signal block (SSB), a channel state information (CSI) reference signal (CSI-RS), a demodulation reference signal (DMRS), a positioning reference signal (PRS) or a phase tracking reference signal (PTRS), etc. As also shown in the figure, the uplink reference signal may include a sounding reference signal (SRS), a DMRS, or a PTRS, etc.

[0079] The SSB may carry information used for initial network acquisition and synchronization, such as a primary synchronization signal (PSS), a secondary synchronization signal (SSS), a PBCH, and a PBCH DMRS. The SSB is sometimes referred to as a synchronization signal / PBCH (SS / PBCH) block. In some aspects, the network node 110 may send multiple SSBs on multiple corresponding beams, and the SSBs may be used for beam selection.

[0080] The CSI-RS may carry information for downlink channel estimation (e.g., downlink CSI acquisition), which may be used for scheduling, link adaptation, or beam management, etc. The network node 110 may configure a CSI-RS set for the UE 120, and the UE 120 may measure the configured CSI-RS set. Based at least in part on these measurements, the UE 120 may perform channel estimation and may report channel estimation parameters to the network node 110 (e.g., in a CSI report), such as a channel quality indicator (CQI), a precoding matrix indicator (PMI), a CSI-RS resource indicator (CRI), a layer indicator (LI), a rank indicator (RI), or a reference signal received power (RSRP), among other examples. The network node 110 can use the CSI reports to select transmission parameters for downlink communications to the UE 120, such as the number of transmission layers (e.g., rank), a precoding matrix (e.g., a precoder), a modulation and coding scheme (MCS), or to refine the downlink beams (e.g., using a beam refinement procedure or a beam management procedure), among other examples.

[0081] DMRS can carry information used to estimate the radio channel to demodulate the associated physical channel (e.g., PDCCH, PDSCH, PBCH, PUCCH, or PUSCH). The design and mapping of DMRS can be specific to the physical channel that the DMRS is used to estimate. DMRS is UE-specific, can be beamformed, can be restricted to scheduled resources (e.g., instead of being sent over a wideband), and can be sent only when necessary. As shown, DMRS is used for both downlink and uplink communications.

[0082] PTRS can carry information for compensating for oscillator phase noise. Typically, phase noise increases with the increase of the oscillator carrier frequency. Therefore, PTRS can be used at high carrier frequencies (such as millimeter wave frequencies) to mitigate phase noise. PTRS can be used to track the phase of the local oscillator and to achieve suppression of phase noise and common phase error (CPE). As shown, PTRS is used for both downlink communications (e.g., on PDSCH) and uplink communications (e.g., on PUSCH).

[0083] The PRS may carry information for implementing timing or ranging measurements of the UE 120 based on signals sent by the network node 110 to improve observed time difference of arrival (OTDOA) positioning performance. For example, the PRS may be a pseudo-random quadrature phase shift keying (QPSK) sequence mapped in a diagonal pattern with a frequency offset and a time offset to avoid conflicts with cell-specific reference signals and control channels (e.g., PDCCH). In general, the PRS may be designed to improve the detectability of the UE 120, which may need to detect downlink signals from multiple neighboring network nodes in order to perform OTDOA-based positioning. Therefore, the UE 120 may receive the PRS from multiple cells (e.g., a reference cell and one or more neighboring cells), and may report a reference signal time difference (RSTD) based on the OTDOA measurements associated with the PRS received from the multiple cells. In some aspects, the network node 110 may then calculate the position of the UE 120 based on the RSTD measurements reported by the UE 120.

[0084] The SRS may carry information for uplink channel estimation, which may be used for scheduling, link adaptation, pre-decoder selection, or beam management, etc. The network node 110 may configure one or more SRS resource sets for the UE 120, and the UE 120 may send the SRS on the configured SRS resource sets. The SRS resource sets may have configured usages, such as uplink CSI acquisition, downlink CSI acquisition for reciprocity-based operation, uplink beam management, etc. The network node 110 may measure the SRS, may perform channel estimation based at least in part on these measurements, and may use the SRS measurements to configure communications with the UE 120.

[0085] As indicated above, Figure 4 are provided as examples. Other examples can be found in the Figure 4 Different than described.

[0086] In some communication systems, the UE may perform one or more measurements of the channel condition and may generate a channel state feedback (CSF) to send to a network node, which may include a pre-coding vector. The network node may use the CSF to configure the communication on the channel. For example, the network node may select a pre-decoder, schedule communication, select a modulation and decoding scheme, select a beam, or select a quasi-co-location (QCL) parameter for communication on the channel. However, the CSF may have a relatively high level of overhead on the channel. Therefore, the UE may compress the CSF before sending the CSF to the network node. For example, the UE may use an encoder to process the CSF and generate a compressed CSF, which is sent to the network node. The network node receives the compressed CSF and may use a decoder to process the compressed CSF to recover the CSF. Although some aspects are described herein based on the CSF and communication between the UE and the network node, encoders and decoders may be used for other information conveyed between a transmitter device and a receiver device.

[0087] The servers may train encoders and decoders for compressing and decompressing pre-coded vectors, for example, included in a CSF to reduce overhead associated with CSF reporting. For example, a first server may train an encoder and a decoder, provide a training set for the decoder to a second server, and the second server may use the training set to train the decoder. The procedure may be offline training, which may include using one or more generated training sets processed on one or more servers, rather than the training occurring on the UE and network node where the encoder and decoder will be deployed, respectively.

[0088] Each vendor that manufactures and distributes UEs or access points (APs) (e.g., access points may correspond to network nodes described herein) may use a separate set of servers to participate in offline training. In other words, a first UE vendor and a first AP vendor may use one or more first data sets to train a first encoder and decoder, and a second UE vendor and a second AP vendor may use one or more second data sets to separately train a second encoder and decoder. The UE vendors and AP vendors may be the same or different vendor groups. However, performing separate training for each possible pair of UE vendors and AP vendors may result in separate training for each encoder and decoder, which may result in an excessive number of training sessions and an excessive number of encoder models and decoder models.

[0089] Some aspects described herein enable multi-vendor training for cross-node machine learning. For example, a first server associated with a first vendor may communicate with a second server associated with a second vendor to jointly train one or more encoders and decoders. In this way, machine learning training may use a reduced number of training sessions than would be required in separate training for each pair of vendors, thereby reducing the use of computing resources to perform training while reducing the number of encoder models and decoder models. In addition, combining training from multiple vendors may result in improved encoders and decoders, which may enable, for example, better compression and decompression of CSF, thereby improving communication performance on a network using these encoders and decoders.

