Neural network or layer configuration indicators for channel state information schemes

Abstract

Methods, systems, and apparatus for wireless communications are described. A user equipment (UE) may train a first set of layers of a neural network based on a channel estimate using a set of resources. The UE may generate a set of weights for the first set of layers of the neural network based on the training. The UE may receive an indication of an association between a first set of signals and a second set of signals from a first network entity based on the first set of layers of the neural network. The UE may receive a second set of signals from a second network entity and process the second set of signals using the set of weights for the first set of layers based on the association between the first set of signals and the second set of signals.

Description

[0001] Cross-references

[0002] This patent application claims priority to Greek patent application No. 20200100495 filed by Manolakos et al. on August 18, 2020, entitled “NEURAL NETWORKOR LAYER CONFIGURATION INDICATOR FOR A CHANNEL STATE INFORMATION SCHEME,” which is assigned to the assignee of this application.

[0003] introduction

[0004] The following relates to wireless communications, and more particularly to reusing trained layers of a neural network to process signals or channels.

[0005] Wireless communication systems are widely deployed to provide various types of communication content, such as voice, video, packet data, messaging, broadcast, and the like. These systems may be able to support communication with multiple users by sharing available system resources (e.g., time, frequency, and power). Examples of such multiple access systems include fourth generation (4G) systems (such as long term evolution (LTE) systems, advanced LTE (LTE-A) systems, or LTE-A Pro systems), and fifth generation (5G) systems, which may be referred to as NR systems. These systems may employ various technologies, such as code division multiple access (CDMA), time division multiple access (TDMA), frequency division multiple access (FDMA), orthogonal frequency division multiple access (OFDMA), or discrete Fourier transform spread orthogonal frequency division multiplexing (DFT-S-OFDM). A wireless multiple access communication system may include one or more base stations or one or more network access nodes, each of which simultaneously supports communication for multiple communication devices, which may be further referred to as user equipment (UE).

[0006] Overview

[0007] A method for wireless communication at a UE is described. The method may include: training a first set of layers of a neural network based on a channel estimate using a set of resources; generating a set of weights for the first set of layers of the neural network based on the training; receiving an indication of an association between a first set of one or more signals and a second set of one or more signals from a first network entity based on the first set of layers of the neural network; receiving the second set of one or more signals from a second network entity; and processing the second set of one or more signals using the set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals.

[0008] An apparatus for wireless communication at a UE is described. The apparatus may include a processor, a memory in electronic communication with the processor, and instructions stored in the memory. The instructions may be executable by the processor to cause the apparatus to: train a first set of layers of a neural network based on channel estimates using a set of resources; generate a set of weights for the first set of layers of the neural network based on the training; receive an indication of an association between a first set of one or more signals and a second set of one or more signals from a first network entity based on the first set of layers of the neural network; receive the second set of one or more signals from a second network entity; and process the second set of one or more signals using the set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals.

[0009] Another apparatus for wireless communication at a UE is described. The apparatus may include: means for training a first set of layers of a neural network based on a channel estimate using a set of resources; means for generating a set of weights for the first set of layers of the neural network based on the training; means for receiving an indication of an association between a first set of one or more signals and a second set of one or more signals from a first network entity based on the first set of layers of the neural network; means for receiving the second set of one or more signals from a second network entity; and means for processing the second set of one or more signals using the set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals.

[0010] A non-transitory computer-readable medium storing code for wireless communication at a UE is described. The code may include instructions executable by a processor to: train a first set of layers of a neural network based on channel estimation using a set of resources; generate a set of weights for the first set of layers of the neural network based on the training; receive an indication of an association between a first set of one or more signals and a second set of one or more signals from a first network entity based on the first set of layers of the neural network; receive the second set of one or more signals from a second network entity; and process the second set of one or more signals using the set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals.

[0011] Some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein may further include operations, features, means, or instructions for transmitting to a first network entity an indication of the number of layer states that the UE may be able to store, track, train, process, or any combination thereof for one or more of a component carrier, a frequency band, a frequency band combination, or the like.

[0012] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, processing the second group of one or more signals may include operations, features, apparatus, or instructions for decoding the second group of one or more signals using a set of weights for a first layer set based on an association between the first group of one or more signals and the second group of one or more signals.

[0013] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, processing the second group of one or more signals may include operations, features, apparatuses, or instructions for the following actions: demodulating the second group of one or more signals using a set of weights for a first layer set based on an association between the first group of one or more signals and the second group of one or more signals.

[0014] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, processing the second group of one or more signals may include operations, features, apparatus, or instructions for estimating a downlink channel from the second group of one or more signals using a set of weights for a first layer set based on an association between the first group of one or more signals and the second group of one or more signals.

[0015] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, processing the second set of one or more signals may include operations, features, apparatuses, or instructions for compressing the second set of one or more signals using a set of weights for a first set of layers based on an association between the first set of one or more signals and the second set of one or more signals, and training the first set of layers of a neural network, the second set of layers of a neural network, or both based on the compressed second set of one or more signals.

[0016] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, processing the second set of one or more signals may include operations, features, means, or instructions for training a set of layers of a second neural network using a set of weights for a first set of layers of the neural network based on correlations between the first set of one or more signals and the second set of one or more signals.

[0017] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the first group of one or more signals includes one or more of a channel state information (CSI) reference signal (CSI-RS), a synchronization signal block (SSB), a positioning reference signal (PRS), a demodulation reference signal (DMRS), a tracking signal, a data channel, or a control channel.

[0018] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the second set of one or more signals includes one or more of a CSI-RS, an SSB, a PRS, a DMRS, a tracking signal, a data channel, or a control channel.

[0019] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the indication of the association includes a source identifier and a target identifier.

[0020] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the source identifier includes an identifier of the neural network.

[0021] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the target identifier includes an identifier of the second set of one or more signals, a procedure for the second set of one or more signals, or an identifier of the second neural network, or any combination.

[0022] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the source identifier includes an identifier of a signal or protocol corresponding to at least a first set of one or more signals.

[0023] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the source identifier includes an identifier of at least a first set of layers of the neural network.

[0024] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the indication of the association may be received via a higher layer signal, a medium access control (MAC) control element (CE), downlink control information, or both.

[0025] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the first set of one or more signals corresponds to a first component carrier and the second set of one or more signals corresponds to a second component carrier.

[0026] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the first set of one or more signals corresponds to a first frequency band and the second set of one or more signals corresponds to a second frequency band.

[0027] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the first set of one or more signals corresponds to a first frequency band combination, and the second set of one or more signals corresponds to a second frequency band combination.

[0028] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the first set of one or more signals corresponds to a first frequency range and the second set of one or more signals corresponds to a second frequency range.

[0029] In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, the first set of layers of the neural network includes one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

[0030] Some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein may further include operations, features, means, or instructions for receiving a third set of one or more signals from a second network entity, and processing the third set of one or more signals using the set of weights.

[0031] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the second network entity may be another UE, a base station, a transmission and reception point, a server, the first network entity, or any combination thereof.

[0032] A method for wireless communication at a network entity is described. The method may include receiving an indication of a trained set of layers of a neural network from a UE based on a channel estimate on a set of resources; identifying a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers; transmitting to the UE an indication of an association between a first set of one or more signals and a second set of one or more signals based on the trained set of layers of the neural network; and transmitting the second set of one or more signals to the UE.

[0033] An apparatus for wireless communication at a network entity is described. The apparatus may include a processor, a memory in electronic communication with the processor, and instructions stored in the memory. The instructions may be executable by the processor to cause the apparatus to: receive an indication of a trained set of layers of a neural network from a UE based on a channel estimate on a set of resources; identify a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers; transmit to the UE an indication of an association between a first set of one or more signals and a second set of one or more signals based on the trained set of layers of the neural network; and transmit the second set of one or more signals to the UE.

[0034] Another apparatus for wireless communication at a network entity is described. The apparatus may include: means for receiving an indication of a trained set of layers of a neural network from a UE based on a channel estimate on a set of resources; means for identifying a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers; means for transmitting to the UE an indication of an association between a first set of one or more signals and a second set of one or more signals based on the trained set of layers of the neural network; and means for transmitting the second set of one or more signals to the UE.

[0035] A non-transitory computer-readable medium storing code for wireless communication at a network entity is described. The code may include instructions executable by a processor to: receive an indication of a trained set of layers of a neural network from a UE based on a channel estimate on a set of resources; identify a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers; transmit to the UE an indication of an association between a first set of one or more signals and a second set of one or more signals based on the trained set of layers of the neural network; and transmit the second set of one or more signals to the UE.

[0036] Some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein may further include operations, features, apparatus, or instructions for receiving from a UE an indication of the number of layer states that the UE may be able to store, track, train, process, or any combination thereof for one or more of a component carrier, a frequency band, a frequency band combination, or any combination thereof.

[0037] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the first set of one or more signals includes one or more of a CSI-RS, an SSB, or a PRS.

[0038] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the second set of one or more signals includes one or more of a CSI-RS, an SSB, or a PRS.

[0039] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the indication of the association includes a source identifier and a target identifier.

[0040] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the source identifier includes an identifier of the neural network.

[0041] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the source identifier includes an identifier of at least a set of trained layers of the neural network.

[0042] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the source identifier includes an identifier of a signal or protocol corresponding to at least a first set of one or more signals.

[0043] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the target identifier includes an identifier of the second set of one or more signals, a procedure for the second set of one or more signals, or both.

[0044] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, transmitting an indication of an association may include operations, features, apparatuses, or instructions for transmitting an indication of an association via a MAC CE, downlink control information, or both.

[0045] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the first set of one or more signals corresponds to a first component carrier and the second set of one or more signals corresponds to a second component carrier.

[0046] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the first set of one or more signals corresponds to a first frequency band and the second set of one or more signals corresponds to a second frequency band.

[0047] In some examples of the methods, apparatus (devices), and non-transitory computer-readable media described herein, the first set of one or more signals corresponds to a first frequency range and the second set of one or more signals corresponds to a second frequency range.

[0048] In some examples of the methods, apparatuses (devices), and non-transitory computer-readable media described herein, the trained set of layers of the neural network includes one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Illustrated are examples of systems for wireless communication supporting neural network or layer configuration indicators for channel state information (CSI) schemes in accordance with aspects of the present disclosure.

[0051] Figure 2 An example of a wireless communication system supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated.

[0052] Figure 3 An example of a CSI reporting scheme supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is illustrated.

[0053] Figure 4 Examples of supporting neural network or layer configuration indicators for CSI schemes according to aspects of the present disclosure are illustrated.

[0054] Figure 5 Examples of supporting neural network or layer configuration indicators for CSI schemes according to aspects of the present disclosure are illustrated.

[0055] Figure 6 Examples of supporting neural network or layer configuration indicators for CSI schemes according to aspects of the present disclosure are illustrated.

[0056] Figure 7 Examples of supporting neural network or layer configuration indicators for CSI schemes according to aspects of the present disclosure are illustrated.

[0057] Figure 8 Examples of supporting neural network or layer configuration indicators for CSI schemes according to aspects of the present disclosure are illustrated.

[0058] Figure 9 Illustrated are example processes for supporting neural network or layer configuration indicators for CSI schemes in accordance with aspects of the present disclosure.

[0059] Figure 10 Illustrated are example processes for supporting neural network or layer configuration indicators for CSI schemes in accordance with aspects of the present disclosure.

[0060] Figure 11 An example of a process flow supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated.

[0061] Figure 12 and 13 A block diagram of a device supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown.

[0062] Figure 14 A block diagram of a communication manager supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown.

[0063] Figure 15 A diagram of a system including devices supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown.

[0064] Figure 16 and 17 A block diagram of a device supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown.

[0065] Figure 18 A block diagram of a communication manager supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown.

[0066] Figure 19 A diagram of a system including devices supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown.

[0067] Figures 20 to 23 A flow chart illustrating a method of supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown.

[0068] Detailed description

[0069] A user equipment (UE) may measure signals or signaling from a base station and transmit measurement reports to the base station. For example, the UE may measure reference signals and transmit measurement reports of the reference signals to assist the base station in managing the network and improving the channel conditions of the served devices. In some cases, the UE may transmit channel state information (CSI) or channel state feedback (CSF) generated based on measuring the signaling from the base station. Some wireless communication systems support multiple types of CSI. For example, a first type of CSI may be used for a beam selection scheme, in which the UE selects the index of each best possible beam and transmits CSI information to report these indexes. A second type of CSI may be a beam combining scheme, in which the UE also calculates the best linear combination coefficients of various beams and reports the beam index, where the coefficients are used to combine them on a subband basis. The wireless communication system may support at least neural network-based CSI or machine learning-based CSI, which may use machine learning techniques to compress and feedback the channel and interference observed at the UE. For neural network-based CSI reporting, the UE may use a resource set to train one or more layers of the neural network or the entire neural network. The UE generates weights or coefficients at each layer and indicates the set of weights to the base station. The base station can recreate the channel based on the weight set and perform efficient channel maintenance.The layer state can correspond to the weight set obtained through training and contained in the layer.

[0070] The wireless communication system described herein may support the reuse of trained neural networks or neural network layers to process other signals or channels. For example, a UE may train a neural network or neural network layer, and the UE may be indicated an association between two types of signals or processes. The UE may receive a neural network or layer configuration indicator (NNCI) for neural network-based CSI reporting indicating the association. The UE may have trained one or more layers of a neural network using a first set of one or more signals, and these trained layers may be used to efficiently process a second set of one or more signals at the UE. If the UE knows the association between the trained layers and the signals or channels that can be processed by reusing the trained layers, the UE may avoid cold start training of the neural network or layer and save processing power and latency.

[0071] The UE may indicate a number of states that the UE may store, track, or train per component carrier, frequency band, or subband. For example, the UE may train a neural network on one CSI-RS from one gNB and use that training to demodulate another reference signal or channel. Some example associations using previously trained neural network states or layer states may include using CSI-RS training for a demodulation reference signal (DMRS), other CSI-RS, or positioning reference signal (PRS), using synchronization signal blocks (SSBs) training for a CSI-RS or PRS, or using PRS for a DMRS. The NNCI may include a source identifier of a signal or procedure used to train the neural network, a source identifier of an already trained neural network, or a source identifier for a specific layer. The NNCI may be configured semi-statically or dynamically. In some cases, the NNCI may be used for cross-component carrier NNCI relationships, cross-frequency range relationships, cross-band relationships, or cross-band combination relationships.

[0072] Various aspects of the present disclosure are initially described in the context of wireless communication systems. Various aspects of the present disclosure are further illustrated and described by and with reference to apparatus diagrams, system diagrams, and flow diagrams relating to neural networks or layer configuration indicators for CSI schemes.

[0073] Figure 1 An example of a wireless communication system 100 that supports a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated. The wireless communication system 100 may include one or more base stations 105, one or more UEs 115, and a core network 130. In some examples, the wireless communication system 100 may be a Long Term Evolution (LTE) network, an Advanced LTE (LTE-A) network, an LTE-A Pro network, or a New Radio (NR) network. In some examples, the wireless communication system 100 may support enhanced broadband communication, ultra-reliable (e.g., mission-critical) communication, low-latency communication, communication with low-cost and low-complexity devices, or any combination thereof.

[0074] Base stations 105 may be dispersed throughout a geographic area to form wireless communication system 100 and may be different forms of devices or devices with different capabilities. Base stations 105 and UEs 115 may communicate wirelessly via one or more communication links 125. Each base station 105 may provide a coverage area 110 over which UEs 115 and base stations 105 may establish one or more communication links 125. Coverage area 110 may be an example of a geographic area over which base stations 105 and UEs 115 may support signal communication according to one or more radio access technologies.

[0075] The UEs 115 may be dispersed throughout the coverage area 110 of the wireless communication system 100, and each UE 115 may be stationary or mobile, or stationary and mobile at different times. The UEs 115 may be different forms of devices or devices with different capabilities. Figure 1 1. The UE 115 described herein may be capable of communicating with various types of devices, such as other UEs 115, base stations 105, or network equipment (e.g., core network nodes, relays, integrated access and backhaul (IAB) nodes, or other network equipment), such as Figure 1 As shown in .

[0076] Each base station 105 can communicate with the core network 130, with each other, or both. For example, the base stations 105 can interface with the core network 130 via one or more backhaul links 120 (e.g., via S1, N2, N3, or other interfaces). The base stations 105 can communicate with each other directly (e.g., directly between the base stations 105), indirectly (e.g., via the core network 130), or both directly and indirectly over the backhaul links 120 (e.g., via X2, Xn, or other interfaces). In some examples, the backhaul links 120 can be or include one or more radio links. The UE 115 can communicate with the core network 130 via a communication link 155.

[0077] One or more of the base stations 105 described herein may include or may be referred to by one of ordinary skill in the art as a base transceiver station, a radio base station, an access point, a radio transceiver, a Node B, an evolved Node B (eNB), a next generation Node B, or a Gigabit Node B (any of which may be referred to as a gNB), a Home Node B, a Home Evolved Node B, or other suitable terminology.

[0078] UE 115 may include or be referred to as a mobile device, a wireless device, a remote device, a handheld device, or a subscriber device, or some other suitable terminology, where "device" may also be referred to as a unit, a station, a terminal, or a client, etc. UE 115 may also include or be referred to as a personal electronic device, such as a cellular phone, a personal digital assistant (PDA), a tablet computer, a laptop computer, or a personal computer. In some examples, UE 115 may include or be referred to as a wireless local loop (WLL) station, an Internet of Things (IoT) device, an Internet of Everything (IoE) device, or a machine type communication (MTC) device, etc., which may be implemented in various objects, such as appliances or vehicles, meters, etc.

[0079] The UE 115 described herein may be capable of communicating with various types of devices, such as other UEs 115, which may sometimes act as relays, as well as base stations 105 and network equipment including macro eNBs or gNBs, small cell eNBs or gNBs, relay base stations, etc. Figure 1 As shown in .

[0080] The UE 115 and the base station 105 may communicate wirelessly with each other via one or more communication links 125 on one or more carriers. The term "carrier" may refer to a set of radio frequency spectrum resources having a defined physical layer structure for supporting the communication link 125. For example, a carrier for the communication link 125 may include a portion of a radio frequency spectrum band (e.g., a bandwidth portion (BWP)) that operates according to one or more physical layer channels for a given radio access technology (e.g., LTE, LTE-A, LTE-A Pro, NR). Each physical layer channel may carry acquisition signaling (e.g., synchronization signals, system information), control signaling for coordinating carrier operation, user data, or other signaling. The wireless communication system 100 may support communication with the UE 115 using carrier aggregation or multi-carrier operation. The UE 115 may be configured with multiple downlink component carriers and one or more uplink component carriers according to a carrier aggregation configuration. Carrier aggregation may be used with both frequency division duplex (FDD) and time division duplex (TDD) component carriers.

[0081] In some examples (e.g., in a carrier aggregation configuration), a carrier may also have acquisition signaling or control signaling that coordinates the operation of other carriers. A carrier may be associated with a frequency channel (e.g., an Evolved Universal Mobile Telecommunications System Terrestrial Radio Access (E-UTRA) Absolute Radio Frequency Channel Number (EARFCN)) and may be located according to a channel grid for discovery by a UE 115. A carrier may operate in a standalone mode in which initial acquisition and connection may be performed by a UE 115 via the carrier, or a carrier may operate in a non-standalone mode in which the connection is anchored using a different carrier (e.g., a different carrier of the same or different radio access technology).

[0082] The communication link 125 shown in the wireless communication system 100 may include an uplink transmission from the UE 115 to the base station 105, or a downlink transmission from the base station 105 to the UE 115. A carrier may carry downlink or uplink communications (e.g., in FDD mode) or may be configured to carry both downlink and uplink communications (e.g., in TDD mode).

[0083] A carrier may be associated with a particular bandwidth of radio frequency spectrum, and in some examples, the carrier bandwidth may be referred to as the "system bandwidth" of the carrier or wireless communication system 100. For example, the carrier bandwidth may be one of several determined bandwidths (e.g., 1.4, 3, 5, 10, 15, 20, 40, or 80 megahertz (MHz)) of a carrier of a particular radio access technology. Devices of the wireless communication system 100 (e.g., base station 105, UE 115, or both) may have a hardware configuration that supports communication on a particular carrier bandwidth, or may be configurable to support communication on one of a set of carrier bandwidths. In some examples, the wireless communication system 100 may include a base station 105 or UE 115 that supports simultaneous communication via carriers associated with multiple carrier bandwidths. In some examples, each served UE 115 may be configured to operate on a portion (e.g., a subband, a BWP) or all of the carrier bandwidth.

[0084] The signal waveform transmitted on the carrier may include multiple subcarriers (e.g., using a multicarrier modulation (MCM) technique such as orthogonal frequency division multiplexing (OFDM) or discrete Fourier transform spread OFDM (DFT-S-OFDM)). In a system employing MCM techniques, a resource element may include one symbol period (e.g., the duration of one modulation symbol) and one subcarrier, where the symbol period and subcarrier spacing are inversely related. The number of bits carried by each resource element may depend on the modulation scheme (e.g., the order of the modulation scheme, the code rate of the modulation scheme, or both). Thus, the more resource elements received by UE 115 and the higher the order of the modulation scheme, the higher the data rate of UE 115 can be. Wireless communication resources may refer to a combination of radio frequency spectrum resources, time resources, and spatial resources (e.g., spatial layers or beams), and the use of multiple spatial layers may further improve the data rate or data integrity of communications with UE 115.

[0085] One or more parameter designs for a carrier may be supported, where the parameter designs may include subcarrier spacing (Δf) and cyclic prefix. A carrier may be divided into one or more BWPs with the same or different parameter designs. In some examples, a UE 115 may be configured with multiple BWPs. In some examples, a single BWP for a carrier may be active at a given time, and communications for the UE 115 may be limited to the one or more active BWPs.

[0086] The time interval of the base station 105 or the UE 115 can be expressed as a multiple of a basic time unit, which can be, for example, a sampling period T s =1 / (Δf max ·N f ) seconds, where Δf max It can represent the maximum supported subcarrier spacing, and N f The maximum supported discrete Fourier transform (DFT) size may be indicated. Time intervals of communication resources may be organized according to radio frames, each having a specified duration (e.g., 10 milliseconds (ms)). Each radio frame may be identified by a system frame number (SFN) (e.g., ranging from 0 to 1023).

[0087] Each frame may include a plurality of consecutively numbered subframes or time slots, and each subframe or time slot may have the same duration. In some examples, a frame may be divided (e.g., in the time domain) into subframes, and each subframe may be further divided into a number of time slots. Alternatively, each frame may include a variable number of time slots, and the number of time slots may depend on the subcarrier spacing. Each time slot may include a number of code element periods (e.g., depending on the length of the cyclic prefix added before each code element period). In some wireless communication systems 100, a time slot may be further divided into a plurality of mini-time slots containing one or more code elements. Excluding the cyclic prefix, each code element period may contain one or more (e.g., N f The duration of a symbol period may depend on the subcarrier spacing or the operating band.

[0088] A subframe, slot, mini-slot, or symbol may be the smallest scheduling unit (e.g., in the time domain) of the wireless communication system 100 and may be referred to as a transmission time interval (TTI). In some examples, the TTI duration (e.g., the number of symbol periods in a TTI) may be variable. Additionally or alternatively, the smallest scheduling unit of the wireless communication system 100 may be dynamically selected (e.g., in bursts of shortened TTIs (sTTIs)).

[0089] Physical channels may be multiplexed on a carrier according to various techniques. Physical control channels and physical data channels may be multiplexed on a downlink carrier, for example, using one or more of time division multiplexing (TDM), frequency division multiplexing (FDM), or hybrid TDM-FDM techniques. A control region (e.g., a control resource set (CORESET)) for physical control channels may be defined by a number of symbol periods and may extend across the system bandwidth or a subset of the system bandwidth of a carrier. One or more control regions (e.g., CORESETs) may be configured for a set of UEs 115. For example, one or more of UEs 115 may monitor or search the control region for control information according to one or more search space sets, and each search space set may include one or more control channel candidates in one or more aggregation levels arranged in a cascaded manner. An aggregation level for a control channel candidate may refer to the number of control channel resources (e.g., control channel elements (CCEs)) associated with encoded information for a control information format having a given payload size. The search space sets may include a common search space set configured for transmitting control information to multiple UEs 115 and a UE-specific search space set for transmitting control information to a specific UE 115 .

[0090] Each base station 105 may provide communication coverage via one or more cells (e.g., macro cells, small cells, hotspots, or other types of cells, or any combination thereof). The term "cell" may refer to a logical communication entity used to communicate with a base station 105 (e.g., on a carrier) and may be associated with an identifier (e.g., a physical cell identifier (PCID), a virtual cell identifier (VCID), or other) used to distinguish between adjacent cells. In some examples, a cell may also refer to a geographic coverage area 110 or a portion of a geographic coverage area 110 (e.g., a sector) on which the logical communication entity operates. The scope of such a cell may range from a smaller area (e.g., a structure, a subset of structures) to a larger area depending on various factors (such as the capabilities of the base station 105). For example, a cell may be or include a building, a subset of buildings, or an external space between or overlapping geographic coverage areas 110, among other examples.

[0091] A macro cell generally covers a relatively large geographic area (e.g., several kilometers in radius) and may allow unrestricted access to UEs 115 that have a service subscription with a network provider that supports the macro cell. A small cell may be associated with a lower power base station 105 (compared to a macro cell), and the small cell may operate in the same or different (e.g., licensed, unlicensed) frequency band as the macro cell. A small cell may provide unrestricted access to UEs 115 that have a service subscription with the network provider, or may provide restricted access to UEs 115 associated with the small cell (e.g., UEs 115 in a closed subscriber group (CSG), UEs 115 associated with users in a home or office). A base station 105 may support one or more cells and may also support communications over one or more cells using one or more component carriers.

[0092] In some examples, a carrier may support multiple cells, and different cells may be configured according to different protocol types (e.g., MTC, narrowband IoT (NB-IoT), enhanced mobile broadband (eMBB)) that may provide access to different types of devices.

[0093] In some examples, base stations 105 can be mobile and, therefore, provide communication coverage for mobile geographic coverage areas 110. In some examples, different geographic coverage areas 110 associated with different technologies can overlap, but the different geographic coverage areas 110 can be supported by the same base station 105. In other examples, overlapping geographic coverage areas 110 associated with different technologies can be supported by different base stations 105. The wireless communication system 100 can include, for example, a heterogeneous network in which different types of base stations 105 provide coverage for various geographic coverage areas 110 using the same or different radio access technologies.

[0094] The wireless communication system 100 may support synchronous or asynchronous operation. For synchronous operation, the base stations 105 may have similar frame timing, and transmissions from different base stations 105 may be approximately aligned in time. For asynchronous operation, the base stations 105 may have different frame timing, and transmissions from different base stations 105 may not be aligned in time in some examples. The techniques described herein may be used for either synchronous or asynchronous operation.

[0095] Some UEs 115, such as MTC or IoT devices, may be low-cost or low-complexity devices and may provide automated communication between machines (e.g., via machine-to-machine (M2M) communication). M2M communication or MTC may refer to data communication technology that allows devices to communicate with each other or with a base station 105 without human intervention. In some examples, M2M communication or MTC may include communications from devices that incorporate sensors or meters to measure or capture information and relay such information to a central server or application that utilizes the information or presents it to a person interacting with the application. Some UEs 115 may be designed to collect information or implement automated behavior of machines or other devices. Examples of applications for MTC devices include smart metering, inventory monitoring, water level monitoring, equipment monitoring, healthcare monitoring, wilderness survival monitoring, weather and geographic event monitoring, queue management and tracking, remote security sensing, physical access control, and transaction-based commercial charging.

[0096] Some UEs 115 may be configured to employ a reduced power consumption mode of operation, such as half-duplex communication (e.g., a mode that supports unidirectional communication via transmission or reception but not simultaneous transmission and reception). In some examples, half-duplex communication may be performed at a reduced peak rate. Other power saving techniques for UEs 115 include entering a power-saving deep sleep mode when not engaged in active communication, operating over a limited bandwidth (e.g., according to narrowband communication), or a combination of these techniques. For example, some UEs 115 may be configured to operate using a narrowband protocol type that is associated with a defined portion or range (e.g., a set of subcarriers or resource blocks (RBs)) within a carrier, within a guard band of a carrier, or outside a carrier.

[0097] The wireless communication system 100 can be configured to support ultra-reliable communication or low-latency communication or various combinations thereof. For example, the wireless communication system 100 can be configured to support ultra-reliable low-latency communication (URLLC) or mission-critical communication. The UE 115 can be designed to support ultra-reliable, low-latency or critical functions (e.g., mission-critical functions). Ultra-reliable communication can include private communication or group communication and can be supported by one or more mission-critical services such as mission-critical push-to-talk (MCPTT), mission-critical video (MCVideo), or mission-critical data (MCData). Support for mission-critical functions can include prioritization of services, and mission-critical services can be used for public safety or general commercial applications. The terms ultra-reliable, low-latency, mission-critical, and ultra-reliable low-latency can be used interchangeably herein.

[0098] In some examples, UE 115 may also be able to communicate directly with other UEs 115 over a device-to-device (D2D) communication link 135 (e.g., using a peer-to-peer (P2P) or D2D protocol). One or more UEs 115 utilizing D2D communication may be within the geographic coverage area 110 of base station 105. Other UEs 115 in such a group may be outside the geographic coverage area 110 of base station 105 or otherwise unable to receive transmissions from base station 105. In some examples, groups of UEs 115 communicating via D2D communication may utilize a one-to-many (1:M) system, in which each UE 115 transmits to every other UE 115 in the group. In some examples, base station 105 facilitates the scheduling of resources for D2D communication. In other cases, D2D communication is performed between UEs 115 without involving base station 105.

[0099] In some systems, the D2D communication link 135 can be an example of a communication channel (such as a sidelink communication channel) between vehicles (e.g., UE 115). In some examples, the vehicles can communicate using vehicle-to-everything (V2X) communication, vehicle-to-vehicle (V2V) communication, or some combination of these communications. The vehicles can signal information related to traffic conditions, signal scheduling, weather, safety, emergency situations, or any other information related to the V2X system. In some examples, the vehicles in the V2X system can use vehicle-to-network (V2N) communication to communicate with roadside infrastructure (such as roadside units), with the network, or with both, via one or more network nodes (e.g., base station 105).

[0100] The core network 130 may provide user authentication, access authorization, tracking, Internet Protocol (IP) connectivity, and other access, routing, or mobility functions. The core network 130 may be an evolved packet core (EPC) or a 5G core (5GC), and the EPC or 5GC may include at least one control plane entity (e.g., a mobility management entity (MME), an access and mobility management function (AMF)) that manages access and mobility, and at least one user plane entity (e.g., a serving gateway (S-GW), a packet data network (PDN) gateway (P-GW), or a user plane function (UPF)) that routes packets or interconnects to external networks. The control plane entity may manage non-access stratum (NAS) functions such as mobility, authentication, and bearer management for UEs 115 served by base stations 105 associated with the core network 130. User IP packets may be delivered through the user plane entity, which may provide IP address allocation and other functions. The user plane entity may be connected to the network operator IP service 150. Network operator IP services 150 may include access to the Internet, an intranet, an IP Multimedia Subsystem (IMS), or packet-switched streaming services.

[0101] Some network devices (such as base stations 105) may include subcomponents, such as access network entities 140, which may be examples of access node controllers (ANCs). Each access network entity 140 may communicate with each UE 115 through one or more other access network transport entities 145, which may be referred to as radio heads, smart radio heads, or transmit / receive points (TRPs). Each access network transport entity 145 may include one or more antenna panels. In some configurations, the various functions of each access network entity 140 or base station 105 may be distributed across various network devices (e.g., radio heads and ANCs) or consolidated into a single network device (e.g., base station 105).

[0102] The wireless communication system 100 can operate using one or more frequency bands, typically in the range of 300 megahertz (MHz) to 300 gigahertz (GHz). Generally speaking, the 300 MHz to 3 GHz region is referred to as the ultra-high frequency (UHF) region or the decimeter band because the wavelengths range from approximately 1 decimeter to 1 meter long. UHF waves can be blocked or redirected by buildings and environmental features, but these waves can penetrate various structures sufficiently for macrocells to provide service to UEs 115 located indoors. Transmissions using UHF waves can be associated with smaller antennas and a shorter range (e.g., less than 100 kilometers) compared to transmissions using the lower frequencies and longer wavelengths in the high frequency (HF) or very high frequency (VHF) portions of the spectrum below 300 MHz.

[0103] The wireless communication system 100 may also operate in a super high frequency (SHF) region using a frequency band from 3 GHz to 30 GHz (also known as a centimeter band) or in an extremely high frequency (EHF) region of the spectrum (e.g., from 30 GHz to 300 GHz) (also known as a millimeter band). In some examples, the wireless communication system 100 may support millimeter wave (mmW) communications between the UE 115 and the base station 105, and the EHF antennas of the corresponding devices may be smaller and more closely spaced than the UHF antennas. In some examples, this may facilitate the use of antenna arrays within the device. However, the propagation of EHF transmissions may be subject to even greater atmospheric attenuation and a shorter range than SHF or UHF transmissions. The techniques disclosed herein may be employed across transmissions using one or more different frequency regions, and the use of frequency bands specified across these frequency regions may vary by country or regulatory agency.

[0104] The wireless communication system 100 may utilize both licensed and unlicensed radio frequency spectrum bands. For example, the wireless communication system 100 may employ licensed assisted access (LAA), LTE unlicensed (LTE-U) radio access technology, or NR technology in an unlicensed band, such as the 5 GHz industrial, scientific, and medical (ISM) band. When operating in an unlicensed radio frequency spectrum band, devices (such as base stations 105 and UEs 115) may employ carrier sensing for conflict detection and avoidance. In some examples, operations in the unlicensed band may be based on a carrier aggregation configuration (e.g., LAA) in coordination with component carriers operating in the licensed band. Operations in the unlicensed spectrum may include downlink transmissions, uplink transmissions, P2P transmissions, or D2D transmissions, among others.

[0105] The base station 105 or UE 115 may be equipped with multiple antennas that can be used to employ techniques such as transmit diversity, receive diversity, multiple-input multiple-output (MIMO) communications, or beamforming. The antennas of the base station 105 or UE 115 may be located within one or more antenna arrays or antenna panels that can support MIMO operations or transmit or receive beamforming. For example, one or more base station antennas or antenna arrays may be co-located at an antenna assembly (such as an antenna tower). In some examples, the antennas or antenna arrays associated with the base station 105 may be located at different geographical locations. The base station 105 may have an antenna array having several rows and columns of antenna ports that the base station 105 can use to support beamforming for communications with the UE 115. Similarly, the UE 115 may have one or more antenna arrays that can support various MIMO or beamforming operations. Additionally or alternatively, the antenna panel may support radio frequency beamforming for signals transmitted via the antenna ports.

[0106] The base station 105 or the UE 115 can use MIMO communication to exploit multipath signal propagation and improve spectral efficiency by transmitting or receiving multiple signals via different spatial layers. Such a technique may be referred to as spatial multiplexing. For example, a transmitting device may transmit multiple signals via different antennas or different antenna combinations. Similarly, a receiving device may receive multiple signals via different antennas or different antenna combinations. Each of the multiple signals may be referred to as a separate spatial stream and may carry bits associated with the same data stream (e.g., the same codeword) or different data streams (e.g., different codewords). Different spatial layers may be associated with different antenna ports for channel measurement and reporting. MIMO techniques include single-user MIMO (SU-MIMO), in which multiple spatial layers are transmitted to the same receiving device, and multi-user MIMO (MU-MIMO), in which multiple spatial layers are transmitted to multiple devices.

[0107] Beamforming (which may also be referred to as spatial filtering, directional transmission, or directional reception) is a signal processing technique that can be used at a transmitting device or a receiving device (e.g., base station 105, UE 115) to shape or steer an antenna beam (e.g., a transmit beam, a receive beam) along a spatial path between the transmitting device and the receiving device. Beamforming can be achieved by combining signals communicated via antenna elements of an antenna array so that some signals propagating at a particular orientation relative to the antenna array experience constructive interference, while other signals experience destructive interference. Adjustments to signals communicated via antenna elements can include the transmitting device or the receiving device applying an amplitude offset, a phase offset, or both to the signals carried via the antenna elements associated with that device. The adjustments associated with each antenna element can be defined by a set of beamforming weights associated with a particular orientation (e.g., relative to the antenna array of the transmitting device or the receiving device, or relative to some other orientation).

[0108] The base station 105 or the UE 115 may use beam sweeping techniques as part of a beamforming operation. For example, the base station 105 may use multiple antennas or antenna arrays (e.g., antenna panels) to perform beamforming operations for directional communication with the UE 115. Some signals (e.g., synchronization signals, reference signals, beam selection signals, or other control signals) may be transmitted multiple times by the base station 105 in different directions. For example, the base station 105 may transmit signals according to different sets of beamforming weights associated with different transmission directions. The transmissions in different beam directions may be used (e.g., by a transmitting device (such as the base station 105) or a receiving device (such as the UE 115)) to identify a beam direction for later transmission or reception by the base station 105.

[0109] Some signals, such as data signals associated with a particular recipient device, may be transmitted by base station 105 in a single beam direction, e.g., a direction associated with a recipient device, such as UE 115. In some examples, a beam direction associated with transmissions along a single beam direction may be determined based on signals transmitted in one or more beam directions. For example, UE 115 may receive one or more signals transmitted by base station 105 in different directions and may report to base station 105 an indication of the signal received by UE 115 with the highest signal quality or other acceptable signal quality.

[0110] In some examples, transmission by a device (e.g., by a base station 105 or a UE 115) may be performed using multiple beam directions, and the device may use a combination of digital precoding or radio frequency beamforming to generate a combined beam for transmission (e.g., from the base station 105 to the UE 115). The UE 115 may report feedback indicating precoding weights for one or more beam directions, and the feedback may correspond to a configured number of beams across the system bandwidth or one or more subbands. The base station 105 may transmit a reference signal (e.g., a cell-specific reference signal (CRS), a CSI reference signal (CSI-RS)) that may be precoded or uncoded. The UE 115 may provide feedback for beam selection, which may be a precoding matrix indicator (PMI) or codebook-based feedback (e.g., a multi-panel type codebook, a linear combination type codebook, a port selection type codebook). Although these techniques are described with reference to signals transmitted by base station 105 in one or more directions, UE 115 may use similar techniques to transmit signals multiple times in different directions (e.g., to identify a beam direction for subsequent transmission or reception by UE 115) or to transmit signals in a single direction (e.g., to transmit data to a receiving device).

[0111] A receiving device (e.g., UE 115) may attempt multiple receive configurations (e.g., directional listening) when receiving various signals (such as synchronization signals, reference signals, beam selection signals, or other control signals) from base station 105. For example, the receiving device may attempt multiple receive directions by receiving via different antenna subarrays, processing received signals according to different antenna subarrays, receiving according to different receive beamforming weight sets applied to signals received at multiple antenna elements of an antenna array (e.g., different directional listening weight sets), or processing received signals according to different receive beamforming weight sets applied to signals received at multiple antenna elements of an antenna array, any of which may be referred to as "listening" according to different receive configurations or receive directions. In some examples, the receiving device may use a single receive configuration to receive along a single beam direction (e.g., when receiving data signals). The single receive configuration may be aligned on a beam direction determined based on listening according to different receive configuration directions (e.g., a beam direction determined to have the highest signal strength, highest signal-to-noise ratio (SNR), or other acceptable signal quality based on listening according to multiple beam directions).