[0090] FIG. 5A to FIG. 5D is a diagram illustrating an example 500 associated with offline multi-vendor training for cross-node machine learning according to the present disclosure. Figure 5A As shown, example 500 includes communications between: a set of UE servers (e.g., UE server 180), a set of access point (AP) servers (e.g., AP server 190), a set of UEs (e.g., UE 120), and a set of APs (e.g., network node 110).

[0091] like Figure 5A As further shown by reference numerals 510 and 520, the UE server and the AP server may train an encoder and a decoder, respectively. The encoder may include a neural network or other machine learning model for compressing the CSF. Similarly, the decoder may include a neural network or other machine learning model for decompressing the compressed CSF. For example, a UE server associated with a first vendor may train an encoder, and an AP server associated with the first vendor may train a corresponding decoder. Similarly, a UE server associated with a second vendor may train an encoder, and an AP server associated with the first vendor may train a corresponding decoder.

[0092] In some aspects, the UE server and the AP server may train the encoder and the decoder. For example, the UE server may train the encoder and communicate the encoder training set to the corresponding AP server for training the decoder of the AP server for deployment to the AP associated with the AP server. As shown by reference numerals 530 and 540, the UE server and the AP server may communicate to exchange encoder training data and decoder training data.

[0093] In a first configuration, the UE server and the AP server may perform 1:1 split learning to train the encoder and the decoder using the encoder training data and the decoder training data. Figure 5BAs shown, in 1:1 split learning, the first UE server and the first AP server may be associated with the first UE vendor (Vendor-1) and the first AP vendor (Vendor-A), and the second UE server and the second AP server may be associated with the second UE vendor and the second AP vendor. Although Vendor-1 and Vendor-A may be described as different vendors, Vendor-1 and Vendor-A may be the same vendor, or may be different vendors paired with each other for UE / AP encoder / decoder training. In other words, when Vendor-1 and Vendor-A are the same vendor or when Vendor-1 and Vendor-A are paired, both the first UE and the first AP vendor may be associated with a specific vendor (in this case, the first UE is associated with both Vendor-1 and Vendor-A, and the first AP is associated with both Vendor-A and Vendor-1, which may or may not be the same vendor). The first UE server may provide the baseline truth information and activation information to the first AP server and the second AP server via the first training session and the second training session, and may receive gradient information from the first AP server and the second AP server via the first training session and the second training session. Similarly, the second UE server may provide the reference truth information and activation information to the first AP server and the second AP server via the third training session and the fourth training session, and may receive gradient information from the first AP server and the second AP server via the third training session and the fourth training session. In this case, each UE server and / or each AP server may provide an encoder and / or a decoder to a corresponding group of UEs or APs associated with the UE server and / or the AP server. For example, the first UE server may provide an encoder to a first UE associated with a first vendor, or the second UE server may provide an encoder to a second UE associated with a second vendor, and so on.

[0094] In a second configuration, the UE server and the AP server may perform many-to-many split learning to train the encoder and the decoder using the encoder training data and the decoder training data. Figure 5CAs shown, in many-to-many split learning, the first UE server and the second UE server provide the reference truth information and activation information to the first AP server and the second AP server via the first training session (and receive gradient information from the first AP server and the second AP server). Similarly, the first UE server and the third UE server provide the reference truth information and activation information to the third AP server via the second training session (and receive gradient information from the third AP server). In this case, although servers of multiple vendors may be involved in the training of each encoder or decoder, there is no need for complete overlap (for example, the second UE server associated with the second UE vendor may not have the training session that the third AP server associated with the third AP vendor has).

[0095] In some aspects, in order to allow joint training of the encoder and decoder, each UE server may provide reference truth information to each AP server. The reference truth information may include, for example, information identifying a training set that the UE server can use to train the decoder corresponding to the encoder of the UE server. Additionally or alternatively, the encoder training data and the decoder training data may include gradient information and activation information. For example, the UE server may send the output from the last layer of the encoder to each AP server as activation information. Each AP server may input the received activation information to the corresponding decoder, and may use the reference truth information to determine the loss function value. Each AP server may use the information determined from the loss function to back-propagate the gradient to the input of the decoder. Additionally or alternatively, each AP server may send the gradient to each UE server, and the UE server may back-propagate the gradient to the input of the encoder. In this case, different AP servers may not share decoders with each other (and different UE servers may not share encoders with each other), but may share data used to train the decoder (or encoder).

[0096] In a third configuration, the UE server and the AP server may perform many-to-one federated learning to train the encoder and the decoder using the encoder training data and the decoder training data. Figure 5CAs shown, in many-to-one federated learning, the first UE server and the second UE server may provide gradient information and / or model weights for the first decoder to the first AP server via a first training session, and provide gradient information and / or model weights for the second decoder to the second AP server via a second training session. The first AP server and / or the second AP server may generate a global model and provide the global model to the UE server (as a local copy). In this case, training is performed on the UE server, and the AP server aggregates the training results (e.g., weights or gradients) from the UE server to update the global model. The UE server can train both the UE-specific model (encoder) and the local copy of the global model (decoder).

[0097] As shown in reference numerals 550-1 and 550-2, the UE server and the AP server may deploy encoders and decoders to the UE and the AP, respectively. For example, a UE server of a first vendor may deploy an encoder to a UE of the first UE vendor, and a UE server of a second UE vendor may deploy an encoder to a UE of a second UE vendor. Similarly, decoders may be deployed on a vendor-specific basis. As shown in reference numerals 560, 570, and 580, an inference procedure may occur between the UE and the AP. In the inference procedure, the UE may determine a CSF, compress the CSF using an encoder, and then send the compressed CSF to the AP, and the AP may attempt to decompress the compressed CSF using a decoder corresponding to the encoder.

[0098] As indicated above, FIG. 5A to FIG. 5D are provided as examples. Other examples can be found in relation to FIG. 5A to FIG. 5D Different from what is described.

[0099] Figure 6 600 is a diagram illustrating an example process 600 performed, for example, by a first server according to the present disclosure. Example process 600 is an example in which a first server (eg, UE server 180) performs operations associated with offline multi-vendor training for cross-node machine learning.