[0112] The wireless communication system 100 can be a packet-based network that operates according to a layered protocol stack. In the user plane, the communication of the bearer or packet data convergence protocol (PDCP) layer can be IP-based. The radio link control (RLC) layer can perform packet segmentation and reassembly to communicate on the logical channel. The MAC layer can perform priority handling and multiplex logical channels into transport channels. The MAC layer can also use error detection technology, error correction technology, or both to support retransmission of the MAC layer to improve link efficiency. In the control plane, the radio resource control (RRC) protocol layer can provide the establishment, configuration and maintenance of the RRC connection of the radio bearer that supports user plane data between the UE 115 and the base station 105 or the core network 130. In the physical layer, the transport channel can be mapped to the physical channel.

[0113] UE 115 and base station 105 may support retransmission of data to increase the likelihood that the data is successfully received. Hybrid Automatic Repeat Request (HARQ) feedback is a technique for increasing the likelihood that data is correctly received on communication link 125. HARQ may include a combination of error detection (e.g., using a cyclic redundancy check (CRC)), forward error correction (FEC), and retransmission (e.g., automatic repeat request (ARQ)). HARQ may improve MAC layer throughput in poor radio conditions (e.g., low signal-to-noise ratio conditions). In some examples, a device may support simultaneous slot HARQ feedback, wherein the device may provide HARQ feedback in a particular time slot for data received in a previous symbol in that time slot. In other cases, the device may provide HARQ feedback in a subsequent time slot or based on some other time interval.

[0114] Wireless communication systems are widely deployed to provide various types of communication content, such as voice, video, packet data, messaging, broadcast, and the like. These systems can be multiple-access systems capable of supporting communication with multiple users by sharing available system resources (e.g., time, frequency, and power). A wireless network (e.g., a wireless local area network (WLAN), such as a Wi-Fi (i.e., Institute of Electrical and Electronics Engineers (IEEE) 802.11) network) may include an access point (AP) that can communicate with one or more wireless or mobile devices. An AP can be coupled to a network (such as the Internet) and can enable mobile devices to communicate via the network (or with other devices coupled to the access point). Wireless devices can communicate bidirectionally with network devices. For example, in a WLAN, a device can communicate with an associated AP via a downlink (e.g., a communication link from the AP to the device) and an uplink (e.g., a communication link from the device to the AP). A wireless personal area network (PAN), which may include a Bluetooth connection, can provide short-range wireless connections between two or more paired wireless devices. For example, a wireless device (such as a cellular telephone) may utilize wireless PAN communications to exchange information, such as audio signals, with a wireless head-mounted device.

[0115] The electromagnetic spectrum is typically subdivided into various classes, bands, channels, etc. based on frequency / wavelength. 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). Frequencies between FR1 and FR2 are typically referred to as mid-band frequencies. Although a portion of FR1 is greater than 6 GHz, FR1 is often (interchangeably) referred to as the “sub-6 GHz band” in various documents and articles. Similar naming issues sometimes arise regarding FR2, which is often (interchangeably) referred to as the “millimeter wave” band in various 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).

[0116] In view of the above aspects, unless otherwise specified, it should be understood that the terms "sub-6 GHz" and the like, if used herein, can broadly refer to frequencies that may be less than 6 GHz, may be within FR1, or may include mid-band frequencies. Furthermore, unless otherwise specified, it should be understood that the terms "millimeter wave" and the like, if used herein, can broadly refer to frequencies that may be less than 6 GHz, may be within FR1, or may include mid-band frequencies.

[0117] The wireless communication system 100 may support at least neural network-based CSI, which may use machine learning techniques to compress and feedback the channel and interference observed at the UE 115. For neural network-based CSI reporting, the UE 115 may use a resource set to train one or more layers of the neural network, or the entire neural network. The UE 115 generates weights or coefficients at each layer and indicates the weight set to the base station. The base station may recreate the channel based on the weight set and perform efficient channel maintenance. The layer state may correspond to the weight set contained in the layer obtained through training.

[0118] Wireless communication systems described herein, such as wireless communication systems 100 and 200, may support the reuse of trained neural networks or neural network layers to process other signals or channels. For example, a UE 115 may train a neural network or neural network layer, and the UE 115 may be instructed about an association between two types of signals or processes. The UE 115 may receive a neural network-based CSI report indicating the association. The UE 115 may have trained one or more layers of the neural network using a first set of one or more signals, and these trained layers may be used to efficiently process a second set of one or more signals at the UE 115. The base station 105 may indicate to the UE 115 that the one or more trained layers or neural networks may be reused to process another signal or channel. For example, the base station 105 may transmit an NNCI indicating one or more associations between the first set of one or more signals and the second set of one or more signals. If the UE 115 knows the association between the trained layer and a signal or channel that can be decoded / demodulated / estimated using the trained layer, the UE 115 may avoid cold-start training of the neural network or layer and save processing power and latency by reusing a previously trained layer or neural network.

[0119] The UE 115 may indicate a number of states that the UE 115 may store, track, or train per component carrier, frequency band, or subband. For example, the UE 115 may train a neural network on one CSI-RS from one base station 105 and use the training to demodulate another reference signal or channel. Some example associations using previously trained neural network states or layer states may include using CSI-RS training for DMRS, other CSI-RS, or PRS, using SSB training for CSI-RS or PRS, or using PRS for DMRS. The NNCI may include a source identifier of a signal or procedure used to train the neural network, a source identifier of an already trained neural network, or a source identifier for a particular layer. The NNCI may be configured semi-statically or dynamically. In some cases, the NNCI may be used for cross-component carrier NNCI relationships, cross-frequency range relationships, cross-band relationships, or cross-band combination relationships.

[0120] In various examples, the communication manager 101 may be included in the UE 115 to support a neural network or layer configuration indicator for a CSI scheme. The communication manager 102 may be included in a network entity such as a base station 105, a UE 115, a server, a transmission and reception point.

[0121] In some examples, communications manager 101 may train a first set of layers of a neural network based at least in part on the channel estimate using the set of resources. Communications manager 101 may generate a set of weights for the first set of layers of the neural network based at least in part on the training. Communications manager 101 may receive, from a first network entity, an indication of an association between a first set of one or more signals and a second set of one or more signals based at least in part on the first set of layers of the neural network. Communications manager 101 may receive a second set of one or more signals from a second network entity, and communications manager 101 may process the second set of one or more signals using the set of weights for the first set of layers based at least in part on the association between the first set of one or more signals and the second set of one or more signals. In some cases, communications manager 101 may process the second set of one or more signals using a subset of the weights, a subset of the first set of layers, or both based at least in part on the association between the first set of one or more signals and the second set of one or more signals.

[0122] In some examples, communications manager 102 may receive an indication of a trained set of layers of a neural network from UE 115 based at least in part on a channel estimate on a set of resources. Communications manager 102 may identify a set of weights for the trained set of layers of the neural network based at least in part on the indication of the trained set of layers. Communications manager 102 may transmit to the UE an indication of an association between the first set of one or more signals and the second set of one or more signals based at least in part on the trained set of layers of the neural network, and transmit the second set of one or more signals to the UE.

[0123] Figure 2 An example of a wireless communication system 200 that supports a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is illustrated. In some examples, the wireless communication system 200 can implement aspects of the wireless communication system 100. The wireless communication system 200 can include a UE 115-a and a base station 105-a, which can be referenced to Figure 1 The respective examples of UE 115 and base station 105 are described. In some examples, UE 115-a can be an example of an encoding device, and base station 105-a can be an example of a decoding device. In some other examples, another UE 115 can be an example of a decoding device.

[0124] UE 115-a may measure reference signals, channels, or both to report to a network entity. For example, UE 115-a may measure signaling 205 to determine CSI or CSF, and UE 115-a may transmit a CSI report to base station 105-a to indicate the measured channel conditions. In some cases, signaling 205 may include a reference signal (such as a CSI-RS) that UE 115-a may measure to determine CSI. In some cases, UE 115-a may measure the received power of reference signals from a serving cell and / or a neighboring cell, signal strength of an inter-radio access technology (e.g., Wi-Fi) network, or sensor signals used to detect the location of one or more objects within an environment, as well as other types of signaling. The CSI report may include channel quality information (CQI), a precoding matrix indicator (PMI), a rank indicator, a CSI-RS resource indicator (CRI), an SSB resource indicator (SSBRI), a layer indicator, or any combination thereof.

[0125] Some wireless communication systems support multiple types of CSI. A first type of CSI (e.g., Type 1 CSI) may be used for a beam selection scheme, where the UE 115 selects the best possible beam index and reports CSI information based on the best beam index. A second type of CSI (e.g., Type 2 CSI) may be used for a beam combining scheme, where the UE 115 also calculates the best linear combination coefficients for the various beams and reports the beam index. In some cases of Type 2 CSI, the UE 115 may report the coefficients used to combine the beams. In some cases, Type 2 CSI reporting may occur on a subband or configured subband basis.

[0126] The wireless communication system 200 may support at least a neural network-based CSI reporting scheme. The neural network-based CSI may use machine learning techniques to compress and feedback the channel, including the interference observed at the UE 115. For example, the UE 115-a may determine a set of weights or coefficients that represent a compressed form of a received channel (e.g., a downlink channel, a sidelink channel, etc.). The UE 115-a may report the set of weights to the base station 105-a in a CSI report, and the base station 105-a may be able to reconstruct the channel based on the set of weights. By compressing the channel with respect to the neural network-based CSI report 210, the CSI report 210 may be comprehensive, thereby informing the base station 105-a of both the channel and any interference. In some cases, the neural network-based CSI report may compress the entire channel, or the neural network-based CSI report may have variable granularity or accuracy for individual subbands.

[0127] In some aspects described herein, a coding device (such as a UE 115-a) may train one or more neural networks, or one or more layers of a neural network, to support neural network-based CSI reporting 210. In some cases, the one or more layers or one or more neural networks may be trained to learn the dependence of measured quality on individual parameters, isolate the measured quality through various layers (also referred to as "operations") of the one or more neural networks, and compress the measurements in a manner that limits compression losses. In some aspects, the UE 115-a may use the nature of the number of bits being compressed to construct a step-by-step extraction and compression of each feature (also referred to as a dimension) that affects the number of bits. In some aspects, the number of bits may be associated with samples of one or more reference signals and / or may indicate CSI. For example, the UE 115-a may encode the measurements using one or more extraction operations and compression operations associated with the neural network to produce compressed measurements, where the one or more extraction operations and compression operations are based at least in part on a feature set of the measurements.

[0128] As an example, UE 115-a may be configured to perform one or more operations on samples (e.g., data) received via one or more antennas of an encoding device to compress the samples. In some cases, the term "layer" may be used to refer to the operation performed on the input data. For example, there are fully connected layers, convolutional layers, etc. For example, for a layer AxB(p), A may refer to the number of input features, B may refer to the number of output features, and p may refer to the kernel size, where the kernel size refers to the number of adjacent coefficients that are combined in one dimension. For a one-dimensional convolution, p may be a single value, where p may include multiple values ​​for higher dimensions (e.g., as a tuple).

[0129] In some aspects, the encoding device may identify features to be compressed. In some aspects, the encoding device may perform a first type of operation on a first dimension associated with the features to be compressed. The encoding device may perform a second type of operation on other dimensions (e.g., on all other dimensions). For example, the encoding device may perform a fully connected operation on a first dimension and perform convolution (e.g., point-by-point convolution) on all other dimensions. In some aspects, reference numerals identify operations that include multiple neural network layers and / or operations. The neural networks of the encoding device and the decoding device may be formed by a cascade of one or more of the recited operations.

[0130] In one example, base station 105-a may transmit signaling 205 to UE 115-a. Signaling 205 may include a reference signal, such as a CSI-RS. Signaling 205 may be an input to one or more neural networks, each having one or more layers. For example, 115-a may perform spatial feature extraction 220 on the input. In some cases, UE 115-a may perform tap domain feature extraction 225 on the data. In some examples, UE 115-a may perform tap domain feature extraction before performing spatial feature extraction. In some cases, some features may be extracted simultaneously, or multiple layers of operations may be performed simultaneously. In some aspects, the extraction operation may include multiple operations or layers. For example, the multiple operations may include one or more convolution operations that may be activated or inactive, one or more fully connected operations, etc. In some aspects, the extraction operation may include a residual neural network (ResNet) operation.

[0131] UE 115-a may perform feature compression 230 on one or more features extracted from the input. In some aspects, the compression operation may include one or more operations such as one or more convolution operations, one or more fully connected operations, etc. Feature compression 230 may further compress the two-dimensional spatiotemporal features into a lower dimensional vector (e.g., of size M) for over-the-air transmission. After compression, the bit count of the output may be less than the bit count of the input.

[0132] UE 115-a may perform a quantization operation 235 before transmitting the encoder output 240 over the air to base station 105-a. In some aspects, the encoding device may perform the quantization operation after flattening the output of the compression operation and / or performing the feature compression operation after flattening the output.

[0133] UE 115-a may transmit the compressed measurements to a network entity, such as a server, a TRP, another UE, a base station, etc. Although the examples described herein cite base station 105 as a decoding device, the decoding device may be any network entity. The network entity may be referred to as a "decoding device."

[0134] Base station 105-a may receive the CSI feedback and attempt to reconstruct the channel. For example, base station 105-a may decode the compressed measurement using one or more decompression and reconstruction operations associated with a neural network. The one or more decompression and reconstruction operations may be based at least in part on a feature set of the compressed data set to produce a reconstructed measurement. A decoding device may use the reconstructed measurement as CSI feedback.

[0135] For example, base station 105-a may receive encoder output 240 and perform feature decompression 245 on encoder output 240. Base station 105-a may then perform tap-domain feature reconstruction 250. Base station 105-a may perform spatial feature reconstruction 255 and attempt to reconstruct the channel as received by UE 115-a. In some aspects, the decoding device may perform spatial feature reconstruction before performing tap-domain feature reconstruction. Additionally or alternatively, some features may be extracted, reconstructed, or decompressed simultaneously. After the reconstruction operation, the decoding device may obtain a reconstructed output 260 that is input to the encoding device.

[0136] In some cases, the decoder (e.g., base station 105-a) may follow the reverse order of the encoder (e.g., UE 115-a). For example, if the encoding device follows operations (a, b, c, d), the decoding device may follow the reverse operations (D, C, B, A). In some aspects, the decoding device may perform operations that are completely symmetrical to the operations of the encoding device. This may reduce the number of bits required for the neural network configuration at the UE. In some aspects, the decoding device may perform additional operations in addition to the operations of the encoding device (e.g., convolution operations, fully connected operations, ResNet operations, etc.). In some aspects, the decoding device may perform operations that are asymmetrical to the operations of the encoding device.

[0137] Based at least in part on the coding device using a neural network to encode a data set for uplink communication, the coding device (e.g., a UE) can transmit CSI with a reduced payload. This can save network resources that might otherwise have been used to transmit the full data set as sampled by the coding device.

[0138] The wireless communication system 200 may support the reuse of trained layers or trained neural networks for efficient communication at the device. For example, the UE 115-a may use a set of resources to train a neural network layer or an entire neural network. Each trained layer state may correspond to a set of weights contained in the layer. Signaling may be sent to the UE 115-a to inform the UE 115-a that the UE 115-a can reuse one or more trained layers of the neural network to process another signal or channel. For example, the UE 115-a may use CSI-RS to train a layer of the neural network, and the UE 115-a may be able to use the trained layer to process other types of signals or signaling, such as decoding PRS or SSB. The UE 115-a may decode, demodulate, estimate, or compress another signal or channel, or any combination thereof, based on reusing a previously trained layer or neural network. In some cases, the UE 115-a may use a subset of the weights or layers of the neural network to process other types of signals. For example, UE 115-a may reuse one or more layers of the neural network (e.g., up to all trained layers), or UE 115-a may reuse some of the determined weights of these layers (e.g., up to all determined weights) to process other types of signals.

[0139] 215. The UE 115-a may be informed of an association between reference signals or channels that the UE 115-a may use for enhanced processing. The indication of the association between the first set of one or more signals and the second set of one or more signals may be referred to as NNCI 215. For example, if the UE 115-a is aware of the association, the UE 115-a may avoid performing a cold start of a neural network or layer, which may be used for other purposes. For example, instead of training a new neural network or neural network layer for PRS, the UE 115-a may at least partially use a neural network or neural network layer that has been trained using CSI-RS. Some weights of the neural network trained by CSI-RS may be applied to the neural network or at least one layer of the neural network for PRS. Additionally or alternatively, the UE 115-a may determine weights for at least one layer of the neural network for PRS based on the weights of the neural network for CSI-RS. The UE 115-a may save processing power and latency by reusing a previously trained layer or neural network.

[0140] The UE 115-a may report the number of layer states that may be saved, stored, or tracked at the UE 115-a. The UE 115-a may report the number per component carrier, per frequency band, per subband, or per frequency band combination. For example, the UE 115-a may store one or more associations so that the UE 115-a may reuse multiple different layers or neural networks to process other signals or channels.

[0141] NNCI 215 may include a source identifier and a target identifier for the association. In some cases, the source identifier may refer to a first set of signals, while the target identifier may refer to a second set of signals. In some cases, the source identifier may include a source identifier of a particular signal or procedure (e.g., a CSI process or report ID) that has been used to train the neural network. Additionally or alternatively, the source identifier may include an identifier of the trained neural network. In some cases, the source identifier may include an identifier of a particular layer of the neural network. The target identifier may refer to a signal or procedure that can be operated on or performed using a previously trained neural network or layer.

[0142] UE 115-a may be configured with multiple different types of associations for reusing a neural network or neural network layer. For example, UE 115-a may use CSI-RS to train a neural network or neural network layer. UE 115-a may then reuse the neural network or neural network layer for processing DMRS, another CSI-RS, or PRS. Additionally or alternatively, UE 115-a may use SSB to train a neural network or neural network layer, and UE 115-a may reuse the trained layer or trained neural network for CSI-RS or PRS. Additionally or alternatively, UE 115-a may use PRS to train a neural network or neural network layer, and UE 115-a may reuse the trained layer or trained layer for processing DMRS. Thus, a first set of signals may include CSI-RS, SSB, DMRS, PRS, or any combination thereof. A second set of signals may include CSI-RS, SSB, DMRS, PRS, or any combination thereof.

[0143] In one example, the source identifier may correspond to a CSI-RS and the target identifier may correspond to a DMRS, such that the UE 115-a may reuse a layer or neural network (e.g., trained with CSI-RS) for the DMRS. In another example, the source identifier may correspond to an SSB and the target identifier may correspond to a demodulation, such that the UE 115-a may reuse a layer or neural network (e.g., trained with SSB) for demodulating other types of signals (e.g., CSI-RS or PRS).

[0144] In another example, UE 115-a may receive NNCI 215 indicating that a particular layer of a neural network trained by UE 115-a for CSI is to be reused. UE 115-a may reuse the layer for demodulation, and the target identifier may indicate a downlink shared channel or DMRS. Thus, UE 115-a may use the neural network or neural network layer trained for CSI to demodulate the downlink shared channel or demodulate the DMRS transmitted for the downlink shared channel.

[0145] In another example, UE 115-a may receive NNCI 215 having a source identifier indicating a first neural network and a target identifier indicating a CSI-RS. In some cases, UE 115-a may train a second neural network based on a first set of weights of the first neural network. For example, UE 115-a may generate a second set of weights for the second neural network based on the first set of weights of the first neural network based on the compression thereof.

[0146] NNCI 215 may be configured semi-statically or dynamically. For example, NNCI 215 may be transmitted semi-statically via a MAC CE, configured via an RRC message, or transmitted in downlink control information on a downlink control channel. The association indicated by NNCI 215 may be across component carriers, across frequency bands, or across a combination of frequency bands. In some cases, a specific layer may support cross-component carrier association. In some cases, the association may be across frequency ranges. For example, a source identifier may be associated with a first frequency range, while a target identifier may be associated with a second frequency range. In some cases, a specific layer of the neural network may support cross-frequency range association.

[0147] In some cases, some layers of the neural network may be used for NNCI relationships, while other layers may not. For example, a feature compression layer may be at the end of the encoding chain, and thus may only be applicable to (e.g., relevant to or informative about) the specific signal or protocol on which the neural network was trained. However, earlier layers may track that some long-term components are relevant across channels, signals, protocols, or any combination thereof, and may be used for other protocols or channels.

[0148] In some cases, a first network device (e.g., base station 105-a) may transmit both the first set of signals and the NNCI 215. In some other examples, UE 115-a may train on the signals from the first network device and receive the NNCI from a second network device. UE 115-a may receive the second set of signals (e.g., which may be associated with the first set of signals as indicated by the NNCI) from the first network device, the second network device, or a third network device. In some examples described herein, base station 105 may be described as transmitting NNCI 215. However, any network entity may transmit the NNCI and enable or support UE 115 to reuse previously trained neural network layers or previously trained neural networks.

[0149] Figure 3 An example of a CSI reporting scheme 300 that supports a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated. In some examples, the CSI reporting scheme 300 can implement aspects of the wireless communication system 100.

[0150] The CSI reporting scheme 300 may include an encoding device 301 and a decoding device 302. In some examples, the encoding device 301 may be as described with reference to Figure 1 and 2 The decoding device 302 may be an example of a UE 115 as described above. Figure 1 and 2 Examples of UE 115, base station 105, server or transmission and reception point or another network entity are described.

[0151] The encoding apparatus 301 may include a CSI instance encoder 310, a CSI sequence encoder 320, and a memory 330. The decoding apparatus 302 may include a CSI sequence decoder 360, a memory 370, and a CSI instance decoder 380.

[0152] In some aspects, the encoding device 301 and the decoding device 302 may utilize the correlation of CSI instances over time (temporal aspect) or a sequence of CSI instances to perform a series of channel estimates. The encoding device 301 and the decoding device 302 may save and use previously stored CSI and only encode and decode CSI changes from the previous instance. This may provide less CSI feedback overhead and improve performance. The encoding device 301 may also be able to encode more accurate CSI, and the neural network may be trained with more accurate CSI.

[0153] like Figure 3 As shown in , the CSI instance encoder 310 may encode the CSI instance into intermediate coded CSI for each downlink channel estimate in the downlink channel estimate sequence. The CSI instance encoder 310 (e.g., a feed-forward network) may use neural network encoder weights θ. The intermediate coded CSI may be represented as The CSI sequence encoder 320 (e.g., a long short-term memory (LSTM) network) may determine a previously encoded CSI instance h(t-1) from the memory 330 and compare the intermediate encoded CSI m(t) with the previously encoded CSI instance h(t-1) to determine a change n(t) in the encoded CSI. The change n(t) may be a portion of the channel estimate that is new and may not have been predicted by the decoding device 302. The encoded CSI at this point may be represented by The CSI sequence encoder 320 may provide the change n(t) on a physical uplink shared channel (PUSCH) or a physical uplink control channel (PUCCH), and the encoding device 301 may transmit the change (e.g., information indicating the change) n(t) as encoded CSI to the decoding device 302 on an uplink channel. Because the change is smaller than the entire CSI instance, the encoding device 301 may send a smaller payload for the encoded CSI on the uplink channel while including more detailed information about the change in the encoded CSI. The CSI sequence encoder 320 may generate the encoded CSI h(t) based at least in part on the intermediate encoded CSI m(t) and at least a portion of the previously encoded CSI instance h(t-1). The CSI sequence encoder 320 may store the encoded CSI h(t) in the memory 330.

[0154] The CSI sequence decoder 360 may receive the encoded CSI on the PUSCH or PUCCH. The CSI sequence decoder 360 may determine that only the change n(t) of the CSI is received as the encoded CSI. The CSI sequence decoder 360 may determine the intermediate decoded CSI m(t) based at least in part on the encoded CSI and at least a portion of the previous intermediate decoded CSI instance h(t-1) from the memory 370 and the change. The CSI instance decoder 380 may decode the intermediate decoded CSI m(t) into decoded CSI. The CSI sequence decoder 360 and the CSI instance decoder 380 may use the neural network decoder weights φ. The intermediate decoded CSI may be determined by The CSI sequence decoder 360 may generate decoded CSI h(t) based at least in part on the intermediate decoded CSI m(t) and at least a portion of the previously decoded CSI instance h(t-1). The decoding device 302 may reconstruct a downlink channel estimate from the decoded CSI h(t), and the reconstructed channel estimate may be represented as The CSI sequence decoder 360 may store the decoded CSI h(t) in the memory 370 .

[0155] Because the change n(t) is smaller than the entire CSI instance, the encoding device 301 can send a smaller payload on the uplink channel. For example, if the downlink channel has barely changed from the previous feedback due to low Doppler or minimal motion of the encoding device 301, the output of the CSI sequence encoder can be quite compact. In this way, the encoding device 301 can exploit the correlation of the channel estimate over time. In some aspects, because the output is smaller, the encoding device 301 can include more detailed information about the change in the encoded CSI. In some aspects, the encoding device 301 can transmit an indication (e.g., a flag) to the decoding device 302 that the encoded CSI was encoded in time (CSI change). Alternatively, the encoding device 301 can transmit an indication that the encoded CSI was encoded independently of any previously encoded CSI feedback. The decoding device 302 can decode the encoded CSI without using the previously decoded CSI instance. In some aspects, a device (which may include the encoding device 301 or the decoding device 302) can use the CSI sequence encoder and the CSI sequence decoder to train a neural network model.

[0156] In some aspects, CSI may be a function of a channel estimate (referred to as a channel response) H and interference N. There may be a variety of ways to convey H and N. For example, the encoding device 301 may encode the CSI as N -1 / 2H. The encoding device 301 may encode H and N separately. The encoding device 301 may partially encode H and N separately, and then jointly encode the two partially encoded outputs. It may be advantageous to encode H and N separately. Interference and channel variations may occur on different time scales. In low Doppler scenarios, the channel may be stable, but interference may still change rapidly due to traffic or scheduler algorithms. In high Doppler scenarios, the channel may change faster than the UE's scheduler grouping. In some aspects, a device (which may include the encoding device 301 or the decoding device 302) may use the separately encoded H and N to train a neural network model.

[0157] In some aspects, the reconstructed downlink channel H^ may faithfully reflect the downlink channel H, and this may be referred to as explicit feedback. In some aspects, H^ may capture only the information required for the decoding device 302 to derive rank and precoding. CQI may be fed back separately. In a time-coded scenario, CSI feedback may be expressed as m(t) or n(t). Similar to type-2 CSI feedback, m(t) may be structured as a concatenation of a rank index (RI), a beam index, and coefficients representing amplitude or phase. In some aspects, m(t) may be a quantized version of a real-valued vector. The beam may be predefined (not obtained through training) or may be part of the training (e.g., part of θ and φ and communicated to the encoding device 301 or the decoding device 302).

[0158] In some aspects, the decoding device 302 and the encoding device 301 may maintain multiple encoder and decoder networks, each targeting a different payload size (to achieve a different accuracy-versus-uplink overhead tradeoff). For each CSI feedback, depending on the reconstruction quality and uplink budget (e.g., PUSCH payload size), the encoding device 301 may select, or the decoding device 302 may instruct the encoding device 301 to select, one of the encoders to construct the encoded CSI. The encoding device 301 may transmit an encoder index along with the CSI based at least in part on the encoder selected by the encoding device 301. Similarly, the decoding device 302 and the encoding device 301 may maintain multiple encoder and decoder networks to account for different antenna geometries and channel conditions. Note that although some operations are described with respect to the decoding device 302 and the encoding device 301, these operations may also be performed by another device as part of preconfiguration of the encoder and decoder weights and / or structure.

[0159] The CSI reporting scheme 300 can be implemented to support neural network-based CSI reporting. In some cases, the CSI reporting scheme may support the reuse of trained layers or trained neural networks to process other signals or channels. For example, the decoding device 302 or another network entity may transmit a NNCI (Network Node Identifier) ​​including a source identifier and a target identifier to the encoding device 301. The NNCI may signal to the encoding device 301 that the encoding device 301 may reuse one or more trained layers or neural networks to decode, demodulate, estimate, compress, train, or any combination thereof, another signal or channel. The source identifier may be associated with a first protocol, a first set of signals, a first layer of a neural network, or a first neural network. The target identifier may be associated with a second set of signals or a second protocol. The NNCI may indicate that the trained layer or trained neural network associated with the source identifier may be reused to process a second set of signals or a second protocol indicated by the target identifier. These associations may be indicated to the encoding device 301 so that the encoding device 301 can avoid cold-start training of the neural network or layers for other purposes. By reusing previously trained layers or neural networks, the encoding device 301 can save processing power and latency.

[0160] Figure 4 An example of an example 400 supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated. In some examples, example 400 can implement aspects of wireless communication system 100.

[0161] An encoding device (eg, UE 115, encoding device 301, etc.) may be configured to perform one or more operations on data to compress the data. A decoding device (eg, base station 105, decoding device 302, etc.) may be configured to decode the compressed data to determine information.

[0162] As used herein, a "layer" of a neural network is used to represent an operation on input data. For example, a convolutional layer, a fully connected layer, etc., represent associated operations on the data input to the layer. A convolution AxB operation is an operation that transforms several input features A into several output features B. The "kernel size" refers to the number of adjacent coefficients that are combined in a dimension.

[0163] As used herein, "weights" are used to refer to one or more coefficients used in an operation in a layer for combining rows and / or columns of input data. For example, a fully connected layer operation may have an output y that is determined at least in part based on the product of the input matrix x and the weights A (which may be a matrix) and the sum of the bias values ​​B (which may be a matrix). The term "weights" may be used herein to generally refer to both weights and bias values.

[0164] As shown in example 400, the encoding device may perform a convolution operation on the samples. For example, the encoding device may receive a set of bits structured as a 2x64x32 data set, which indicates IQ samples for tap features (e.g., associated with multipath timing offset) and spatial features (e.g., associated with different antennas of the encoding device). The convolution operation may be a 2x2 operation with a kernel size of 3x3 on the data structure. The output of the convolution operation may be input to a batch normalization (BN) layer, followed by a LeakyReLu activation, thereby providing an output data set with a size of 2x64x32. The encoding device may perform a flattening operation to flatten the bits into a 4096-bit vector. The encoding device may apply a fully connected operation with dimensions 4096xM to the 4096-bit vector to output an M-bit payload. The encoding device may transmit the M-bit payload to the decoding device.

[0165] The decoding device may apply a fully connected operation having dimensions Mx4096 to the M-bit payload to output a 4096-bit vector. The decoding device may reshape the 4096-bit vector to have dimensions 2x64x32. The decoding device may apply one or more refinement network (RefineNet) operations to the reshaped bit vector. For example, the RefineNet operation may include applying a 2x8 convolution operation (e.g., with a kernel size of 3x3) to the output of the BN layer, followed by a LeakyReLU activation, which produces an output dataset having dimensions 8x64x32, applying an 8x16 convolution operation (e.g., with a kernel size of 3x3) to the output of the BN layer, followed by a LeakyReLU activation, which produces an output dataset having dimensions 16x64x32, and / or applying a 16x2 convolution operation (e.g., with a kernel size of 3x3) to the output of the BN layer, followed by a LeakyReLU activation, which produces an output dataset having dimensions 2x64x32. The decoding device may also apply 2x2 convolution operations with kernel sizes of 3 and 3 to generate decoded and / or reconstructed output.

[0166] As indicated above, Figure 4 These are provided as examples only. Other examples may differ from those regarding Figure 4 Examples described.

[0167] As described herein, a coding device operating in a network may measure reference signals, etc., to report to a decoding device. For example, a UE may measure reference signals during a beam management process to report CSI, may measure the received power of reference signals from a serving cell and / or neighboring cells, may measure signal strength across an inter-radio access technology (e.g., WiFi) network, may measure sensor signals for detecting the location of one or more objects within an environment, etc. However, reporting such information to a network entity may consume communication and / or network resources.

[0168] In some aspects described herein, a coding device (e.g., a UE) may train one or more neural networks to learn the dependencies of these measured qualities on individual parameters, isolate these measured qualities through various layers (also referred to as "operations") of the one or more neural networks, and compress these measurements in a manner that limits compression losses.

[0169] In some aspects, the encoding device may use the nature of the number of bits being compressed to construct a method for progressively extracting and compressing each feature (also referred to as a dimension) that affects the number of bits. In some aspects, the number of bits may be associated with samples of one or more reference signals and / or may indicate CSI.

[0170] Based at least in part on using a neural network to encode and decode a data set for uplink communication, the encoding device can transmit the CSF or CSI with a reduced payload. This can save network resources that might otherwise have been used to transmit the full data set as sampled by the encoding device.

[0171] Example 400 may support the reuse of one or more trained layers of a neural network to process other signals or channels. For example, the decoding device 302 or another network entity may transmit an indication (such as an NNCI) including a source identifier and a target identifier of an associated signal, channel, or layer to the encoding device 301. The NNCI may indicate that the encoding device 301 may reuse one or more trained layers or neural networks to decode, demodulate, estimate, compress, train, or any combination thereof the indicated signal or channel (e.g., corresponding to the target identifier). The source identifier may be associated with a first procedure, a first set of signals, a first layer of a neural network, or a first neural network. The target identifier may be associated with a second set of signals or a second procedure. The NNCI may indicate that the trained layer or trained neural network associated with the source identifier may be reused to process a second set of signals or a second procedure, as indicated by the target identifier.

[0172] Figure 5An example of an example 500 of an example supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is illustrated. In some examples, example 500 can implement aspects of wireless communication system 100. An encoding device (e.g., UE 115, encoding device 301, etc.) can be configured to perform one or more operations on samples (e.g., data) received via one or more antennas of the encoding device to compress the samples. A decoding device (e.g., base station 105, decoding device 302, etc.) can be configured to decode the compressed samples to determine information, such as a CSF.

[0173] In some aspects, the encoding device may identify features to be compressed. In some aspects, the encoding device may perform a first type of operation on a first dimension associated with the features to be compressed. The encoding device may perform a second type of operation on other dimensions (e.g., on all other dimensions). For example, the encoding device may perform a fully connected operation on the first dimension and perform convolution (e.g., point-by-point convolution) on all other dimensions.

[0174] In some aspects, reference numerals identify operations that include multiple neural network layers and / or operations. The neural networks of the encoding device and the decoding device can be formed by concatenating one or more of the recited operations.

[0175] As indicated by reference numeral 505, the encoding device may perform spatial feature extraction on the data. As indicated by reference numeral 510, the encoding device may perform tap domain feature extraction on the data. In some aspects, the encoding device may perform tap domain feature extraction before performing spatial feature extraction. In some aspects, the extraction operation may include multiple operations. For example, the multiple operations may include one or more convolution operations that may or may not be activated, one or more fully connected operations, etc. In some aspects, the extraction operation may include a residual neural network (ResNet) operation.

[0176] As shown by reference numeral 515, the encoding device may compress the one or more features that have been extracted. In some aspects, the compression operation may include one or more operations, such as one or more convolution operations, one or more fully connected operations, etc. After compression, the bit count of the output may be less than the bit count of the input.

[0177] The encoding device may perform a quantization operation, as indicated by reference numeral 520. In some aspects, the encoding device may perform the quantization operation after flattening the output of the compression operation and / or performing a fully connected operation after flattening the output.

[0178] As indicated by reference numeral 525, the encoding device may perform feature decompression. As indicated by reference numeral 530, the decoding device may perform tap-domain feature reconstruction. As indicated by reference numeral 535, the decoding device may perform spatial feature reconstruction. In some aspects, the decoding device may perform spatial feature reconstruction before performing tap-domain feature reconstruction. After the reconstruction operation, the decoding device may output a reconstructed version of the input of the encoding device.

[0179] In some aspects, the decoding device may perform operations in the reverse order of the operations performed by the encoding device. For example, if the encoding device follows operations (a, b, c, d), the decoding device may follow the reverse operations (D, C, B, A). In some aspects, the decoding device may perform operations that are completely symmetrical to the operations of the encoding device. This may reduce the number of bits required for the neural network configuration at the UE. In some aspects, the decoding device may perform additional operations in addition to the operations of the encoding device (e.g., convolution operations, fully connected operations, ResNet operations, etc.). In some aspects, the decoding device may perform operations that are asymmetrical to the operations of the encoding device.

[0180] Based at least in part on the coding device using a neural network to encode a data set for uplink communication, the coding device (e.g., a UE) can transmit the CSF with a reduced payload. This can save network resources that might otherwise have been used to transmit the full data set as sampled by the coding device.

[0181] Example 500 may support the reuse of one or more trained layers of a neural network to process other signals or channels. For example, a decoding device or another network entity may transmit an indication (such as an NNCI) including a source identifier and a target identifier of an associated signal, channel, or layer to an encoding device. The NNCI may indicate that the encoding device may reuse one or more trained layers or neural networks to decode, demodulate, estimate, compress, train, or any combination thereof the indicated signal or channel (e.g., corresponding to the target identifier). The source identifier may be associated with a first procedure, a first set of signals, a first layer of a neural network, or a first neural network. The target identifier may be associated with a second set of signals or a second procedure. The NNCI may indicate that the trained layer or trained neural network associated with the source identifier may be reused to process a second set of signals or a second procedure indicated by the target identifier.

[0182] As indicated above, Figure 5 These are provided as examples only. Other examples may differ from those regarding Figure 5 Examples described.

[0183] Figure 6An example of an example 600 supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated. In some examples, example 600 can implement aspects of wireless communication system 100.

[0184] An encoding device (e.g., UE 115, encoding device 301, etc.) may be configured to perform one or more operations on samples (e.g., data) received via one or more antennas of the encoding device to compress the samples. A decoding device (e.g., base station 105, decoding device 302, etc.) may be configured to decode the compressed samples to determine information, such as a CSF.

[0185] As shown by example 600, the encoding device may receive samples from the antennas. For example, the encoding device may receive a data set of size 64x64 based at least in part on the number of antennas, the number of samples per antenna, and the tap characteristics.

[0186] The encoding device may perform spatial feature extraction, short-term (tap) feature extraction, and the like. In some aspects, this may be achieved by using a 1-dimensional convolution operation that is fully connected in the spatial dimension (to extract spatial features) and a simple convolution with a small kernel size (e.g., 3) in the tap dimension (to extract short-tap features). The output of such a 64xW 1-dimensional convolution operation may be a Wx64 matrix.

[0187] The encoding device may perform one or more ResNet operations. The one or more ResNet operations may further refine the spatial features and / or temporal features. In some aspects, the ResNet operation may include multiple operations associated with the features. For example, the ResNet operation may include multiple (e.g., 3) 1-dimensional convolution operations, skip connections (e.g., between the input of the ResNet and the output of the ResNet to avoid applying a 1-dimensional convolution operation), a summation operation of a path through multiple 1-dimensional convolution operations and a path through a skip connection, etc. In some aspects, the plurality of 1-dimensional convolution operations may include: a Wx256 convolution operation with a kernel size of 3, the output of which is input to a BN layer, followed by a LeakyReLU activation, which produces an output dataset of size 256x64; a 256x512 convolution operation with a kernel size of 3, the output of which is input to a BN layer, followed by a LeakyReLU activation, which produces an output dataset of size 512x64; and a 512xW convolution operation with a kernel size of 3, the output of which has a BN dataset of size Wx64. The output from one or more ResNet operations may be a Wx64 matrix.