[0100] like Figure 6 As shown, in some aspects, process 600 may include sending encoder training data about an encoder to a set of second servers associated with training a set of decoders and associated with a set of vendors, wherein the first server is associated with a particular vendor in the set of vendors (block 610). For example, the first server (e.g., using Figure 8The communication manager 182 and / or sending component 804 depicted in the figure may send encoder training data about the encoder to a set of second servers associated with training a set of decoders and associated with a set of suppliers, wherein the first server is associated with a particular supplier in the set of suppliers, as described above.

[0101] like Figure 6 As further shown, in some aspects, process 600 may include receiving decoder training data associated with the configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data (block 620). Figure 8 The communication manager 182 and / or receiving component 802 depicted in FIG. 8 may receive decoder training data associated with the configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data, as described above.

[0102] Process 600 may include additional aspects, such as any single aspect or any combination of aspects described below and / or in conjunction with one or more other processes described elsewhere herein.

[0103] In a first aspect, the encoder training data comprises at least one of: ground truth data, activation data, or gradient data.

[0104] In a second aspect, either alone or in combination with the first aspect, the decoder training data comprises gradient data.

[0105] In a third aspect, alone or in combination with one or more of the first and second aspects, process 600 includes training an encoder using a training set for the encoder, and sending the encoder training data includes sending information identifying one or more outputs from the training of the encoder.

[0106] In a fourth aspect, alone or in combination with one or more of the first to third aspects, process 600 includes training an encoder using decoder training data.

[0107] In a fifth aspect, alone or in combination with one or more of the first to fourth aspects, process 600 includes deploying an encoder to at least one UE associated with a particular vendor.

[0108] In a sixth aspect, either alone or in combination with one or more of the first to fifth aspects, at least one UE deployed with an encoder is associated with at least one access point associated with a specific vendor deployed with a decoder in the group of decoders.

[0109] In the seventh aspect, alone or in combination with one or more of the first to sixth aspects, other decoders in the group of decoders are deployed to one or more other access points associated with one or more other UEs, and other encoders trained by one or more other first servers are deployed to the one or more other UEs, wherein the one or more other access points and the one or more other UEs are associated with one or more suppliers in the group of suppliers other than a specific supplier.

[0110] In an eighth aspect, either alone or in combination with one or more of aspects one to seven, the group of decoders comprises a single global decoder corresponding to a plurality of encoders and associated with the group of suppliers, wherein the plurality of encoders comprises an encoder corresponding to a UE associated with the particular supplier, and at least one other encoder corresponding to at least one other UE associated with at least one other supplier in the group of suppliers.

[0111] In a ninth aspect, alone or in combination with one or more of the first to eighth aspects, the encoder is associated with channel state feedback compression.

[0112] although Figure 6 An example block diagram of process 600 is shown, but in some aspects, process 600 may include Figure 6 Additional blocks, fewer blocks, different blocks, or blocks arranged in a different manner than those depicted in the process 600. Additionally or alternatively, two or more of the blocks in the process 600 may be performed in parallel.

[0113] Figure 7 700 is a diagram illustrating an example process 700 performed, for example, by a second server according to the present disclosure. Example process 700 is an example in which a second server (eg, AP server 190) performs operations associated with offline multi-vendor training for cross-node machine learning.

[0114] like Figure 7 As shown, in some aspects, process 700 may include: receiving encoder training data for a plurality of encoders from a plurality of first servers associated with a plurality of encoders and a set of suppliers, wherein a second server is associated with a particular supplier in the set of suppliers (block 710). For example, the second server (e.g., using Figure 8 The communication manager 192 and / or receiving component 802 depicted in FIG. 1 may receive encoder training data about the multiple encoders from multiple first servers associated with the multiple encoders and a group of suppliers, wherein a second server is associated with a specific supplier in the group of suppliers, as described above.

[0115] like Figure 7As further shown, in some aspects, process 700 may include: sending decoder training data for a decoder associated with a configuration of the plurality of encoders to the plurality of first servers and based at least in part on receiving the encoder training data (block 720). Figure 8 The communication manager 192 and / or the sending component 804 depicted in FIG. 1 may send decoder training data for the decoders associated with the configuration of the plurality of encoders to the plurality of first servers and based at least in part on receiving the encoder training data, as described above.

[0116] Process 700 may include additional aspects, such as any single aspect or any combination of aspects described below and / or in conjunction with one or more other processes described elsewhere herein.

[0117] In a first aspect, the encoder training data comprises at least one of: ground truth data, activation data, or gradient data.

[0118] In a second aspect, either alone or in combination with the first aspect, the decoder training data comprises gradient data.

[0119] In a third aspect, alone or in combination with one or more of the first and second aspects, process 700 includes training a decoder using encoder training data, and sending the decoder training data includes sending information identifying one or more outputs from the training of the decoder.

[0120] In a fourth aspect, either alone or in combination with one or more of aspects 1 to 3, process 700 comprises determining a loss function associated with a decoder based at least in part on encoder training data, and sending the decoder training data comprises sending one or more outputs from determining the loss function associated with the decoder.

[0121] In a fifth aspect, alone or in combination with one or more of the first to fourth aspects, process 700 includes deploying a decoder to at least one access point associated with a particular vendor.

[0122] In a sixth aspect, either alone or in combination with one or more of the first to fifth aspects, at least one access point deployed with a decoder is associated with at least one UE associated with a specific vendor deployed with an encoder among the plurality of encoders.

[0123] In the seventh aspect, alone or in combination with one or more of the first to sixth aspects, other encoders among the multiple encoders are deployed to one or more other UEs associated with one or more other access points, and one or more other decoders trained by one or more other second servers are deployed to the one or more other access points, wherein the one or more other access points and the one or more other UEs are associated with one or more suppliers other than a specific supplier in the group of suppliers.

[0124] In the eighth aspect, alone or in combination with one or more of the first to seventh aspects, the decoder is a single global decoder corresponding to the multiple encoders and the group of suppliers, wherein the multiple encoders include a first encoder corresponding to the UE and the second supplier, and at least one second encoder corresponding to at least one other UE and corresponding to at least one supplier in the group of suppliers other than the specific supplier.

[0125] In a ninth aspect, alone or in combination with one or more of the first to eighth aspects, the decoder is associated with channel state feedback decompression or reconstruction.

[0126] although Figure 7 An example block diagram of process 700 is shown, but in some aspects, process 700 may include Figure 7 Additional blocks, fewer blocks, different blocks, or blocks arranged in a different manner than those depicted in the process 700. Additionally or alternatively, two or more blocks of the blocks of process 700 may be performed in parallel.