[0188] The encoding device may perform a WxV convolution operation on the output from one or more ResNet operations. The WxV convolution operation may include a point-by-point (e.g., tap-by-tap) convolution operation. The WxV convolution operation may compress spatial features into a reduced dimension for each tap. The WxV convolution operation has an input of W features and an output of V features. The output from the WxV convolution operation may be a Vx64 matrix.

[0189] The encoding device may perform a flattening operation to flatten the Vx64 matrix into a 64V element vector. The encoding device may perform a 64VxM fully connected operation to further compress the spatial-temporal feature dataset into a low-dimensional vector of size M for transmission over the air to the decoding device. The encoding device may perform quantization before transmitting the low-dimensional vector of size M over the air to map the transmitted samples to discrete values ​​for the low-dimensional vector of size M.

[0190] The decoding device may perform an Mx64V fully connected operation to decompress a low-dimensional vector of size M into a space-time feature dataset. The decoding device may perform a reshape operation to reshape the 64V element vector into a 2-dimensional Vx64 matrix. The decoding device may perform a VxW (with a kernel of 1) convolution operation on the output from the reshape operation. The VxW convolution operation may include a point-by-point (e.g., tap-by-tap) convolution operation. The VxW convolution operation may decompress the spatial features from the reduced dimension for each tap. The VxW convolution operation has an input of V features and an output of W features. The output from the VxW convolution operation may be a Wx64 matrix.

[0191] The decoding device may perform one or more ResNet operations. The one or more ResNet operations may further decompress spatial features and / or temporal features. In some aspects, the ResNet operation may include multiple (e.g., 3) 1-dimensional convolution operations, skip connections (e.g., to avoid applying 1-dimensional convolution operations), summation operations of paths through multiple convolution operations and paths through skip connections, etc. The output from the one or more ResNet operations may be a Wx64 matrix.

[0192] The decoding device can perform spatial and temporal feature reconstruction. In some aspects, this can be achieved by using a 1-dimensional convolution operation that is fully connected in the spatial dimension (to reconstruct spatial features) and a simple convolution with a small kernel size (e.g., 3) in the tap dimension (to reconstruct short tap features). The output from the 64xW convolution operation can be a 64x64 matrix. In some aspects, the values ​​of M, W, and / or V can be configurable to adjust the weights of features, payload size, etc.

[0193] Example 600 may support the reuse of one or more trained layers of a neural network to process other signals or channels. Example 600 may include examples of a neural network at an encoding device and a neural network at a decoding device that may be used for neural network-based CSI reporting. In some cases, the processes or layers described in Example 600 may generate a set of weights that can be reused to accelerate processing of different signals or channels. For example, a decoding device or another network entity may transmit an indication (such as an NNCI) including a source identifier and a target identifier for an associated signal, channel, or layer to an encoding device. The NNCI may indicate that the encoding device may reuse one or more trained layers or neural networks to decode, demodulate, estimate, compress, train, or any combination thereof the indicated signal or channel (e.g., corresponding to the target identifier). The source identifier may be associated with a first procedure, a first set of signals, a first layer of a neural network, or a first neural network. The target identifier may be associated with a second set of signals or a second procedure. The NNCI may indicate that the trained layer or trained neural network associated with the source identifier may be reused to process a second set of signals or a second procedure indicated by the target identifier.

[0194] As indicated above, Figure 6 These are provided as examples only. Other examples may differ from those regarding Figure 6 Examples described.

[0195] Figure 7 An example of an example 700 supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated. In some examples, example 700 can implement aspects of wireless communication system 100.

[0196] An encoding device (e.g., UE 115, encoding device 301, etc.) may be configured to perform one or more operations on samples (e.g., data) received via one or more antennas of the encoding device to compress the samples. A decoding device (e.g., base station 105, decoding device 302, etc.) may be configured to decode the compressed samples to determine information, such as a CSF. As shown by example 700, features may be compressed and decompressed in sequence. For example, an encoding device may extract and compress features associated with an input to produce a payload, and then a decoding device may extract and compress features associated with the payload to reconstruct the input. The encoding and decoding operations may be symmetric (as shown) or asymmetric.

[0197] As shown by example 700, the encoding device may receive samples from the antennas. For example, the encoding device may receive a data set of size 256x64 based at least in part on the number of antennas, the number of samples per antenna, and the tap characteristics. The encoding device may reshape the data into a (64x64x4) data set.

[0198] The encoding device may perform a 2-dimensional 64x128 convolution operation (with kernel sizes of 3 and 1). In some aspects, the 64x128 convolution operation may perform spatial feature extraction associated with the decoding device antenna dimension, short-term (tap) feature extraction associated with the decoding device (e.g., base station) antenna dimension, etc. In some aspects, this may be achieved by using a 2D convolution layer that is fully connected in the decoding device antenna dimension, has a small kernel size (e.g., 3) in the tap dimension, and a simple convolution operation with a small kernel size (e.g., 1) in the encoding device antenna dimension. The output from the 64xW convolution operation may be a matrix of size (128x64x4).

[0199] The encoding device may perform one or more ResNet operations. The one or more ResNet operations may further refine the spatial features associated with the decoding device and / or the temporal features associated with the decoding device. In some aspects, the ResNet operation may include multiple operations associated with the features. For example, the ResNet operation may include multiple (e.g., 3) 2D convolution operations, skip connections (e.g., between the input of the ResNet and the output of the ResNet to avoid applying a 2D convolution operation), a summation operation of a path through multiple 2D convolution operations and a path through a skip connection, etc. In some aspects, the plurality of 2D convolution operations may include: a Wx2W convolution operation with kernel sizes of 3 and 1, the output of which is input to a BN layer, followed by a LeakyReLU activation, which produces an output dataset of size 2Wx64xV; a 2Wx4W convolution operation with kernel sizes of 3 and 1, the output of which is input to a BN layer, followed by a LeakyReLU activation, which produces an output dataset of size 4Wx64xV; and a 4WxW convolution operation with kernel sizes of 3 and 1, the output of which has a BN dataset of size (128x64x4). The output from one or more ResNet operations may be a matrix of size (128x64x4).

[0200] The encoding device may perform a 2-dimensional 128xV convolution operation (with kernel sizes of 1 and 1) on the output from the one or more ResNet operations. The 128xV convolution operation may include a point-by-point (e.g., tap-by-tap) convolution operation. The WxV convolution operation may compress the spatial features associated with the decoding device into a reduced dimension for each tap. The output from the 128xV convolution operation may be a matrix of size (4x64xV).

[0201] The encoding device may perform a 2-dimensional 4x8 convolution operation (with kernel sizes of 3 and 1). In some aspects, the 4x8 convolution operation may perform spatial feature extraction associated with the encoding device antenna dimensions, short-term (tap) feature extraction associated with the encoding device antenna dimensions, etc. The output from the 4x8 convolution operation may be a matrix of size (8x64xV).

[0202] The encoding device may perform one or more ResNet operations. The one or more ResNet operations may further refine the spatial features associated with the encoding device and / or the temporal features associated with the encoding device. In some aspects, the ResNet operation may include multiple operations associated with the features. For example, the ResNet operation may include multiple (e.g., 3) 2D convolution operations, skip connections (e.g., to avoid applying 2D convolution operations), summation operations of paths through multiple 2D convolution operations and paths through skip connections, etc. The output from the one or more ResNet operations may be a matrix of size (8x64xV).

[0203] The encoding device may perform a 2-dimensional 8xU convolution operation (with kernel sizes of 1 and 1) on the output from the one or more ResNet operations. The 8xU convolution operation may include a point-by-point (e.g., tap-by-tap) convolution operation. The 8xU convolution operation may compress the spatial features associated with the decoding device into a reduced dimension for each tap. The output from the 128xV convolution operation may be a matrix of size (Ux64xV).

[0204] The encoding device may perform a flattening operation to flatten a matrix of size (Ux64xV) into a 64UV element vector. The encoding device may perform a 64UVxM fully connected operation to further compress the 2-dimensional space-time feature data set into a low-dimensional vector of size M for transmission over the air to the decoding device. The encoding device may perform quantization before transmitting the low-dimensional vector of size M over the air to map the transmitted samples to discrete values ​​for the low-dimensional vector of size M.

[0205] The decoding device may perform an Mx64UV fully connected operation to decompress the low-dimensional vector of size M into a spatial-temporal feature dataset. The decoding device may perform a reshape operation to reshape the 64UV element vector into a matrix of size (Ux64xV). The decoding device may perform a 2-dimensional Ux8 (with a 1,1 kernel) convolution operation on the output from the reshape operation. The Ux8 convolution operation may include a point-by-point (e.g., tap-by-tap) convolution operation. The Ux8 convolution operation may decompress the spatial features from the reduced dimension for each tap. The output from the Ux8 convolution operation may be a dataset of size (8x64xV).

[0206] The decoding device may perform one or more ResNet operations. The one or more ResNet operations may further decompress the spatial features and / or temporal features associated with the encoding device. In some aspects, the ResNet operation may include multiple (e.g., 3) 2D convolution operations, skip connections (e.g., to avoid applying 2D convolution operations), summation operations of paths through multiple 2D convolution operations and paths through skip connections, etc. The output from the one or more ResNet operations may be a dataset of size (8x64xV).

[0207] The decoding device may perform a 2-dimensional 8x4 convolution operation (with kernel sizes of 3 and 1). In some aspects, the 8x4 convolution operation may perform spatial feature reconstruction in the antenna dimension of the encoding device, as well as short-term feature reconstruction. The output from the 8x4 convolution operation may be a dataset of size (Vx64x4).

[0208] The decoding device may perform a 2-dimensional Vx128 (with a kernel of 1) convolution operation on the output from the 2-dimensional 8x4 convolution operation to reconstruct the tap features and spatial features associated with the decoding device. The Vx128 convolution operation may include a point-by-point (e.g., tap-by-tap) convolution operation. The Vx128 convolution operation may decompress the spatial features associated with the decoding device antenna from a reduced dimension for each tap. The output from the Ux8 convolution operation may be a matrix of size (128x64x4).

[0209] The decoding device may perform one or more ResNet operations. The one or more ResNet operations may further decompress spatial features and / or temporal features associated with the decoding device. In some aspects, the ResNet operation may include multiple (e.g., 3) 2D convolution operations, skip connections (e.g., to avoid applying 2D convolution operations), summation operations of paths through multiple 2D convolution operations and paths through skip connections, etc. The output from the one or more ResNet operations may be a matrix of size (128x64x4).

[0210] The decoding device may perform a 2-dimensional 128x64 convolution operation (with kernel sizes of 3 and 1). In some aspects, the 128x64 convolution operation may perform spatial feature reconstruction, short-term feature reconstruction, etc. associated with the antenna dimensions of the decoding device. The output from the 128x64 convolution operation may be a dataset of size (64x64x4).

[0211] In some aspects, the values ​​of M, V, and / or U may be configurable to adjust the weighting of features, payload size, etc. For example, the value of M may be 32, 64, 128, 256, or 512, the value of V may be 16, and / or the value of U may be 1.

[0212] Example 700 may support the reuse of one or more trained layers of a neural network to process other signals or channels. Example 700 may include examples of a neural network at an encoding device and a neural network at a decoding device that may be used for neural network-based CSI reporting. In some cases, the process or layer described in Example 600 may generate a set of weights that can be reused to efficiently process different signals or channels.

[0213] As indicated above, Figure 7 These are provided as examples only. Other examples may differ from those regarding Figure 7 Examples described.

[0214] Figure 8 An example of an example 800 supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated. In some examples, example 800 can implement aspects of wireless communication system 100.

[0215] The encoding device (e.g., UE 115, encoding device 301, etc.) can be configured to perform one or more operations on samples (e.g., data) received via one or more antennas of the encoding device to compress the samples. The decoding device (e.g., base station 105, decoding device 302, etc.) can be configured to decode the compressed samples to determine information, such as CSF. The encoding device and decoding device operations can be asymmetric. In other words, the decoding device can have a greater number of layers than the decoding device.

[0216] As shown by example 800, the encoding device may receive samples from the antennas. For example, the encoding device may receive a data set of size 64x64 based at least in part on the number of antennas, the number of samples per antenna, and the tap characteristics.

[0217] The encoding device may perform a 64xW convolution operation (with a kernel size of 1). In some aspects, the 64xW convolution operation may be fully connected in the antenna, may be a convolution in the tap, and so on. The output from the 64xW convolution operation may be a Wx64 matrix. The encoding device may perform one or more WxW convolution operations (with a kernel size of 1 or 3). The output from one or more WxW convolution operations may be a Wx64 matrix. The encoding device may perform a convolution operation (with a kernel size of 1). In some aspects, one or more WxW convolution operations may perform spatial feature extraction, short-term (tap) feature extraction, and the like. In some aspects, the WxW convolution operation may be a series of 1-dimensional convolution operations.

[0218] The encoding device may perform a flattening operation to flatten the Wx64 matrix into a 64W element vector. The encoding device may perform a 4096xM fully connected operation to further compress the spatial-temporal feature dataset into a small-size vector of size M for transmission over the air to the decoding device. The encoding device may perform quantization before transmitting the low-dimensional vector of size M over the air to map the transmitted samples to discrete values ​​for the low-dimensional vector of size M.

[0219] The decoding device may perform a 4096xM fully connected operation to decompress the low-dimensional vector of size M into a spatial-temporal feature dataset. The decoding device may perform a reshape operation to reshape the 6W element vector into a Wx64 matrix.

[0220] The decoding device may perform one or more ResNet operations. The one or more ResNet operations may decompress spatial features and / or temporal features. In some aspects, the ResNet operation may include multiple (e.g., 3) 1-dimensional convolution operations, skip connections (e.g., between the input of the ResNet and the output of the ResNet to avoid applying the 1-dimensional convolution operation), summation operations of paths through multiple 1-dimensional convolution operations and paths through the skip connections, etc. In some aspects, the multiple 1-dimensional convolution operations may include: a Wx256 convolution operation with a kernel size of 3, the output of which is input to a BN layer, followed by a LeakyReLU activation, which produces an output dataset of size 256x64; a 256x512 convolution operation with a kernel size of 3, the output of which is input to a BN layer, followed by a LeakyReLU activation, which produces an output dataset of size 512x64; and a 512xW convolution operation with a kernel size of 3, the output of which has a BN dataset of size Wx64. The output from one or more ResNet operations can be a Wx64 matrix.

[0221] The decoding device may perform one or more WxW convolution operations (with a kernel size of 1 or 3). The output from the one or more WxW convolution operations may be a Wx64 matrix. The encoding device may perform a convolution operation (with a kernel size of 1). In some aspects, the WxW convolution operation may perform spatial feature reconstruction, short-term (tap) feature reconstruction, etc. In some aspects, the WxW convolution operation may be a series of 1-dimensional convolution operations.

[0222] The encoding device may perform a Wx64 convolution operation (with a kernel size of 1). In some aspects, the Wx64 convolution operation may be a 1-dimensional convolution operation. The output from the 64xW convolution operation may be a 64x64 matrix. In some aspects, the values ​​of M and / or W may be configurable to adjust the weights of features, payload size, etc.

[0223] As indicated above, Figure 8These are provided as examples only. Other examples may differ from those regarding Figure 8 Examples described.

[0224] Figure 9 An example of an example process 900 supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is illustrated. In some examples, example process 900 can implement aspects of wireless communication system 100.

[0225] Example process 900 is an example process in which a first device (eg, an encoding device, a UE 115, Figures 12 to 15 An example of a device 1205, 1305, 1405, or 1505, etc.) performing operations associated with encoding a data set using a neural network.

[0226] like Figure 9 As shown in , in some aspects, process 900 may include encoding a data set using one or more extraction operations and compression operations associated with a neural network, the one or more extraction operations and compression operations being based at least in part on a feature set of the data set to produce a compressed data set (block 910). For example, a first device (e.g., communications manager 1215) may encode a data set using one or more extraction operations and compression operations associated with a neural network, the one or more extraction operations and compression operations being based at least in part on a feature set of the data set to produce a compressed data set, as described above.

[0227] like Figure 9 As further shown in , in some aspects, process 900 may include transmitting the compressed data set to the second device (block 920). For example, the first device (eg, using transmitter 1220) may transmit the compressed data set to the second device, as described above.

[0228] Process 900 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.

[0229] In a first aspect, the data set is based at least in part on sampling of one or more reference signals.

[0230] In a second aspect, alone or in combination with the first aspect, transmitting the compressed data set to the second device includes transmitting CSI feedback to the second device.

[0231] In a third aspect, alone or in combination with one or more of the first and second aspects, process 900 includes identifying a feature set of a data set, wherein the one or more extraction operations and compression operations include: a first type of operation performed on a dimension associated with a feature in the feature set of the data set, and a second type of operation performed on remaining dimensions associated with other features in the feature set of the data set, the second type of operation being different from the first type of operation.

[0232] In a fourth aspect, alone or in combination with one or more of the first to third aspects, the first type of operation comprises a one-dimensional fully connected layer operation, and the second type of operation comprises a convolution operation.

[0233] In a fifth aspect, either alone or in combination with one or more of the first to fourth aspects, the one or more extraction operations and compression operations include multiple operations including one or more of convolution operations, fully connected layer operations, or residual neural network operations.

[0234] In the sixth aspect, alone or in combination with one or more of the first to fifth aspects, the one or more extraction operations and compression operations include a first extraction operation and a first compression operation performed on a first feature in a feature set of a data set, and a second extraction operation and a second compression operation performed on a second feature in the feature set of the data set.

[0235] In a seventh aspect, alone or in combination with one or more of the first to sixth aspects, process 900 includes performing one or more additional operations on the intermediate data set output after performing the one or more extraction operations and compression operations.

[0236] In an eighth aspect, alone or in combination with one or more of the first to seventh aspects, the one or more additional operations include one or more of a quantization operation, a flattening operation, or a fully connected operation.

[0237] In a ninth aspect, alone or in combination with one or more of the first to eighth aspects, the feature set of the data set comprises one or more of spatial features or tap-domain features.

[0238] In the tenth aspect, alone or in combination with one or more of the first to ninth aspects, the one or more extraction operations and compression operations include one or more of the following: spatial feature extraction using a one-dimensional convolution operation, temporal feature extraction using a one-dimensional convolution operation, a residual neural network operation for refining the extracted spatial features, a residual neural network operation for refining the extracted temporal features, a point-by-point convolution operation for compressing the extracted spatial features, a point-by-point convolution operation for compressing the extracted temporal features, a flattening operation for flattening the extracted spatial features, a flattening operation for flattening the extracted temporal features, or a compression operation for compressing one or more of the extracted temporal features or the extracted spatial features into a low-dimensional vector for transmission.

[0239] In the eleventh aspect, alone or in combination with one or more of the first to tenth aspects, the one or more extraction operations and compression operations include: a first feature extraction operation associated with one or more features associated with the second device, a first compression operation for compressing one or more features associated with the second device, a second feature extraction operation associated with one or more features associated with the first device, and a second compression operation for compressing one or more features associated with the first device.

[0240] although Figure 9 Example blocks of process 900 are shown, but in some aspects, process 900 may include Figure 9 900. Additionally or alternatively, two or more blocks of process 900 may be executed in parallel.

[0241] Figure 10 An example of an example process 1000 for supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is illustrated. In some examples, the example process 1000 can implement aspects of the wireless communication system 100. The example process 1000 is a process in which a second device (e.g., a decoding device, a UE 115, a base station 105, Figures 12 to 15 Device 1205, 1305, 1405 or 1505, or Figures 16 to 19 An example of a device 1605, 1705, 1805, or 1905, etc.) performing operations associated with using a neural network to decode a data set.

[0242] like Figure 10 As shown in , in some aspects, process 1000 may include receiving a compressed data set from a first device (block 1010). For example, the second device (eg, using receiver 1210) may receive the compressed data set from the first device, as described above.

[0243] like Figure 10As further shown in FIG. 1 , in some aspects, process 1000 may include decoding the compressed data set using one or more decompression operations and reconstruction operations associated with the neural network, the one or more decompression operations and reconstruction operations based at least in part on the feature set of the compressed data set to produce a reconstructed data set (block 1020). For example, the second device (e.g., using the communication manager 1215) may decode the compressed data set using one or more decompression operations and reconstruction operations associated with the neural network, the one or more decompression operations and reconstruction operations based at least in part on the feature set of the compressed data set to produce a reconstructed data set, as described above.

[0244] Process 1000 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.

[0245] In a first aspect, decoding a compressed data set using one or more decompression operations and reconstruction operations includes performing the one or more decompression operations and reconstruction operations based at least in part on an assumption that the first device generated the compressed data set using a set of operations that are symmetric to the one or more decompression operations and reconstruction operations, or performing the one or more decompression operations and reconstruction operations based at least in part on an assumption that the first device generated the compressed data set using a set of operations that are asymmetric to the one or more decompression operations and reconstruction operations.

[0246] In a second aspect, alone or in combination with the first aspect, the compressed data set is based at least in part on sampling of one or more reference signals by the first device.

[0247] In a third aspect, alone or in combination with one or more of the first and second aspects, receiving the compressed data set includes receiving CSI feedback from the first device.

[0248] In a fourth aspect, alone or in combination with one or more of the first to third aspects, the one or more decompression operations and reconstruction operations include: a first type of operation performed on a dimension associated with a feature in a feature set of the compressed data set, and a second type of operation performed on the remaining dimensions associated with other features in the feature set of the compressed data set, the second type of operation being different from the first type of operation.

[0249] In a fifth aspect, alone or in combination with one or more of the first to fourth aspects, the first type of operation comprises a one-dimensional fully connected layer operation, and wherein the second type of operation comprises a convolution operation.

[0250] In a sixth aspect, either alone or in combination with one or more of the first to fifth aspects, the one or more decompression operations and reconstruction operations include multiple operations including one or more of convolution operations, fully connected layer operations, or residual neural network operations.

[0251] In the seventh aspect, alone or in combination with one or more of the first to sixth aspects, the one or more decompression operations and reconstruction operations include: a first operation performed on a first feature in a feature set of the compressed data set, and a second operation performed on a second feature in the feature set of the compressed data set.

[0252] In an eighth aspect, alone or in combination with one or more of the first to seventh aspects, process 1000 includes performing a reshape operation on the compressed data set.

[0253] In a ninth aspect, alone or in combination with one or more of the first to eighth aspects, the feature set of the compressed data set includes one or more of spatial features or tap-domain features.

[0254] In a tenth aspect, either alone or in combination with one or more of the first to ninth aspects, the one or more decompression operations and reconstruction operations include one or more of a feature decompression operation, a temporal feature reconstruction operation, or a spatial feature reconstruction operation.

[0255] In the eleventh aspect, alone or in combination with one or more of the first to tenth aspects, the one or more decompression operations and reconstruction operations include: a first feature reconstruction operation performed on one or more features associated with the first device, and a second feature reconstruction operation performed on one or more features associated with the second device.

[0256] although Figure 10 Example blocks of process 1000 are shown, but in some aspects, process 1000 may include Figure 10 1000. Additionally or alternatively, two or more blocks of process 1000 may be executed in parallel.

[0257] Figure 11An example of a process flow 1100 for supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is illustrated. In some examples, process flow 1100 can implement aspects of wireless communication system 100. Process flow 1100 can include UE 115-b and base station 105-b, which can be respective examples of UE 115 and base station 105. In some cases, UE 115-b can be an example of an encoding device described herein, and base station 105-b can be an example of a decoding device described herein. In some additional or alternative examples, UE 115, a server, a transmission point, or another network entity can be an example of a decoding device. For example, process flow 1100 can be implemented by two UEs 115, a UE 115 and a server, a UE 115 and a transmission point, a UE 115 and a network entity, or any combination thereof, wherein UE 115 can receive NNCI and process signals of any one or more of the described devices or entities based on the NNCI.

[0258] At 1105, base station 105-b may transmit a first set of one or more signals to UE 115-b. UE 115-b may train a first set of layers of a neural network based on channel estimation using the resource set. For example, UE 115-b may train the first set of layers of the neural network based on measurements, estimates, etc. of the first set of one or more signals. UE 115-b may generate a set of weights for the first set of layers of the neural network based on the training. In some cases, the set of weights may correspond to a compressed representation of the channel. For example, the set of weights may be used to reconstruct the channel as received by UE 115-b. In some cases, the channel may be an example of a downlink channel from base station 105-b to UE 115-b. In some other examples, the channel may be an example of a sidelink channel between UE 115-b and another UE 115. For example, these techniques may be applicable to sidelink communications (such as in a V2X system).

[0259] At 1115, UE 115-b may transmit CSI to base station 105-b based on the first set of layers of the trained neural network. For example, UE 115-b may transmit a neural network-based CSI report to base station 105-b. The neural network-based CSI report may include at least a set of weights corresponding to the channel. Base station 105-b may be able to reconstruct the channel based on the set of weights and identify channel characteristics and interference at UE 115-b on the channel.

[0260] At 1120, base station 105-b may transmit NNCI to UE 115-b. The NNCI may indicate an association between a first set of one or more signals and a second set of one or more signals based on a set of trained layers of the neural network. For example, base station 105-b may transmit NNCI to indicate that UE 115-b may reuse one or more trained layers or neural networks for the purpose of decoding, demodulating, estimating, or compressing another signal or channel. The NNCI may indicate an association between reference signals or channels, or an association between a layer or neural network and certain signals, processes, or channels. If UE 115-b is aware of the association, UE 115-b may avoid a cold start when training the neural network or layer using the second set of one or more signals. UE 115-b may save processing power and latency by reusing previously trained layers or neural networks.

[0261] At 1125, base station 105-b may transmit a second set of one or more signals to UE 115-b. In some cases, base station 105-b may transmit both the NNCI and the second set of one or more signals. In some other examples, different devices or network entities may transmit the NNCI and the second set of one or more signals.

[0262] At 1130 , UE 115 - b may process the second set of one or more signals using the set of weights for the first set of layers based on a correlation between the first set of one or more signals and the second set of one or more signals.

[0263] In one example, the first set of one or more signals may be CSI-RS and the second set of one or more signals may be DMRS. UE 115-b may have trained one or more layers of a neural network on CSI-RS, and the NNCI may indicate that one or more trained layers may be reused to process DMRS. UE 115-b may receive the DMRS and process the DMRS based on the one or more trained layers. For example, UE 115-b may demodulate the DMRS using a set of weights of a neural network trained by the CSI-RS. This may reduce the amount of time UE 115-b will spend demodulating the DMRS. Additionally or alternatively, UE 115-b may train one or more layers of a neural network based on DMRS. By reusing some weights of the CSI-RS neural network, UE 115-b may increase the rate at which the DMRS neural network is trained.

[0264] Figure 12A block diagram 1200 of a device 1205 supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The device 1205 can be an example of aspects of a UE 115 as described herein. The device 1205 may include a receiver 1210, a communication manager 1215, and a transmitter 1220. The device 1205 may also include a processor. Each of these components may be in communication with each other (e.g., via one or more buses).

[0265] The receiver 1210 may receive information such as packets, user data, or control information associated with various information channels (e.g., control channels, data channels, and information related to neural network or layer configuration indicators for CSI schemes, etc.). The information may be passed to other components of the device 1205. The receiver 1210 may be a reference Figure 15 Examples of various aspects of the described transceiver 1515. The receiver 1210 may utilize a single antenna or a collection of antennas.

[0266] The communication manager 1215 may train a first set of layers of a neural network based on the channel estimate using the set of resources; generate a set of weights for the first set of layers of the neural network based on the training; receive an indication of an association between a first set of one or more signals and a second set of one or more signals from a first network entity based on the first set of layers of the neural network; receive a second set of one or more signals from a second network entity; and process the second set of one or more signals using the set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals. The communication manager 1215 may be an example of aspects of the communication manager 1510 described herein.

[0267] The communication manager 1215 may be an example of a means for performing various aspects of the neural network or layer configuration indicator or CSI scheme described herein. The communication manager 1215 or its subcomponents may be implemented in hardware, code (e.g., software or firmware) executed by a processor, or any combination thereof. If implemented in code executed by a processor, the functions of the communication manager 1215 or its subcomponents may be performed by a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device designed to perform the functions described in this disclosure, discrete gate or transistor logic, discrete hardware components, or any combination thereof.

[0268] In some examples, the communication manager 1215 may be configured to perform various operations (eg, training, receiving, determining, transmitting, processing) using or otherwise coordinating with the receiver 510, the transmitter 520, or both.

[0269] The communication manager 1215 or its subcomponents can be physically located at various locations, including being distributed such that portions of functionality are implemented by one or more physical components at different physical locations. In some examples, according to various aspects of the present disclosure, the communication manager 1215 or its subcomponents can be separate and distinct components. In some examples, according to various aspects of the present disclosure, the communication manager 1215 or its subcomponents can be combined with one or more other hardware components (including, but not limited to, input / output (I / O) components, a transceiver, a network server, another computing device, one or more other components described in the present disclosure, or a combination thereof).

[0270] The transmitter 1220 may transmit signals generated by other components of the device 1205. In some examples, the transmitter 1220 may be co-located with the receiver 1210 in a transceiver module. For example, the transmitter 1220 may be a reference Figure 15 Examples of various aspects of the described transceiver 1515. The transmitter 1220 may utilize a single antenna or a collection of antennas.

[0271] Figure 13 A block diagram 1300 of a device 1305 supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The device 1305 can be an example of aspects of the device 1205 or UE 115 as described herein. The device 1305 may include a receiver 1310, a communication manager 1315, and a transmitter 1345. The device 1305 may also include a processor. Each of these components may be in communication with each other (e.g., via one or more buses).

[0272] The receiver 1310 may receive information such as packets, user data, or control information associated with various information channels (e.g., control channels, data channels, and information related to neural network or layer configuration indicators for CSI schemes, etc.). The information may be passed to other components of the device 1305. The receiver 1310 may be a reference Figure 15 Examples of various aspects of the described transceiver 1515. The receiver 1310 may utilize a single antenna or a collection of antennas.

[0273] The communication manager 1315 can be an example of aspects of the communication manager 1215 as described herein. The communication manager 1315 can include a layer training component 1320, a weight generation component 1325, an association indication receiving component 1330, an association signal receiving component 1335, and an association signal processing component 1340. The communication manager 1315 can be an example of aspects of the communication manager 1510 as described herein.

[0274] The association signal receiving component 1335 can receive a second set of one or more signals from the second network entity. The layer training component 1320 can train a first set of layers of the neural network based on the channel estimate using the resource set. The association signal processing component 1340 can process the second set of one or more signals using a set of weights for the first set of layers based on a correlation between the first set of one or more signals and the second set of one or more signals. The weight generating component 1325 can generate a set of weights for the first set of layers of the neural network based on the training. The association indication receiving component 1330 can receive an indication of an association between the first set of one or more signals and the second set of one or more signals from the first network entity based on the first set of layers of the neural network.

[0275] The transmitter 1345 may transmit signals generated by other components of the device 1305. In some examples, the transmitter 1345 may be co-located with the receiver 1310 in a transceiver module. Figure 15 Examples of various aspects of the described transceiver 1515. The transmitter 1345 may utilize a single antenna or a collection of antennas.

[0276] Figure 14 A block diagram 1400 of a communication manager 1405 supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown. The communication manager 1405 can be an example of aspects of the communication manager 1215, the communication manager 1315, or the communication manager 1510 described herein. The communication manager 1405 can include a layer training component 1410, a weight generation component 1415, an association indication receiving component 1420, an association signal receiving component 1425, an association signal processing component 1430, and a layer state capability component 1435. Each of these modules can communicate directly or indirectly with each other (e.g., via one or more buses).

[0277] Layer training component 1410 can train a first set of layers of a neural network based on the channel estimate using the resource set. In some cases, the first set of one or more signals includes one or more of a CSI-RS, an SSB, a PRS, a DMRS, a tracking signal, a data channel, or a control channel. In some cases, the first set of layers of the neural network includes one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

[0278] Weight generating component 1415 can generate a set of weights for a first set of layers of the neural network based on the training. Association indication receiving component 1420 can receive, from a first network entity, an indication of an association between the first set of one or more signals and the second set of one or more signals based on the first set of layers of the neural network. In some cases, the indication of the association includes a source identifier and a target identifier. In some cases, the source identifier includes an identifier of the neural network. In some cases, the target identifier includes an identifier of the second set of one or more signals, a protocol for the second set of one or more signals, an identifier of the second neural network, or any combination thereof.

[0279] In some cases, the source identifier comprises an identifier of a signal or protocol corresponding to at least a first set of one or more signals. In some cases, the source identifier comprises an identifier of at least a first set of layers of a neural network.

[0280] In some cases, the indication of the association is received via a higher layer signal, a MAC CE, downlink control information, or both.In some cases, the second network entity is another UE, a base station, a transmission and reception point, a server, the first network entity, or any combination thereof.

[0281] The association signal receiving component 1425 can receive a second set of one or more signals from the second network entity. In some cases, the second set of one or more signals includes one or more of a CSI-RS, an SSB, a PRS, a DMRS, a tracking signal, a data channel, or a control channel.

[0282] In some cases, the first set of one or more signals corresponds to a first component carrier, and the second set of one or more signals corresponds to a second component carrier. In some cases, the first set of one or more signals corresponds to a first frequency band, and the second set of one or more signals corresponds to a second frequency band. In some cases, the first set of one or more signals corresponds to a first frequency band combination, and the second set of one or more signals corresponds to a second frequency band combination. In some cases, the first set of one or more signals corresponds to a first frequency range, and the second set of one or more signals corresponds to a second frequency range.

[0283] The associated signal processing component 1430 may process the second set of one or more signals using the set of weights for the first layer set based on the correlation between the first set of one or more signals and the second set of one or more signals. In some examples, the associated signal processing component 1430 may decode the second set of one or more signals using the set of weights for the first layer set based on the correlation between the first set of one or more signals and the second set of one or more signals. In some examples, the associated signal processing component 1430 may demodulate the second set of one or more signals using the set of weights for the first layer set based on the correlation between the first set of one or more signals and the second set of one or more signals.

[0284] In some examples, the correlation signal processing component 1430 may estimate a channel from the second set of one or more signals using a set of weights for the first set of layers based on a correlation between the first set of one or more signals and the second set of one or more signals. In some examples, the correlation signal processing component 1430 may compress the second set of one or more signals using a set of weights for the first set of layers based on a correlation between the first set of one or more signals and the second set of one or more signals. In some examples, the correlation signal processing component 1430 may train a first set of layers of the neural network, a second set of layers of the neural network, or both based on compressing the second set of one or more signals. In some examples, the correlation signal processing component 1430 may train a set of layers of the second neural network using a set of weights for the first set of layers of the neural network based on a correlation between the first set of one or more signals and the second set of one or more signals.

[0285] In some examples, associated signal processing component 1430 can receive a third set of one or more signals from the second network entity. In some examples, associated signal processing component 1430 can process the third set of one or more signals using a set of weights.

[0286] The layer state capability component 1435 can transmit to the first network entity an indication of the number of layer states that the UE is capable of storing, tracking, training, processing, or any combination thereof for one or more of component carriers, frequency bands, frequency band combinations.

[0287] Figure 15A diagram of a system 1500 including a device 1505 that supports a neural network or layer configuration indicator for a CSI scheme in accordance with various aspects of the present disclosure is shown. The device 1505 may be an example of, or include components of, the device 1205, device 1305, or UE 115 as described herein. The device 1505 may include components for two-way voice and data communications, including components for transmitting and receiving communications, including a communication manager 1510, a transceiver 1515, an antenna 1520, a memory 1525, and a processor 1535. These components may be in electronic communication via one or more buses (e.g., bus 1540).

[0288] The communication manager 1510 may use a set of resources to train a first set of layers of a neural network based on a channel estimate; generate a set of weights for the first set of layers of the neural network based on the training; receive an indication of an association between a first set of one or more signals and a second set of one or more signals from a first network entity based on the first set of layers of the neural network; receive a second set of one or more signals from a second network entity; and process the second set of one or more signals using the set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals.

[0289] The transceiver 1515 can communicate bidirectionally via one or more antennas, wired or wireless links, as described above. For example, the transceiver 1515 can represent a wireless transceiver and can communicate bidirectionally with another wireless transceiver. The transceiver 1515 can also include a modem to modulate packets and provide the modulated packets to the antenna for transmission, and demodulate packets received from the antenna.

[0290] In some cases, a wireless device may include a single antenna 1520. However, in some cases, the device may have more than one antenna 1520, which may be capable of transmitting or receiving multiple wireless transmissions concurrently.

[0291] The memory 1525 may include random access memory (RAM) and read-only memory (ROM). The memory 1525 may store computer-readable, computer-executable code 1530 including instructions that, when executed, cause the processor to perform the various functions described herein. In some cases, the memory 1525 may include, among other things, a basic input / output system (BIOS), which may control basic hardware or software operations, such as interaction with peripheral components or devices.

[0292] The code 1530 may include instructions for implementing various aspects of the present disclosure, including instructions for supporting wireless communications. The code 1530 may be stored in a non-transitory computer-readable medium, such as system memory or other types of memory. In some cases, the code 1530 may not be directly executed by the processor 1535, but may cause the computer (e.g., when compiled and executed) to perform the functions described herein.

[0293] The processor 1535 may include an intelligent hardware device (e.g., a general-purpose processor, a DSP, a CPU, a microcontroller, an ASIC, an FPGA, a programmable logic device, discrete gate or transistor logic components, discrete hardware components, or any combination thereof). In some cases, the processor 1535 may be configured to operate a memory array using a memory controller. In other cases, the memory controller may be integrated into the processor 1535. The processor 1535 may be configured to execute computer-readable instructions stored in a memory (e.g., memory 1525) to cause the device 1505 to perform various functions (e.g., various functions or tasks supporting a neural network or layer configuration indicator for a CSI scheme).

[0294] Figure 16 A block diagram 1600 of a device 1605 supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The device 1605 can be an example of aspects of a base station 105 as described herein. The device 1605 may include a receiver 1610, a communication manager 1615, and a transmitter 1620. The device 1605 may also include a processor. Each of these components may be in communication with each other (e.g., via one or more buses).

[0295] The receiver 1610 may receive information such as packets, user data, or control information associated with various information channels (e.g., control channels, data channels, and information related to neural network or layer configuration indicators for CSI schemes, etc.). The information may be passed to other components of the device 1605. The receiver 1610 may be a reference Figure 19 Examples of various aspects of the described transceiver 1920. The receiver 1610 may utilize a single antenna or a collection of antennas.

[0296] The communication manager 1615 may receive an indication of a trained set of layers of the neural network from the UE based on the channel estimate on the resource set; identify a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers; transmit a second set of one or more signals to the UE; and transmit an indication of an association between the first set of one or more signals and the second set of one or more signals to the UE based on the trained set of layers of the neural network. The communication manager 1615 may be an example of aspects of the communication manager 1910 described herein.