[0127] Figure 8 800 is a diagram of an example device 800 for wireless communication according to the present disclosure. Device 800 may be a server, or a server may include device 800. For example, device 800 may be a first server (e.g., UE server 180), and device 806 may be a second server (e.g., AP server 190) or a UE (e.g., UE 120). Additionally or alternatively, device 800 may be a second server (e.g., AP server 190), and device 806 may be a first server (e.g., UE server 180) or a network node (e.g., network node 110).

[0128] In some aspects, the apparatus 800 includes a receiving component 802 and a sending component 804 that can communicate with each other (e.g., via one or more buses and / or one or more other components). As shown, the apparatus 800 can use the receiving component 802 and the sending component 804 to communicate with another apparatus 806 (such as another server, UE, base station, or another wireless communication device). As further shown, the apparatus 800 may include a communication manager 182 or 192. The communication manager 182 / 192 may include one or more of a machine learning component 808, a deployment component 810, or a determination component 812, etc.

[0129] In some aspects, the apparatus 800 may be configured to perform FIG. 5A to FIG. 5D Additionally or alternatively, the apparatus 800 may be configured to perform one or more processes described herein, such as Figure 6 The process of 600 Figure 7 In some aspects, Figure 8 The apparatus 800 and / or one or more components shown in FIG. 8 may include a combination of Figure 2 Additionally or alternatively, Figure 8 One or more of the components shown in the figure may be combined with Figure 2 Additionally or alternatively, one or more components in a set of components may be implemented at least in part as software stored in a memory. For example, a component (or a portion of a component) may be implemented as instructions or codes stored in a non-transitory computer-readable medium and executable by a controller or processor to perform the function or operation of the component.

[0130] The receiving component 802 may receive communications from the device 806, such as reference signals, control information, data communications, or combinations thereof. The receiving component 802 may provide the received communications to one or more other components of the device 800. In some aspects, the receiving component 802 may perform signal processing (such as filtering, amplification, demodulation, analog-to-digital conversion, demultiplexing, deinterleaving, demapping, equalization, interference cancellation, or decoding, etc.) on the received communications and may provide the processed signals to one or more other components of the device 800. In some aspects, the receiving component 802 may include combining Figure 2 One or more antennas, modems, demodulators, MIMO detectors, receive processors, controllers / processors, memories, or combinations thereof of the described servers.

[0131] The transmitting component 804 may transmit communications, such as reference signals, control information, data communications, or combinations thereof, to the device 806. In some aspects, one or more other components of the device 800 may generate communications and may provide the generated communications to the transmitting component 804 for transmission to the device 806. In some aspects, the transmitting component 804 may perform signal processing (such as filtering, amplification, modulation, digital-to-analog conversion, multiplexing, interleaving, mapping, or encoding, etc.) on the generated communications and may transmit the processed signals to the device 806. In some aspects, the transmitting component 804 may include combining Figure 2 One or more antennas, modems, modulators, transmit MIMO processors, transmit processors, controllers / processors, memories, or combinations thereof of the described servers. In some aspects, the transmit component 804 can be co-located with the receive component 802 in a transceiver.

[0132] The sending component 804 can send encoder training data for the encoder to a set of second servers associated with training a set of decoders and associated with a set of vendors, wherein the first server is associated with a particular vendor in the set of vendors. The receiving component 802 can receive decoder training data associated with the configuration of the encoder from the set of second servers and based at least in part on sending the encoder training data.

[0133] The machine learning training component 808 can train the encoder using a training set for the encoder. The machine learning training component 808 can train the encoder using decoder training data. The deployment component 810 can deploy the encoder to at least one UE associated with a particular vendor.

[0134] The receiving component 802 can receive encoder training data for a plurality of encoders from a plurality of first servers associated with a plurality of encoders and a set of vendors, wherein a second server is associated with a particular vendor in the set of vendors. The sending component 804 can send decoder training data of a decoder associated with a configuration of the plurality of encoders to the plurality of first servers and based at least in part on receiving the encoder training data.

[0135] The machine learning training component 808 can train the decoder using the encoder training data. The determining component 812 can determine a loss function associated with the decoder based at least in part on the encoder training data. The deploying component 810 can deploy the decoder to at least one access point associated with a particular vendor.

[0136] Figure 8 The number and arrangement of components shown in the diagram are provided as examples. In practice, there may be Figure 8 The components shown in the figure may include additional components, fewer components, different components, or components arranged in a different manner. Figure 8Two or more components shown in may be implemented in a single component, or Figure 8 The single component shown in may be implemented as multiple distributed components. Additionally or alternatively, Figure 8 A set (one or more) of components shown in the figure may be described as being executed by Figure 8 One or more functions performed by another set of components shown in FIG.

[0137] Fig. 9 is a diagram of an example apparatus 900 for wireless communication according to the present disclosure. Apparatus 900 may be a UE, or a UE may include apparatus 900. In some aspects, apparatus 900 includes a receiving component 902 and a transmitting component 904 that may communicate with each other (e.g., via one or more buses and / or one or more other components). As shown, apparatus 900 may communicate with another apparatus 906 (such as a UE, a base station, or another wireless communication device) using receiving component 902 and transmitting component 904. As further shown, apparatus 900 may include a communication manager 140. Communication manager 140 may include an encoder component 908, among other things.

[0138] In some aspects, the apparatus 900 may be configured to perform FIG. 5A to FIG. 5D Additionally or alternatively, the apparatus 900 may be configured to perform one or more of the processes described herein. In some aspects, Fig. 9 The apparatus 900 and / or one or more components shown in FIG. 1 may include a combination of Figure 2 Additionally or alternatively, Fig. 9 One or more of the components shown in the figure may be combined with Figure 2 Additionally or alternatively, one or more components in a set of components may be implemented at least in part as software stored in a memory. For example, a component (or a portion of a component) may be implemented as instructions or codes stored in a non-transitory computer-readable medium and executable by a controller or processor to perform the function or operation of the component.

[0139] The receiving component 902 may receive communications from the device 906, such as reference signals, control information, data communications, or combinations thereof. The receiving component 902 may provide the received communications to one or more other components of the device 900. In some aspects, the receiving component 902 may perform signal processing (such as filtering, amplification, demodulation, analog-to-digital conversion, demultiplexing, deinterleaving, demapping, equalization, interference cancellation, or decoding, etc.) on the received communications and may provide the processed signals to one or more other components of the device 900. In some aspects, the receiving component 902 may include combining Figure 2One or more antennas, modems, demodulators, MIMO detectors, receive processors, controllers / processors, memories, or combinations thereof of the described UE.

[0140] The transmitting component 904 may transmit communications, such as reference signals, control information, data communications, or combinations thereof, to the device 906. In some aspects, one or more other components of the device 900 may generate communications and may provide the generated communications to the transmitting component 904 for transmission to the device 906. In some aspects, the transmitting component 904 may perform signal processing (such as filtering, amplification, modulation, digital-to-analog conversion, multiplexing, interleaving, mapping, or encoding, etc.) on the generated communications and may transmit the processed signals to the device 906. In some aspects, the transmitting component 904 may include combining Figure 2 One or more antennas, modems, modulators, transmit MIMO processors, transmit processors, controllers / processors, memories, or combinations thereof of the described UE. In some aspects, the transmit component 904 can be co-located with the receive component 902 in a transceiver.

[0141] Receiving component 902 can receive an encoder for use by encoder component 908. Receiving component 902 can receive one or more signals used to generate a CSF. Encoder component 908 can encode the CSF for transmission to device 906. Transmitting component 904 can transmit the encoded CSF to device 906.

[0142] Fig. 9 The number and arrangement of components shown in the diagram are provided as examples. In practice, there may be Fig. 9 The components shown in the figure may include additional components, fewer components, different components, or components arranged in a different manner. Fig. 9 Two or more components shown in may be implemented in a single component, or Fig. 9 The single component shown in may be implemented as multiple distributed components. Additionally or alternatively, Fig. 9 A set (one or more) of components shown in the figure may be described as being executed by Fig. 9 One or more functions performed by another set of components shown in FIG.

[0143] Fig.101 is a diagram of an example apparatus 1000 for wireless communication according to the present disclosure. Apparatus 1000 may be a network node, or a network node may include apparatus 1000. In some aspects, apparatus 1000 includes a receiving component 1002 and a transmitting component 1004 that may communicate with each other (e.g., via one or more buses and / or one or more other components). As shown, apparatus 1000 may communicate with another apparatus 1006 (such as a UE, a base station, or another wireless communication device) using receiving component 1002 and transmitting component 1004. As further shown, apparatus 1000 may include a communication manager 150. Communication manager 150 may include a decoder component 1008, among other things.

[0144] In some aspects, the apparatus 1000 may be configured to perform FIG. 5A to FIG. 5D Additionally or alternatively, the apparatus 1000 may be configured to perform one or more of the processes described herein. In some aspects, Fig.10 The apparatus 1000 and / or one or more components shown in FIG. 1 may include a combination of Figure 2 Additionally or alternatively, Fig.10 One or more of the components shown in the figure may be combined with Figure 2 Additionally or alternatively, one or more components in a set of components may be implemented at least in part as software stored in a memory. For example, a component (or a portion of a component) may be implemented as instructions or codes stored in a non-transitory computer-readable medium and executable by a controller or processor to perform the function or operation of the component.

[0145] The receiving component 1002 may receive communications, such as reference signals, control information, data communications, or combinations thereof, from the device 1006. The receiving component 1002 may provide the received communications to one or more other components of the device 1000. In some aspects, the receiving component 1002 may perform signal processing (such as filtering, amplification, demodulation, analog-to-digital conversion, demultiplexing, deinterleaving, demapping, equalization, interference cancellation, or decoding, etc.) on the received communications and may provide the processed signals to one or more other components of the device 1000. In some aspects, the receiving component 1002 may include combining Figure 2 One or more antennas, modems, demodulators, MIMO detectors, receive processors, controllers / processors, memories, or combinations thereof, of the described network nodes.

[0146] The transmitting component 1004 may transmit communications, such as reference signals, control information, data communications, or combinations thereof, to the device 1006. In some aspects, one or more other components of the device 1000 may generate communications and may provide the generated communications to the transmitting component 1004 for transmission to the device 1006. In some aspects, the transmitting component 1004 may perform signal processing (such as filtering, amplification, modulation, digital-to-analog conversion, multiplexing, interleaving, mapping, or encoding, etc.) on the generated communications and may transmit the processed signals to the device 1006. In some aspects, the transmitting component 1004 may include combining Figure 2 One or more antennas, modems, modulators, transmit MIMO processors, transmit processors, controllers / processors, memories, or combinations thereof of the described network nodes. In some aspects, the transmit component 1004 can be co-located with the receive component 1002 in a transceiver.

[0147] The receiving component 1002 can receive a decoder for use by the decoder component 1008. The receiving component 1002 can receive an encoded CSF from the device 1006. The decoder component 1008 can decode the CSF. The sending component 1004 can send one or more communications using communication and / or communication configuration information associated with the CSF to set the communication configuration of the device 1006 according to the CSF.

[0148] Fig.10 The number and arrangement of components shown in the diagram are provided as examples. In practice, there may be Fig.10 The components shown in the figure may include additional components, fewer components, different components, or components arranged in a different manner. Fig.10 Two or more components shown in may be implemented in a single component, or Fig.10 The single component shown in may be implemented as multiple distributed components. Additionally or alternatively, Fig.10 A set (one or more) of components shown in the figure may be described as being executed by Fig.10 One or more functions performed by another set of components shown in FIG.

[0149] Fig.11 1 is a diagram illustrating an example 1100 of training and using a machine learning model in conjunction with offline multi-vendor training for cross-node machine learning. A machine learning system may be used to perform the machine learning model training and use described herein. The machine learning system may include, or may be included in, a computing device, a server, a cloud computing environment, etc., such as a first server (e.g., UE server 180) or a second server (e.g., AP server 190) described in more detail elsewhere herein.

[0150] As shown by reference numeral 1105, the observation set can be used to train the machine learning model. The observation set can be obtained from training data (e.g., historical data) such as data collected during one or more processes described herein. In some implementations, the machine learning system can receive the observation set (e.g., as input) from the UE 120 or the network node 110, as described elsewhere herein.

[0151] As shown by reference numeral 1110, the observation set includes a feature set. The feature set may include a variable set, and the variables may be referred to as features. A particular observation may include a set of variable values ​​(or feature values) corresponding to the variable set. In some implementations, the machine learning system may determine the variables of the observation set and / or the variable values ​​of the particular observation based on input received from the UE 120 or the network node 110. For example, the machine learning system may identify the feature set (e.g., one or more features and / or feature values) by extracting the feature set from structured data, by performing natural language processing to extract the feature set from unstructured data, and / or by receiving input from an operator.