[0297] The communication manager 1615 may be an example of a device for performing various aspects of the neural network or layer configuration indicator for the CSI scheme described herein. The communication manager 1615 or its subcomponents may be implemented in hardware, code executed by a processor (e.g., software or firmware), or any combination thereof. If implemented in code executed by a processor, the functions of the communication manager 1615 or its subcomponents may be performed by a general-purpose processor, DSP, ASIC, FPGA or other programmable logic device designed to perform the functions described in this disclosure, discrete gate or transistor logic, discrete hardware components, or any combination thereof.

[0298] The communication manager 1615 or its subcomponents may be implemented in hardware, in code executed by a processor (e.g., software or firmware), or any combination thereof. If implemented in code executed by a processor, the functions of the communication manager 1615 or its subcomponents may be performed by a general-purpose processor, a DSP, an ASIC, an FPGA or other programmable logic device designed to perform the functions described in this disclosure, discrete gate or transistor logic, discrete hardware components, or any combination thereof.

[0299] The communication manager 1615 or its subcomponents can be physically located at various locations, including being distributed such that portions of functionality are implemented by one or more physical components at different physical locations. In some examples, according to various aspects of the present disclosure, the communication manager 1615 or its subcomponents can be separate and distinct components. In some examples, according to various aspects of the present disclosure, the communication manager 1615 or its subcomponents can be combined with one or more other hardware components (including, but not limited to, input / output (I / O) components, a transceiver, a network server, another computing device, one or more other components described in the present disclosure, or a combination thereof).

[0300] In some examples, the communication manager 1615 may be configured to perform various operations (eg, receive, identify, determine, transmit) using or otherwise coordinating with the receiver 510, the transmitter 520, or both.

[0301] The transmitter 1620 may transmit signals generated by other components of the device 1605. In some examples, the transmitter 1620 may be co-located with the receiver 1610 in a transceiver module. For example, the transmitter 1620 may be a reference Figure 19 Examples of aspects of the described transceiver 1920. The transmitter 1620 may utilize a single antenna or a collection of antennas.

[0302] Figure 17A block diagram 1700 of a device 1705 supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The device 1705 can be an example of aspects of the device 1605 or base station 105 as described herein. The device 1705 may include a receiver 1710, a communication manager 1715, and a transmitter 1740. The device 1705 may also include a processor. Each of these components may be in communication with each other (e.g., via one or more buses).

[0303] The receiver 1710 may receive information such as packets, user data, or control information associated with various information channels (e.g., control channels, data channels, and information related to neural network or layer configuration indicators for CSI schemes, etc.). The information may be passed to other components of the device 1705. The receiver 1710 may be a reference Figure 19 Examples of various aspects of the described transceiver 1920. The receiver 1710 may utilize a single antenna or a collection of antennas.

[0304] The communication manager 1715 can be an example of aspects of the communication manager 1615 as described herein. The communication manager 1715 can include a trained layer indication component 1720, a weight identification component 1725, an association signaling component 1730, and an association indication component 1735. The communication manager 1715 can be an example of aspects of the communication manager 1910 as described herein.

[0305] A trained layer indicating component 1720 may receive an indication of a trained set of layers of the neural network from the UE based on the channel estimate on the resource set. A weight identifying component 1725 may identify a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers. An association signal transmitting component 1730 may transmit a second set of one or more signals to the UE. An association indicating component 1735 may transmit an indication of an association between the first set of one or more signals and the second set of one or more signals to the UE based on the trained set of layers of the neural network.

[0306] The transmitter 1740 may transmit signals generated by other components of the device 1705. In some examples, the transmitter 1740 may be co-located with the receiver 1710 in a transceiver module. For example, the transmitter 1740 may be a reference Figure 19 Examples of aspects of the described transceiver 1920. The transmitter 1740 may utilize a single antenna or a collection of antennas.

[0307] Figure 18A block diagram 1800 of a communication manager 1805 supporting a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown. The communication manager 1805 can be an example of aspects of the communication manager 1615, the communication manager 1715, or the communication manager 1910 described herein. The communication manager 1805 can include a trained layer indication component 1810, a weight identification component 1815, an association signaling component 1820, an association indication component 1825, and a layer state capability component 1830. Each of these modules can communicate directly or indirectly with each other (e.g., via one or more buses).

[0308] The trained layer indication component 1810 can receive an indication of a set of trained layers of the neural network from the UE based on the channel estimate on the resource set. In some cases, the first set of one or more signals includes one or more of CSI-RS, SSB, or PRS.

[0309] The weight identification component 1815 can identify a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers. The associated signal transmission component 1820 can transmit a second set of one or more signals to the UE. In some cases, the second set of one or more signals includes one or more of a CSI-RS, an SSB, or a PRS.

[0310] The association indication component 1825 can transmit an indication of an association between the first set of one or more signals and the second set of one or more signals to the UE based on the trained set of layers of the neural network. In some examples, the association indication component 1825 can transmit the indication of the association via a MAC CE, downlink control information, or both. In some cases, the indication of the association includes a source identifier and a target identifier.

[0311] In some cases, the source identifier comprises an identifier of a neural network. In some cases, the source identifier comprises an identifier of at least a set of trained layers of the neural network. In some cases, the source identifier comprises an identifier of a signal or a procedure corresponding to at least a first set of one or more signals. In some cases, the target identifier comprises an identifier of a second set of one or more signals, a procedure for the second set of one or more signals, or both.

[0312] In some cases, the first set of one or more signals corresponds to a first component carrier and the second set of one or more signals corresponds to a second component carrier. In some cases, the first set of one or more signals corresponds to a first frequency band and the second set of one or more signals corresponds to a second frequency band. In some cases, the first set of one or more signals corresponds to a first frequency range and the second set of one or more signals corresponds to a second frequency range. In some cases, the set of trained layers of the neural network includes one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

[0313] The layer state capability component 1830 can receive from the UE an indication of the number of layer states that the UE is capable of storing, tracking, training, processing, or any combination thereof for one or more of component carriers, frequency bands, frequency band combinations.

[0314] Figure 19 A diagram of a system 1900 including a device 1905 that supports a neural network or layer configuration indicator for a CSI scheme in accordance with aspects of the present disclosure is shown. Device 1905 may be an example of or include components of device 1605, device 1705, or base station 105 as described herein. Device 1905 may include components for two-way voice and data communications, including components for transmitting and receiving communications, including a communication manager 1910, a network communication manager 1915, a transceiver 1920, an antenna 1925, a memory 1930, a processor 1940, and an inter-station communication manager 1945. These components may be in electronic communication via one or more buses (e.g., bus 1950).

[0315] The communication manager 1910 may receive an indication of a trained set of layers of a neural network from the UE based on a channel estimate on a resource set; identify a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers; transmit a second set of one or more signals to the UE; and transmit an indication of an association between the first set of one or more signals and the second set of one or more signals to the UE based on the trained set of layers of the neural network.

[0316] The network communications manager 1915 may manage communications with the core network (eg, via one or more wired backhaul links). For example, the network communications manager 1915 may manage the delivery of data communications for client devices, such as one or more UEs 115.

[0317] The transceiver 1920 can communicate bidirectionally via one or more antennas, wired or wireless links, as described above. For example, the transceiver 1920 can represent a wireless transceiver and can communicate bidirectionally with another wireless transceiver. The transceiver 1920 can also include a modem to modulate packets and provide the modulated packets to the antenna for transmission, and to demodulate packets received from the antenna.

[0318] In some cases, a wireless device may include a single antenna 1925. However, in some cases, the device may have more than one antenna 1925, which may be capable of transmitting or receiving multiple wireless transmissions concurrently.

[0319] Memory 1930 may include RAM and ROM. Memory 1930 may store computer-readable, computer-executable code 1935 including instructions that, when executed, cause the processor to perform the various functions described herein. In some cases, memory 1930 may include, among other things, BIOS, which may control basic hardware or software operations, such as interaction with peripheral components or devices.

[0320] The code 1935 may include instructions for implementing various aspects of the present disclosure, including instructions for supporting wireless communications. The code 1935 may be stored in a non-transitory computer-readable medium, such as system memory or other types of memory. In some cases, the code 1935 may not be directly executed by the processor 1940, but may cause a computer (e.g., when compiled and executed) to perform the functions described herein.

[0321] The processor 1940 may include an intelligent hardware device (e.g., a general-purpose processor, a DSP, a CPU, a microcontroller, an ASIC, an FPGA, a programmable logic device, a discrete gate or transistor logic component, a discrete hardware component, or any combination thereof). In some cases, the processor 1940 may be configured to operate a memory array using a memory controller. In other cases, the memory controller may be integrated into the processor 1940. The processor 1940 may be configured to execute computer-readable instructions stored in a memory (e.g., memory 1930) to cause the device 1905 to perform various functions (e.g., various functions or tasks supporting a neural network or layer configuration indicator for a CSI scheme).

[0322] The inter-site communication manager 1945 can manage communications with other base stations 105 and can include a controller or scheduler for controlling communications with the UE 115 in coordination with the other base stations 105. For example, the inter-site communication manager 1945 can coordinate the scheduling of transmissions to the UE 115 for various interference mitigation techniques, such as beamforming or joint transmission. In some examples, the inter-site communication manager 1945 can provide an X2 interface within an LTE / LTE-A wireless communication network technology to provide communications between the base stations 105.

[0323] Figure 20 A flow chart illustrating a method 2000 for supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The operations of the method 2000 may be implemented by a UE 115 or components thereof as described herein. For example, the operations of the method 2000 may be implemented by a UE 115 or components thereof as described herein. Figures 12 to 15 In some examples, the UE may execute an instruction set to control the functional elements of the UE to perform the described functions. Additionally or alternatively, the UE may use dedicated hardware to perform various aspects of the described functions.

[0324] At 2005, the UE may train a first set of layers of a neural network based on the channel estimate using the resource set. The operations of 2005 may be performed according to the methods described herein. In some examples, aspects of the operations of 2005 may be performed as described with reference to Figures 12 to 15 The described layer training components are executed.

[0325] At 2010, the UE may generate a set of weights for a first set of layers of a neural network based on the training. The operations of 2010 may be performed according to the methods described herein. In some examples, aspects of the operations of 2010 may be performed as described with reference to Figures 12 to 15 The weight generation component described above is performed.

[0326] At 2015, the UE may receive an indication of an association between the first set of one or more signals and the second set of one or more signals from the first network entity. The operations of 2015 may be performed according to the methods described herein. In some examples, aspects of the operations of 2015 may be performed as described with reference to Figures 12 to 15 The described association instructs the receiving component to perform.

[0327] At 2020, the UE may receive a second set of one or more signals from the second network entity. The operations of 2020 may be performed according to the methods described herein. In some examples, aspects of the operations of 2020 may be performed as described with reference to Figures 12 to 15 The described associated signal receiving component is executed.

[0328] At 2025, the UE processes the second set of one or more signals using a set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals. The operations of 2025 may be performed according to the methods described herein. In some examples, aspects of the operations of 2025 may be as described with reference to Figures 12 to 15 The described associated signal processing components are performed.

[0329] Figure 21 A flow chart illustrating a method 2100 for supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The operations of the method 2100 may be implemented by a UE 115 or components thereof as described herein. For example, the operations of the method 2100 may be implemented by a UE 115 or components thereof as described herein. Figures 12 to 15 In some examples, the UE may execute an instruction set to control the functional elements of the UE to perform the described functions. Additionally or alternatively, the UE may use dedicated hardware to perform various aspects of the described functions.

[0330] At 2105, the UE may transmit to the base station an indication of the number of layer states that the UE is capable of storing, tracking, training, processing, or any combination thereof for one or more of the component carriers, frequency bands, frequency band combinations. The operations of 2105 may be performed according to the methods described herein. In some examples, aspects of the operations of 2105 may be performed as described with reference to Figures 12 to 15 The described layer state capability components are executed.

[0331] At 2110, the UE may train a first set of layers of a neural network based on the channel estimate using the resource set. The operations of 2110 may be performed according to the methods described herein. In some examples, aspects of the operations of 2110 may be performed as described with reference to Figures 12 to 15 The described layer training components are executed.

[0332] At 2115, the UE may generate a set of weights for a first set of layers of a neural network based on the training. The operations of 2115 may be performed according to the methods described herein. In some examples, aspects of the operations of 2115 may be as described with reference to Figures 12 to 15 The weight generation component described above is performed.

[0333] At 2120, the UE may receive, from the first network entity, an indication of an association between the first set of one or more signals and the second set of one or more signals based on the first set of layers of the neural network. The operations of 2120 may be performed according to the methods described herein. In some examples, aspects of the operations of 2120 may be performed as described with reference to Figures 12 to 15 The described association instructs the receiving component to perform.

[0334] At 2125, the UE may receive a second set of one or more signals from the second network entity. The operations of 2125 may be performed according to the methods described herein. In some examples, aspects of the operations of 2125 may be performed as described with reference to Figures 12 to 15 The described associated signal receiving component is executed.

[0335] At 2130, the UE processes the second set of one or more signals using a set of weights for the first set of layers based on the association between the first set of one or more signals and the second set of one or more signals. The operations of 2130 may be performed according to the methods described herein. In some examples, aspects of the operations of 2130 may be as described with reference to Figures 12 to 15 The described associated signal processing components are performed.

[0336] Figure 22 A flow chart illustrating a method 2200 for supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The operations of the method 2200 may be implemented by a base station 105 or components thereof as described herein. For example, the operations of the method 2200 may be implemented by a base station 105 or components thereof as described herein. Figures 16 to 19 In some examples, a base station may execute an instruction set to control functional elements of the base station to perform the described functions. Additionally or alternatively, the base station may use dedicated hardware to perform various aspects of the described functions.

[0337] At 2205, the base station may receive an indication of a set of trained layers of a neural network from the UE based on the channel estimate on the resource set. The operations of 2205 may be performed according to the methods described herein. In some examples, aspects of the operations of 2205 may be performed as described with reference to Figures 16 to 19 The described trained layers instruct the components to perform.

[0338] At 2210, the base station may identify a set of weights for a trained set of layers of a neural network based on an indication of a trained set of layers. The operations of 2210 may be performed according to the methods described herein. In some examples, aspects of the operations of 2210 may be performed as described with reference to Figures 16 to 19 The weights described identify the components to perform.

[0339] At 2215, the base station may transmit an indication of an association between the first set of one or more signals and the second set of one or more signals to the UE based on the trained set of layers of the neural network. The operations of 2215 may be performed according to the methods described herein. In some examples, aspects of the operations of 2215 may be as described with reference to Figures 16 to 19 The described association instructs the components to perform.

[0340] At 2220, the base station may transmit a second set of one or more signals to the UE. The operations of 2220 may be performed according to the methods described herein. In some examples, aspects of the operations of 2220 may be performed as described with reference to Figures 16 to 19 The described associated signal transmission components are performed.

[0341] Figure 23 A flow chart illustrating a method 2300 for supporting a neural network or layer configuration indicator for a CSI scheme according to aspects of the present disclosure is shown. The operations of the method 2300 may be implemented by a base station 105 or components thereof as described herein. For example, the operations of the method 2300 may be implemented by a base station 105 or components thereof as described herein. Figures 16 to 19 In some examples, a base station may execute an instruction set to control functional elements of the base station to perform the described functions. Additionally or alternatively, the base station may use dedicated hardware to perform various aspects of the described functions.

[0342] At 2305, the base station may receive from the UE an indication of the number of layer states that the UE is capable of storing, tracking, training, processing, or any combination thereof for one or more of the component carriers, frequency bands, frequency band combinations. The operations of 2305 may be performed according to the methods described herein. In some examples, aspects of the operations of 2305 may be performed as described with reference to Figures 16 to 19 The described layer state capability components are executed.

[0343] At 2310, the base station may receive an indication of a set of trained layers of a neural network from the UE based on the channel estimate on the resource set. The operations of 2310 may be performed according to the methods described herein. In some examples, aspects of the operations of 2310 may be performed as described with reference to Figures 16 to 19 The described trained layers instruct the components to perform.

[0344] At 2315, the base station may identify a set of weights for the trained set of layers of the neural network based on the indication of the trained set of layers. The operations of 2315 may be performed according to the methods described herein. In some examples, aspects of the operations of 2315 may be performed as described with reference to Figures 16 to 19 The weights described identify the components to perform.

[0345] At 2320, the base station may transmit an indication of an association between the first set of one or more signals and the second set of one or more signals to the UE based on the trained set of layers of the neural network. The operations of 2320 may be performed according to the methods described herein. In some examples, aspects of the operations of 2320 may be as described with reference to Figures 16 to 19 The described association instructs the components to perform.

[0346] At 2325, the base station may transmit a second set of one or more signals to the UE. The operations of 2325 may be performed according to the methods described herein. In some examples, aspects of the operations of 2325 may be performed as described with reference to Figures 16 to 19 The described associated signal transmission components are performed.

[0347] It should be noted that the methods described herein describe possible implementations, and that the operations and steps may be rearranged or otherwise modified and other implementations are possible. Furthermore, aspects from two or more methods may be combined.

[0348] Example 1: A method for wireless communication at a UE, comprising: training a first set of layers of a neural network based at least in part on a channel estimate using a set of resources; generating a set of weights for the first set of layers of the neural network based at least in part on the training; receiving an indication of an association between a first group of one or more signals and a second group of one or more signals from a first network entity; receiving the second group of one or more signals from a second network entity; and processing the second group of one or more signals using the set of weights for the first set of layers based at least in part on the association between the first group of one or more signals and the second group of one or more signals.

[0349] Example 2: The method of Example 1 further comprises: transmitting to the base station an indication of the number of layer states that the UE can store, track, train, process, or any combination thereof for one or more of the component carriers, frequency bands, frequency band combinations.

[0350] Example 3: A method as in Example 1 or 2, wherein processing the second group of one or more signals includes: decoding the second group of one or more signals using a set of weights for the first layer set based at least in part on an association between the first group of one or more signals and the second group of one or more signals.

[0351] Example 4: A method as in any of Examples 1 to 3, wherein processing the second group of one or more signals includes: demodulating the second group of one or more signals using a set of weights for the first layer set based at least in part on an association between the first group of one or more signals and the second group of one or more signals.

[0352] Example 5: A method as in any of Examples 1 to 4, wherein processing the second group of one or more signals includes estimating a channel from the second group of one or more signals using a set of weights for a first layer set based at least in part on an association between the first group of one or more signals and the second group of one or more signals.

[0353] Example 6: A method as in any of Examples 1 to 5, wherein processing the second group of one or more signals includes: compressing the second group of one or more signals using a set of weights for the first set of layers based at least in part on an association between the first group of one or more signals and the second group of one or more signals; and training the first set of layers of the neural network, the second set of layers of the neural network, or both based at least in part on compressing the second group of one or more signals.

[0354] Example 7: A method as in any of Examples 1 to 6, wherein processing the second set of one or more signals includes: training a set of layers of a second neural network using a set of weights for a first set of layers of the neural network based at least in part on an association between the first set of one or more signals and the second set of one or more signals.

[0355] Example 8: A method as in any one of Examples 1 to 7, wherein the first group of one or more signals includes one or more of a channel state information reference signal, a synchronization signal block or a positioning reference signal, a demodulation reference signal, a tracking signal, a data channel, or a control channel.

[0356] Example 9: A method as in any one of Examples 1 to 8, wherein the second group of one or more signals includes one or more of a channel state information reference signal, a synchronization signal block or a positioning reference signal, a demodulation reference signal, a tracking signal, a data channel, or a control channel.

[0357] Example 10: The method of any of Examples 1 to 9, wherein the indication of the association includes a source identifier and a target identifier.

[0358] Example 11: The method of any of Examples 1 to 10, wherein the source identifier comprises an identifier of the neural network.

[0359] Example 12: The method of any of Examples 1 to 10, wherein the target identifier comprises an identifier of the second set of one or more signals, a protocol for the second set of one or more signals, or an identifier of the second neural network, or any combination.