[0152] As an example, a feature set for an observation set may include CSF as a first feature, codeword as a second feature, channel measurement as a third feature, etc. As shown, for a first observation, the first feature may have a value CSF1, the second feature may have a value CW1, the third feature may have a value CM1, etc. These features and feature values ​​are provided as examples and may be different in other examples.

[0153] As indicated by reference numeral 1115, a set of observations can be associated with a target variable. A target variable can represent a variable having a numeric value, can represent a variable having a numeric value that falls within a range of values ​​or has some discrete possible values, can represent a variable that can be selected from one of a plurality of options (e.g., one of a plurality of classes, classifications, or labels), and / or can represent a variable having a Boolean value. A target variable can be associated with a target variable value, and the target variable value can be specific to an observation. In example 1100, the target variable is an encoded CSF, which has a value of E-CSF1 for the first observation.

[0154] The feature sets and target variables described above are provided as examples, and other examples may be different from those described above. For example, for a target variable of decoded CSF, the feature set may include encoded CSF.

[0155] The target variable may represent the value that the machine learning model is being trained to predict, and the feature set may represent the variables that are input to the trained machine learning model to predict the value of the target variable. The observation set may include the target variable values ​​so that the machine learning model may be trained to recognize patterns in the feature set that lead to the target variable values. A machine learning model that is trained to predict the value of the target variable may be referred to as a supervised learning model.

[0156] In some implementations, the machine learning model may be trained on an observation set that does not include a target variable. This may be referred to as an unsupervised learning model. In this case, the machine learning model may learn patterns from the observation set without labeling or supervision, and may provide outputs indicative of such patterns, such as by identifying groups of related items within the observation set using clustering and / or association.

[0157] As shown by reference numeral 1120, the machine learning system can train a machine learning model using the observation set and using one or more machine learning algorithms, such as a regression algorithm, a decision tree algorithm, a neural network algorithm, a k-nearest neighbor algorithm, a support vector machine algorithm, etc. After training, the machine learning system can store the machine learning model as a trained machine learning model 1125 for use in analyzing new observations.

[0158] As an example, the machine learning system may obtain training data for an observation set from multiple second servers associated with multiple vendors, and may train an encoder for a specific vendor for use by network nodes 110 associated with the multiple vendors based at least in part on, for example, gradient data output from a trained decoder model.

[0159] As shown at reference numeral 1130, the machine learning system may apply the trained machine learning model 1125 to the new observation, such as by receiving the new observation and inputting the new observation into the trained machine learning model 1125. As shown, as an example, the new observation may include a CSF N As the first feature, CW N As the second feature, CM N as a third feature, and so on. The machine learning system may apply the trained machine learning model 1125 to the new observation to generate an output (e.g., a result). The type of output may depend on the type of machine learning model and / or the type of machine learning task being performed. For example, the output may include a predicted value of a target variable, such as when supervised learning is employed. Additionally or alternatively, the output may include information identifying a cluster to which the new observation belongs and / or information indicating a degree of similarity between the new observation and one or more other observations, such as when unsupervised learning is employed.

[0160] As an example, the trained machine learning model 1125 may predict the value of the target variable E-CSF for a newly observed coded CSF N , as shown by reference number 1135. Based on the prediction, the machine learning system may provide a first recommendation, may provide an output for determining the first recommendation, may perform a first automated action, and / or may cause the first automated action to be performed (e.g., by commanding another device to perform the automated action), etc.

[0161] In some implementations, the trained machine learning model 1125 can be retrained using feedback information. For example, feedback can be provided to the machine learning model. The feedback can be associated with actions performed based on recommendations provided by the trained machine learning model 1125 and / or automated actions performed or caused by the trained machine learning model 1125. In other words, the recommendations and / or actions output by the trained machine learning model 1125 can be used as input to retrain the machine learning model (e.g., a feedback loop can be used to train and / or update the machine learning model). For example, the feedback information can include the output of a trained decoder model (when training an encoder model) (and the feedback information can include the output of a trained decoder model when training a decoder model).

[0162] In this way, a machine learning system can apply a rigorous and automated process to generate an encoder or decoder. The machine learning system can enable recognition and / or identification of tens, hundreds, thousands, or millions of features and / or feature values ​​for tens, hundreds, thousands, or millions of observations across multiple vendors, thereby increasing accuracy and consistency relative to obtaining observations from a single vendor and reducing latency associated with generating an encoder or decoder.

[0163] As indicated above, Fig.11 are provided as examples. Other examples can be combined with Fig.11 Different than described.

[0164] The following provides an overview of some aspects of the disclosure:

[0165] Aspect 1: A wireless communication method performed by a first server, the method comprising: sending encoder training data about an encoder to a group of second servers associated with training a group of decoders and associated with a group of suppliers, wherein the first server is associated with a specific supplier in the group of suppliers; and receiving decoder training data associated with the configuration of the encoder from the group of second servers and at least in part based on sending the encoder training data.

[0166] Aspect 2: The method according to aspect 1, wherein the encoder training data comprises at least one of the following: reference truth data, activation data or gradient data.

[0167] Aspect 3: The method according to any one of aspects 1 to 2, wherein the decoder training data comprises gradient data.

[0168] Aspect 4: According to the method described in any one of Aspects 1 to 3, the method further includes: training the encoder using a training set for the encoder; and wherein sending the encoder training data includes: sending information identifying one or more outputs from the training of the encoder.

[0169] Aspect 5: According to the method according to any one of aspects 1 to 4, the method further comprises: using the decoder training data to train the encoder.

[0170] Aspect 6: The method according to any one of aspects 1 to 5, further comprising: deploying the encoder to at least one user equipment (UE) associated with the specific vendor.

[0171] Aspect 7: The method according to aspect 6, wherein the at least one UE deployed with the encoder is associated with at least one access point associated with the specific vendor deployed with a decoder from the set of decoders.

[0172] Aspect 8: A method according to Aspect 7, wherein other decoders in the group of decoders are deployed to one or more other access points associated with one or more other UEs, and other encoders trained by one or more other first servers are deployed to the one or more other UEs, wherein the one or more other access points and the one or more other UEs are associated with one or more suppliers in the group of suppliers other than the specific supplier.

[0173] Aspect 9: A method according to Aspect 6, wherein the set of decoders includes a single global decoder corresponding to multiple encoders and associated with the set of suppliers, wherein the multiple encoders include the encoder corresponding to the UE associated with the specific supplier, and at least one other encoder corresponding to at least one other UE associated with at least one other supplier in the set of suppliers.

[0174] Aspect 10: The method according to any one of aspects 1 to 9, wherein the encoder is associated with channel state feedback compression.

[0175] Aspect 11: A wireless communication method performed by a second server, the method comprising: receiving encoder training data about the multiple encoders from multiple first servers associated with the multiple encoders and a group of suppliers, wherein the second server is associated with a specific supplier in the group of suppliers; and sending decoder training data of a decoder associated with the configuration of the multiple encoders to the multiple first servers and at least partially based on receiving the encoder training data.

[0176] Aspect 12: The method according to aspect 11, wherein the encoder training data comprises at least one of the following: ground truth data, activation data or gradient data.

[0177] Aspect 13: A method according to any one of Aspects 11 to 12, wherein the decoder training data comprises gradient data.

[0178] Aspect 14: A method according to any one of Aspects 11 to 13, the method further comprising: using the encoder training data to train the decoder; and wherein sending the decoder training data comprises: sending information identifying one or more outputs from the training of the decoder.

[0179] Aspect 15: According to the method described in any one of Aspects 11 to 14, the method further includes: determining a loss function associated with the decoder based at least in part on the encoder training data; and wherein sending the decoder training data includes: sending one or more outputs from determining the loss function associated with the decoder.

[0180] Aspect 16: The method according to any one of aspects 11 to 15, further comprising: deploying the decoder to at least one access point associated with the specific vendor.

[0181] Aspect 17: The method according to aspect 16, wherein the at least one access point where the decoder is deployed is associated with at least one user equipment (UE) associated with the specific vendor where an encoder of the plurality of encoders is deployed.

[0182] Aspect 18: A method according to Aspect 17, wherein other encoders among the multiple encoders are deployed to one or more other UEs associated with one or more other access points, and one or more other decoders trained by one or more other second servers are deployed to the one or more other access points, wherein the one or more other access points and the one or more other UEs are associated with one or more suppliers among the group of suppliers other than the specific supplier.

[0183] Aspect 19: A method according to Aspect 16, wherein the decoder is a single global decoder corresponding to the multiple encoders and the group of suppliers, wherein the multiple encoders include a first encoder corresponding to a user equipment (UE) and the second supplier, and at least one second encoder corresponding to at least one other UE and corresponding to at least one supplier in the group of suppliers other than the specific supplier.

[0184] Aspect 20: The method according to any one of aspects 11 to 19, wherein the decoder is associated with channel state feedback decompression or reconstruction.

[0185] Aspect 21: An apparatus for performing wireless communications at a device, the apparatus comprising: a processor; a memory coupled to the processor; and instructions stored in the memory and executable by the processor to cause the apparatus to perform one or more of the methods described in Aspects 1 to 10.

[0186] Aspect 22: A device for wireless communication, the device comprising: a memory and one or more processors coupled to the memory, the one or more processors configured to execute the method according to one or more of aspects 1 to 10.

[0187] Aspect 23: An apparatus for wireless communication, the apparatus comprising: at least one means for performing the method according to one or more of aspects 1 to 10.

[0188] Aspect 24: A non-transitory computer-readable medium storing code for wireless communication, the code comprising instructions executable by a processor to perform the method according to one or more of aspects 1 to 10.

[0189] Aspect 25: A non-transitory computer-readable medium storing an instruction set for wireless communication, the instruction set comprising one or more instructions, which, when executed by one or more processors of a device, causes the device to perform one or more of the methods described in aspects 1 to 10.

[0190] Aspect 26: An apparatus for performing wireless communications at a device, the apparatus comprising: a processor; a memory coupled to the processor; and instructions stored in the memory and executable by the processor to cause the apparatus to perform one or more of the methods described in Aspects 11 to 20.

[0191] Aspect 27: A device for wireless communication, the device comprising: a memory and one or more processors coupled to the memory, the one or more processors configured to perform the method according to one or more of aspects 11 to 20.

[0192] Aspect 28: An apparatus for wireless communication, the apparatus comprising: at least one means for performing the method according to one or more of aspects 11 to 20.

[0193] Aspect 29: A non-transitory computer-readable medium storing code for wireless communication, the code comprising instructions executable by a processor to perform the method according to one or more of aspects 11 to 20.

[0194] Aspect 25: A non-transitory computer-readable medium storing an instruction set for wireless communication, the instruction set comprising one or more instructions, which, when executed by one or more processors of a device, causes the device to perform one or more of the methods described in aspects 11 to 20.

[0195] The foregoing disclosure provides illustration and description, but is not intended to be exhaustive or to limit the aspects to the precise forms disclosed. Modifications and variations may be made in light of the above disclosure or may be acquired from practice of these aspects.

[0196] As used herein, the term "component" is intended to be broadly interpreted as a combination of hardware and / or hardware and software. Whether referred to as software, firmware, middleware, microcode, hardware description language or other names, "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, executable files, threads of execution, procedures and / or functions, etc. As used herein, a "processor" is implemented in a combination of hardware and / or hardware and software. It will be apparent that the systems and / or methods described herein can be implemented by a combination of hardware and / or hardware and software in different forms. The actual dedicated control hardware or software code used to implement these systems and / or methods does not limit various aspects. Therefore, there is no reference to a specific software code herein to describe the operation and behavior of the system and / or method, because those skilled in the art will understand that software and hardware can be designed to implement the system and / or method based at least in part on the description herein.

[0197] As used herein, "satisfying a threshold" may refer to a value being greater than a threshold, greater than or equal to a threshold, less than a threshold, less than or equal to a threshold, equal to a threshold, not equal to a threshold, etc., depending on the context.

[0198] Although the specific combination of features is described in the claims and / or disclosed in the specification, these combinations are not intended to limit the disclosure of various aspects. Many of these features can be combined in a manner that is not specifically described in the claims and / or is not disclosed in the specification. The disclosure of various aspects includes each dependent claim combined with each other claim in the claim set. As used herein, the phrase "at least one of" the list of items refers to any combination of these items (it includes a single member). As an example, "at least one of a, b or c" is intended to cover a, b, c, a+b, a+c, b+c and a+b+c, and any combination with multiple identical elements (for example, a+a, a+a+a, a+a+b, a+a+c, a+b+b, a+c+c, b+b, b+b+b, b+b+c, c+c and c+c+c, or any other ordering of a, b and c).