[0360] Example 13: The method of any of Examples 1 to 10, wherein the source identifier comprises an identifier of a signal or a protocol corresponding to at least the first set of one or more signals.

[0361] Example 14: The method of any of Examples 1 to 13, wherein the source identifier comprises an identifier of at least a first set of layers of the neural network.

[0362] Example 15: The method of any one of Examples 1 to 14, wherein the indication of the association is received via a higher layer signal, a MAC CE, downlink control information, or both.

[0363] Example 16: The method of any of Examples 1 to 15, wherein the first set of one or more signals corresponds to a first component carrier and the second set of one or more signals corresponds to a second component carrier.

[0364] Example 17: The method of any of Examples 1 to 16, wherein the first set of one or more signals corresponds to a first frequency band and the second set of one or more signals corresponds to a second frequency band.

[0365] Example 18: The method of any of Examples 1 to 17, wherein the first set of one or more signals corresponds to a first frequency band combination and the second set of one or more signals corresponds to a second frequency band combination.

[0366] Example 19: The method of any of Examples 1 to 18, wherein the first set of one or more signals corresponds to a first frequency range and the second set of one or more signals corresponds to a second frequency range.

[0367] Example 20: The method of any one of Examples 1 to 19, wherein the first set of layers of the neural network includes one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

[0368] Example 21: The method of any of Examples 1 to 20, further comprising: receiving a third set of one or more signals from a second network entity; and processing the third set of one or more signals using the set of weights.

[0369] Example 22: The method of any one of Examples 1 to 21, wherein the second network entity is another UE, a base station, a transmission and reception point, a server, the first network entity, or any combination thereof.

[0370] Example 23: A method for wireless communication at a network entity, comprising: receiving an indication of a trained set of layers of a neural network from a user equipment (UE) based at least in part on a channel estimate on a set of resources; identifying a set of weights for the trained set of layers of the neural network based at least in part on the indication of the trained set of layers; transmitting to the UE an indication of an association between a first set of one or more signals and a second set of one or more signals based at least in part on the trained set of layers of the neural network; and transmitting the second set of one or more signals to the UE.

[0371] Example 24: The method of Example 23 further comprises: receiving from the UE an indication of a number of layer states that the UE is capable of storing, tracking, training, processing, or any combination thereof for one or more of the component carriers, frequency bands, frequency band combinations.

[0372] Example 25: The method of Example 23 or 24, wherein the first set of one or more signals includes one or more of a channel state information reference signal, a synchronization signal block, or a positioning reference signal.

[0373] Example 26: The method of any one of Examples 23 to 25, wherein the second set of one or more signals includes one or more of a channel state information reference signal, a synchronization signal block, or a positioning reference signal.

[0374] Example 27: The method of any one of Examples 23 to 26, wherein the indication of the association comprises a source identifier and a target identifier.

[0375] Example 28: The method of any of Examples 23 to 27, wherein the source identifier comprises an identifier of the neural network.

[0376] Example 29: The method of any of Examples 23 to 27, wherein the source identifier comprises an identifier of at least a set of trained layers of the neural network.

[0377] Example 30: The method of any of Examples 23 to 27, wherein the source identifier comprises an identifier of a signal or a procedure corresponding to at least the first set of one or more signals.

[0378] Example 31: The method of any of Examples 23 to 30, wherein the target identifier comprises an identifier of the second set of one or more signals, a procedure for the second set of one or more signals, or both.

[0379] Example 32: The method of any one of Examples 23 to 31, wherein transmitting the indication of the association comprises transmitting the indication of the association via a MAC CE, downlink control information, or both.

[0380] Example 33: The method of any of Examples 23 to 32, wherein the first set of one or more signals corresponds to a first component carrier and the second set of one or more signals corresponds to a second component carrier.

[0381] Example 34: The method of any of Examples 23 to 33, wherein the first set of one or more signals corresponds to a first frequency band and the second set of one or more signals corresponds to a second frequency band.

[0382] Example 35: The method of any of Examples 23 to 34, wherein the first set of one or more signals corresponds to a first frequency range and the second set of one or more signals corresponds to a second frequency range.

[0383] Example 36: The method of any of Examples 23 to 35, wherein the trained set of layers of the neural network comprises one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

[0384] Example 46: An apparatus for wireless communication, comprising at least one means for performing the method of any of Examples 1 to 22.

[0385] Example 47: An apparatus for wireless communication, comprising a processor and a memory coupled to the processor, the processor and the memory configured to perform the method of any one of Examples 1 to 22.

[0386] Example 48: A non-transitory computer-readable medium storing code for wireless communication, the code comprising instructions executable by a processor to perform the method of any of Examples 1 to 22.

[0387] Example 46: An apparatus for wireless communication, comprising at least one means for performing the method of any of Examples 23 to 36.

[0388] Example 47: An apparatus for wireless communication, comprising a processor and a memory coupled to the processor, the processor and the memory configured to perform the method of any one of Examples 23 to 36.

[0389] Example 48: A non-transitory computer-readable medium storing code for wireless communication, the code comprising instructions executable by a processor to perform the method of any of Examples 23 to 36.

[0390] Although aspects of LTE, LTE-A, LTE-A Pro, or NR systems may be described for example purposes, and LTE, LTE-A, LTE-A Pro, or NR terminology may be used throughout much of the description, the techniques described herein may also be applicable to networks other than LTE, LTE-A, LTE-A Pro, or NR networks. For example, the techniques described may be applicable to various other wireless communication systems, such as Ultra Mobile Broadband (UMB), Institute of Electrical and Electronics Engineers (IEEE) 802.11 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802.20, Flash-OFDM, and other systems and radio technologies not explicitly mentioned herein.

[0391] The information and signals described herein may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referred to throughout this description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

[0392] The various illustrative blocks and components described in conjunction with the disclosure herein may be implemented or performed with a general purpose processor, a DSP, an ASIC, a CPU, an FPGA or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices (e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration).

[0393] The functions described herein may be implemented in hardware, software executed by a processor, firmware, or any combination thereof. If implemented in software executed by a processor, each function may be stored on or transmitted by a computer-readable medium as one or more instructions or code. Other examples and implementations fall within the scope of this disclosure and the appended claims. For example, due to the nature of software, the functions described herein may be implemented using software executed by a processor, hardware, firmware, hardwiring, or any combination thereof. Features that implement the functions may also be physically located in various locations, including being distributed so that parts of the functions are implemented at different physical locations.

[0394] Computer-readable media include both non-transient computer storage media and communication media, which include any media that facilitates a computer program to be transferred from one place to another. Non-transient storage media can be any available medium that can be accessed by a general or special-purpose computer. As an example and not limitation, non-transient computer-readable media may include RAM, ROM, electrically erasable programmable ROM (EEPROM), flash memory, compact disc (CD) ROM or other optical disc storage, magnetic disk storage or other magnetic storage devices, or any other non-transient medium that can be used to carry or store the desired program code means of an instruction or data structure form and can be accessed by a general or special-purpose computer, or a general or special-purpose processor. Similarly, any connection is also properly referred to as a computer-readable medium. For example, if software is transmitted from a website, server or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwaves, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are included in the definition of computer-readable media. Disk and disc, as used herein, include CDs, laser discs, optical discs, digital versatile discs (DVDs), floppy disks, and Blu-ray discs, where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations of the above are also included within the scope of computer-readable media.

[0395] As used herein (including in the claims), "or" used in a list of items (e.g., a list of items followed by a phrase such as "at least one of" or "one or more of") indicates an inclusive list, so that, for example, a list of at least one of A, B, or C means A or B or C or AB or AC or BC or ABC (i.e., A and B and C). Likewise, as used herein, the phrase "based on" should not be read as referencing a closed set of conditions. For example, an example step described as "based on condition A" could be based on both condition A and condition B without departing from the scope of this disclosure. In other words, as used herein, the phrase "based on" should be read in the same manner as the phrase "based at least in part on."

[0396] In the accompanying drawings, similar components or features may have the same reference number. In addition, components of the same type may be distinguished by following the reference number with a dash and a second reference number that distinguishes between the similar components. If only the first reference number is used in the specification, the description applies to any of the similar components having the same first reference number, regardless of the second reference number or other subsequent reference numbers.

[0397] The description set forth herein in conjunction with the accompanying drawings describes example configurations and does not represent all examples that can be implemented or fall within the scope of the claims. The term "example" as used herein means "serving as an example, instance, or illustration" and does not mean "better than" or "better than other examples." This detailed description includes specific details to provide an understanding of the described techniques. However, these techniques can be practiced without these specific details. In some instances, known structures and devices are shown in block diagram form to avoid obscuring the concepts of the described examples.

[0398] The description herein is provided to enable one of ordinary skill in the art to make or use the present disclosure. Various modifications to the present disclosure will be readily apparent to those of ordinary skill in the art, and the general principles defined herein may be applied to other variations without departing from the scope of the present disclosure. Thus, the present disclosure is not limited to the examples and designs described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for wireless communication at a user equipment (UE), comprising: training a first set of layers of a neural network based at least in part on the channel estimate using the set of resources; generating a set of weights for the first set of layers of the neural network based at least in part on the training; receiving, from the first network entity, an indication of an association between the first set of one or more signals and the second set of one or more signals; receiving the second set of one or more signals from a second network entity; as well as Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, the second set of one or more signals is processed using the set of weights of the first set of layers.

2. The method of claim 1, further comprising: An indication of a number of layer states that the UE is capable of storing, tracking, training, processing, or any combination thereof for one or more of component carriers, frequency bands, frequency band combinations, is transmitted to the first network entity.

3. The method of claim 1 , wherein processing the second set of one or more signals comprises: Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, the second set of one or more signals is decoded using the set of weights for the first set of layers.

4. The method of claim 1 , wherein processing the second set of one or more signals comprises: Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, the second set of one or more signals is demodulated using the set of weights for the first set of layers.

5. The method of claim 1 , wherein processing the second set of one or more signals comprises: A downlink channel is estimated from the second set of one or more signals using the set of weights for the first set of layers based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals.

6. The method of claim 1 , wherein processing the second set of one or more signals comprises: compressing the second set of one or more signals using the set of weights for the first set of layers based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals; as well as The first set of layers of the neural network, the second set of layers of the neural network, or both are trained based at least in part on compressing the second set of one or more signals.

7. The method of claim 1 , wherein processing the second set of one or more signals comprises: Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, a set of layers of a second neural network is trained using the set of weights for the first set of layers of the neural network.

8. The method of claim 1, wherein the first set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block or positioning reference signal, a demodulation reference signal, a tracking signal, a data channel, or a control channel.

9. The method of claim 1, wherein the second set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block or positioning reference signal, a demodulation reference signal, a tracking signal, a data channel, or a control channel.

10. The method of claim 1, wherein the indication of the association comprises a source identifier and a target identifier.

11. The method of claim 10, wherein the source identifier comprises an identifier of the neural network.

12. The method of claim 10, wherein the target identifier comprises an identifier of the second set of one or more signals, a protocol for the second set of one or more signals, or an identifier of a second neural network, or any combination.

13. The method of claim 10, wherein the source identifier comprises an identifier of a signal or protocol corresponding to at least the first set of one or more signals, an identifier of at least the first set of layers of the neural network, or both.

14. The method of claim 1, wherein the indication of the association is received via a higher layer signal, a medium access control (MAC) control element (CE), downlink control information, or both.

15. The method of claim 1 , wherein the first group of one or more signals corresponds to a first component carrier and the second group of one or more signals corresponds to a second component carrier, or the first group of one or more signals corresponds to a first frequency band and the second group of one or more signals corresponds to a second frequency band, or the first group of one or more signals corresponds to a first frequency band combination and the second group of one or more signals corresponds to a second frequency band combination, or the first group of one or more signals corresponds to a first frequency range and the second group of one or more signals corresponds to a second frequency range, or any combination thereof.

16. The method of claim 1, wherein the first set of layers of the neural network comprises one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

17. The method of claim 1, further comprising: receiving a third set of one or more signals from the second network entity; as well as The third set of one or more signals is processed using the set of weights.

18. The method of claim 1, wherein the second network entity is another UE, a base station, a transmission and reception point, a server, the first network entity, or any combination thereof.

19. A method for wireless communication at a network entity, comprising: receiving an indication of a set of trained layers of a neural network based at least in part on the channel estimate over the set of resources; identifying a set of weights for the trained set of layers of the neural network based at least in part on the indication of the trained set of layers; transmitting an indication of an association between a first set of one or more signals and a second set of one or more signals based at least in part on the trained set of layers of the neural network; as well as The second set of one or more signals is transmitted.

20. The method of claim 19, further comprising: An indication of a number of layer states that a user equipment (UE) is capable of storing, tracking, training, processing, or any combination thereof for one or more of component carriers, frequency bands, frequency band combinations is received.

21. The method of claim 19, wherein the first set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block, or a positioning reference signal.

22. The method of claim 19, wherein the second set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block, or a positioning reference signal.

23. The method of claim 19, wherein the indication of the association comprises a source identifier and a target identifier.

24. The method of claim 23, wherein the source identifier comprises an identifier of the neural network, an identifier of at least the set of trained layers of the neural network, an identifier of a signal or protocol corresponding to at least the first set of one or more signals, or any combination thereof.

25. The method of claim 23, wherein the target identifier comprises an identifier of the second set of one or more signals, a protocol for the second set of one or more signals, or both.

26. The method of claim 19, wherein transmitting the indication of the association comprises: The indication of the association is transmitted via a medium access control (MAC) control element (CE), downlink control information, or both.

27. The method of claim 19, wherein the first group of one or more signals corresponds to a first component carrier and the second group of one or more signals corresponds to a second component carrier, or the first group of one or more signals corresponds to a first frequency band and the second group of one or more signals corresponds to a second frequency band, or the first group of one or more signals corresponds to a first frequency range and the second group of one or more signals corresponds to a second frequency range, or any combination thereof.

28. The method of claim 19, wherein the set of trained layers of the neural network comprises one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

29. An apparatus for wireless communication at a user equipment (UE), comprising: processor; as well as a memory coupled to the processor, the processor and the memory being configured to: training a first set of layers of a neural network based at least in part on the channel estimate using the set of resources; generating a set of weights for the first set of layers of the neural network based at least in part on the training; receiving, from the first network entity, an indication of an association between the first set of one or more signals and the second set of one or more signals; receiving the second set of one or more signals from a second network entity; as well as Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, the second set of one or more signals is processed using the set of weights for the first set of layers.

30. The apparatus of claim 29, wherein the processor and memory are further configured to: An indication of a number of layer states that the UE is capable of storing, tracking, training, processing, or any combination thereof for one or more of component carriers, frequency bands, frequency band combinations, is transmitted to the first network entity.

31. The apparatus of claim 29, wherein the processor and memory being configured to process the second set of one or more signals comprises the processor and memory being configured to: Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, the second set of one or more signals is decoded using the set of weights for the first set of layers.

32. The apparatus of claim 29, wherein the processor and memory being configured to process the second set of one or more signals comprises the processor and memory being configured to: Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, the second set of one or more signals is demodulated using the set of weights for the first set of layers.

33. The apparatus of claim 29, wherein the processor and memory being configured to process the second set of one or more signals comprises the processor and memory being configured to: A downlink channel is estimated from the second set of one or more signals using the set of weights for the first set of layers based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals.

34. The apparatus of claim 29, wherein the processor and memory being configured to process the second set of one or more signals comprises the processor and memory being configured to: compressing the second set of one or more signals using the set of weights for the first set of layers based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals; and The first set of layers of the neural network, the second set of layers of the neural network, or both are trained based at least in part on compressing the second set of one or more signals.

35. The apparatus of claim 29, wherein the processor and memory being configured to process the second set of one or more signals comprises the processor and memory being configured to: Based at least in part on the correlation between the first set of one or more signals and the second set of one or more signals, a set of layers of a second neural network is trained using the set of weights for the first set of layers of the neural network.

36. The apparatus of claim 29, wherein the first set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block or positioning reference signal, a demodulation reference signal, a tracking signal, a data channel, or a control channel.

37. The apparatus of claim 29, wherein the second set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block or positioning reference signal, a demodulation reference signal, a tracking signal, a data channel, or a control channel.

38. The apparatus of claim 29, wherein the indication of the association comprises a source identifier and a target identifier.

39. The apparatus of claim 38, wherein the source identifier comprises an identifier of the neural network.

40. The apparatus of claim 38, wherein the target identifier comprises an identifier of the second set of one or more signals, a protocol for the second set of one or more signals, or an identifier of a second neural network, or any combination.

41. The apparatus of claim 38, wherein the source identifier comprises an identifier of a signal or protocol corresponding to at least the first set of one or more signals, an identifier of at least the first set of layers of the neural network, or both.

42. The apparatus of claim 29, wherein the indication of the association is received via a higher layer signal, a medium access control (MAC) control element (CE), downlink control information, or both.

43. The apparatus of claim 29, wherein the first group of one or more signals corresponds to a first component carrier and the second group of one or more signals corresponds to a second component carrier, or the first group of one or more signals corresponds to a first frequency band and the second group of one or more signals corresponds to a second frequency band, or the first group of one or more signals corresponds to a first frequency band combination and the second group of one or more signals corresponds to a second frequency band combination, or the first group of one or more signals corresponds to a first frequency range and the second group of one or more signals corresponds to a second frequency range, or any combination thereof.

44. The apparatus of claim 29, wherein the first set of layers of the neural network comprises one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

45. The apparatus of claim 29, wherein the processor and memory are further configured to: receiving a third set of one or more signals from the second network entity; and The third set of one or more signals is processed using the set of weights.

46. ​​The apparatus of claim 29, wherein the second network entity is another UE, a base station, a transmission and reception point, a server, the first network entity, or any combination thereof.

47. An apparatus for wireless communication at a network entity, comprising: processor; as well as a memory coupled to the processor, the processor and the memory being configured to: receiving an indication of a set of trained layers of a neural network based at least in part on the channel estimate over the set of resources; identifying a set of weights for the trained set of layers of the neural network based at least in part on the indication of the trained set of layers; transmitting an indication of an association between a first set of one or more signals and a second set of one or more signals based at least in part on the trained set of layers of the neural network; as well as The second set of one or more signals is transmitted.

48. The apparatus of claim 47, wherein the processor and memory are further configured to: An indication of a number of layer states that a user equipment (UE) is capable of storing, tracking, training, processing, or any combination thereof for one or more of component carriers, frequency bands, frequency band combinations is received.

49. The apparatus of claim 47, wherein the first set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block, or a positioning reference signal.

50. The apparatus of claim 47, wherein the second set of one or more signals comprises one or more of a channel state information reference signal, a synchronization signal block, or a positioning reference signal.

51. The apparatus of claim 47, wherein the indication of the association comprises a source identifier and a target identifier.

52. The apparatus of claim 51 , wherein the source identifier comprises an identifier of the neural network, an identifier of at least the set of trained layers of the neural network, an identifier of a signal or protocol corresponding to at least the first set of one or more signals, or any combination thereof.

53. The apparatus of claim 51, wherein the target identifier comprises an identifier of the second set of one or more signals, a procedure for the second set of one or more signals, or both.

54. The apparatus of claim 47, wherein the processor and memory being configured to transmit the indication of the association comprises the processor and memory being configured to: The indication of the association is transmitted via a medium access control (MAC) control element (CE), downlink control information, or both.

55. The apparatus of claim 47, wherein the first group of one or more signals corresponds to a first component carrier and the second group of one or more signals corresponds to a second component carrier, or the first group of one or more signals corresponds to a first frequency band and the second group of one or more signals corresponds to a second frequency band, or the first group of one or more signals corresponds to a first frequency range and the second group of one or more signals corresponds to a second frequency range, or any combination thereof.

56. The apparatus of claim 47, wherein the set of trained layers of the neural network comprises one or more residual neural network layers or one or more convolutional neural network layers, or any combination thereof.

Images