[0199] Any element, action or instruction used herein should not be interpreted as critical or necessary unless clearly stated. In addition, as used herein, the article "one" is intended to include one or more items, and can be used interchangeably with "one or more". In addition, as used herein, the article "said" is intended to include one or more items connected to the article "said", and can be used interchangeably with "one or more". In addition, as used herein, the terms "group" and "group" are intended to include one or more items, and can be used interchangeably with "one or more". If only one item is intended to be referred to, the phrase "only one" or similar terms will be used. Moreover, as used herein, the terms "have", "have" etc. are intended to be open terms, which do not limit the elements they modify (for example, "elements with" A can also have B). In addition, the phrase "based on" is intended to represent "based at least in part", unless otherwise explicitly stated. Furthermore, as used herein, the term "or" when used in a series is intended to be open-ended and used interchangeably with "and / or" unless expressly stated otherwise (eg, if used in conjunction with "either" or "only one of").

Claims

1. A first server, comprising: Memory; and one or more processors, the one or more processors coupled to the memory, the one or more processors configured to: sending encoder training data about an encoder to a set of second servers associated with training a set of decoders and associated with a set of vendors, wherein the first server is associated with a particular vendor in the set of vendors; as well as Decoder training data associated with a configuration of the encoder is received from the set of second servers and based at least in part on sending the encoder training data.

2. The first server of claim 1, wherein the encoder training data comprises at least one of: Benchmark truth data, Activation data or Gradient data.

3. The first server of claim 1, wherein the decoder training data comprises gradient data.

4. The first server of claim 1, wherein the one or more processors are further configured to: training the encoder using a training set for the encoder; and Wherein, in order to send the encoder training data, the one or more processors are configured to: Information identifying one or more outputs from the training of the encoder is sent.

5. The first server of claim 1, wherein the one or more processors are further configured to: The encoder is trained using the decoder training data.

6. The first server of claim 1, wherein the one or more processors are further configured to: The encoder is deployed to at least one user equipment (UE) associated with the particular vendor. 7 . The first server of claim 6 , wherein the at least one UE on which the encoder is deployed is associated with at least one access point associated with the specific vendor on which a decoder from the set of decoders is deployed.

8. The first server according to claim 7, wherein other decoders in the group of decoders are deployed to one or more other access points associated with one or more other UEs, and other encoders trained by one or more other first servers are deployed to the one or more other UEs, wherein the one or more other access points and the one or more other UEs are associated with one or more suppliers in the group of suppliers other than the specific supplier.

9. A first server according to claim 6, wherein the set of decoders includes a single global decoder corresponding to multiple encoders and associated with the set of suppliers, wherein the multiple encoders include the encoder corresponding to the UE associated with the specific supplier, and at least one other encoder corresponding to at least one other UE associated with at least one other supplier in the set of suppliers.

10. The first server of claim 1, wherein the encoder is associated with channel state feedback compression.

11. A second server, the second server comprising: Memory; and one or more processors, the one or more processors coupled to the memory, the one or more processors configured to: receiving encoder training data for a plurality of encoders from a plurality of first servers associated with a plurality of encoders and a set of vendors, wherein the second server is associated with a particular vendor from the set of vendors; as well as Decoder training data for a decoder associated with a configuration of the plurality of encoders is sent to the plurality of first servers and based at least in part on receiving the encoder training data.

12. The second server according to claim 11, wherein the encoder training data comprises at least one of the following: Benchmark truth data, Activation data or Gradient data.

13. The second server of claim 11, wherein the decoder training data comprises gradient data.

14. The second server according to claim 11, wherein the one or more processors are further configured to: training the decoder using the encoder training data; and Wherein, in order to send the decoder training data, the one or more processors are configured to: Information identifying one or more outputs from the training of the decoder is transmitted.

15. The second server according to claim 11, wherein the one or more processors are further configured to: determining a loss function associated with the decoder based at least in part on the encoder training data; and Wherein, in order to send the decoder training data, the one or more processors are configured to: One or more outputs from determining the loss function associated with the decoder are transmitted.

16. The second server according to claim 11, wherein the one or more processors are further configured to: The decoder is deployed to at least one access point associated with the particular vendor.

17. The second server of claim 16, wherein the at least one access point where the decoder is deployed is associated with at least one user equipment (UE) associated with the specific vendor where an encoder of the plurality of encoders is deployed.

18. The second server of claim 17, wherein other encoders among the plurality of encoders are deployed to one or more other UEs associated with one or more other access points, and one or more other decoders trained by one or more other second servers are deployed to the one or more other access points, wherein the one or more other access points and the one or more other UEs are associated with one or more suppliers among the group of suppliers other than the specific supplier.

19. A second server according to claim 16, wherein the decoder is a single global decoder corresponding to the multiple encoders and the group of suppliers, wherein the multiple encoders include a first encoder corresponding to a user equipment (UE) and the specific supplier, and at least one second encoder corresponding to at least one other UE and corresponding to at least one supplier in the group of suppliers other than the specific supplier.

20. The second server of claim 11, wherein the decoder is associated with channel state feedback decompression or reconstruction.

21. A method performed by a first server, the method comprising: sending encoder training data about an encoder to a set of second servers associated with training a set of decoders and associated with a set of vendors, wherein the first server is associated with a particular vendor in the set of vendors; as well as Decoder training data associated with a configuration of the encoder is received from the set of second servers and based at least in part on sending the encoder training data.

22. The method of claim 21, wherein the encoder training data comprises at least one of: Benchmark truth data, Activation data or Gradient data.

23. The method of claim 21, the decoder training data comprising gradient data.

24. The method according to claim 21, further comprising: training the encoder using a training set for the encoder; and Wherein sending the encoder training data comprises: Information identifying one or more outputs from the training of the encoder is sent.

25. The method according to claim 21, further comprising: The encoder is trained using the decoder training data.

26. The method according to claim 21, further comprising: The encoder is deployed to at least one user equipment (UE) associated with the particular vendor.

27. A method performed by a second server, the method comprising: receiving encoder training data for a plurality of encoders from a plurality of first servers associated with a plurality of encoders and a set of vendors, wherein the second server is associated with a particular vendor from the set of vendors; as well as Decoder training data for a decoder associated with a configuration of the plurality of encoders is sent to the plurality of first servers and based at least in part on receiving the encoder training data.

28. The method of claim 27, wherein the encoder training data comprises at least one of: Benchmark truth data, Activation data or Gradient data.

29. The method of claim 27, wherein the decoder training data comprises gradient data.

30. The method of claim 27, further comprising: training the decoder using the encoder training data; and Wherein sending the decoder training data comprises: Information identifying one or more outputs from the training of the decoder is transmitted.

Images