Configurable metrics for channel state compression and feedback

Abstract

Methods, systems, and apparatus for conducting wireless communications are described. Generally, the techniques described at a user equipment (UE) provide for efficiently reporting channel state information (CSI) to a base station with an appropriate level of accuracy. Specifically, the base station may indicate to the UE an accuracy level for reporting the CSI. The UE may encode the CSI using a first neural network, and the base station may decode the CSI using a second neural network. The first neural network and the second neural network may form a neural network pair, and the UE may train the neural network pair based on the accuracy level indicated by the base station. For example, the base station may indicate a loss function that corresponds to the accuracy level that the UE is to use to report the CSI, and the UE may use the loss function to train the neural network pair.

Description

Technical Field

[0001] The following relates generally to wireless communications and, more particularly, to configurable metrics for channel state compression and feedback. Background Art

[0002] Wireless communication systems are widely deployed to provide various types of communication content, such as voice, video, packet data, messaging, broadcast, etc. These systems can 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-Advanced, LTE-A) systems, or LTE-APro systems), and fifth generation (5G) systems, which may be referred to as new radio (NR) systems. These systems may employ techniques 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).

[0003] A wireless multiple-access communication system may include one or more base stations or one or more network access nodes, each base station or node simultaneously supporting communication for multiple communication devices, which may also be referred to as user equipment (UE). In some wireless communication systems, the UE may be configured to report channel state information (CSI) to the base station to indicate the downlink channel condition, and the base station may use the CSI to improve the quality of downlink transmissions to the UE. For example, the CSI may include a channel quality indicator (CQI), and the base station may use the CQI to identify appropriate parameters (e.g., modulation and coding scheme (MCS)) for transmitting downlink data to the UE. Summary of the Invention

[0004] The described techniques relate to improved methods, systems, devices, and apparatuses that support configurable metrics for channel state compression and feedback. Generally, the techniques described at a user equipment (UE) provide for efficiently reporting channel state information (CSI) to a base station with an appropriate level of accuracy. Specifically, the base station may indicate to the UE an accuracy level for reporting the CSI. The UE may use a first neural network to encode the CSI, and the base station may use a second neural network to decode the CSI. The first neural network and the second neural network may form a neural network pair, and the UE may train the neural network pair based on the accuracy level indicated by the base station. For example, the base station may indicate a loss metric or function corresponding to the accuracy level at which the UE is to report the CSI, and the UE may use the loss metric or function to train the neural network pair. Using these techniques, the base station is able to configure the UE to report the CSI with an appropriate level of accuracy.

[0005] A method for wireless communication at a UE is described. The method may include: receiving, from a base station, an indication of an accuracy level for reporting channel state feedback to the base station; receiving downlink data or a reference signal from the base station; and reporting, to the base station based on the downlink data or the reference signal, the channel state feedback corresponding to the indicated accuracy level.

[0006] An apparatus for wireless communication at a user equipment terminal (UE) is described. The apparatus may include a processor, a memory coupled to the processor, and instructions stored in the memory. The instructions may be executed by the processor to cause the apparatus to: receive, from a base station, an indication of an accuracy level for reporting channel state feedback to the base station; receive downlink data or a reference signal from the base station; and report, to the base station based on the downlink data or the reference signal, the channel state feedback corresponding to the indicated accuracy level.

[0007] Another apparatus for wireless communication at a UE is described. The apparatus may include means for receiving, from a base station, an indication of an accuracy level for reporting channel state feedback to the base station; receiving downlink data or a reference signal from the base station; and reporting, to the base station based on the downlink data or the reference signal, the channel state feedback corresponding to the indicated accuracy level.

[0008] 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: receive, from a base station, an indication of an accuracy level for reporting channel state feedback to the base station; receive downlink data or a reference signal from the base station; and report, to the base station, the channel state feedback corresponding to the indicated accuracy level based on the downlink data or the reference signal.

[0009] In some examples of the methods, apparatus, and non-transitory computer-readable media described herein, receiving an indication of an accuracy level for reporting channel state feedback may include operations, features, components, or instructions for performing the following steps: receiving an indication of a loss function corresponding to the accuracy level for training a neural network pair, the neural network pair comprising a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder, the method further comprising training the neural network pair using the loss function. Because the UE can receive the indication of the loss function from the base station, the UE can report CSI feedback to the base station with an appropriate accuracy level.

[0010] In some examples of the methods, apparatus, and non-transitory computer-readable media described herein, training the neural network pair using the loss function may include operations, features, components, or instructions for iteratively entering a channel state feedback input into the neural network pair and identifying a channel state feedback output from the neural network pair; determining a difference between the channel state feedback input and the channel state feedback output for each iteration using the loss function, wherein the difference comprises a loss; and adjusting coefficients of the neural network pair for each iteration based on the determination to minimize the difference between the channel state feedback input and the channel state feedback output. Because the UE can train the neural network pair using the loss function corresponding to the indicated accuracy level, the UE can report CSI feedback with an appropriate level of accuracy to minimize unnecessary overhead.

[0011] In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, reporting the channel state feedback corresponding to the indicated level of accuracy may include operations, features, components, or instructions for: encoding the channel state feedback using the first neural network at the encoder based on the training; and reporting the encoded channel state feedback. Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for: transmitting coefficients of the second neural network to the base station for decoding the channel state feedback based on the training. Because the UE can transmit the coefficients of the second neural network for decoding the channel state feedback to the base station, the UE can train the neural network pair without exchanging signaling with the base station, thereby reducing overhead.

[0012] Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for: receiving an indication from the base station to train a set of neural network pairs based on a set of accuracy levels, the set of neural network pairs including the neural network pair; and training each neural network pair in the set of neural network pairs based on a corresponding accuracy level in the set of accuracy levels. In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, receiving the indication of the accuracy level may include operations, features, components, or instructions for: receiving an indication to use the neural network pair in the set of neural network pairs for reporting the channel state feedback. Because the UE can train a set of neural network pairs, the base station can dynamically indicate the accuracy level to the UE, and the UE can select a neural network pair in the set of neural network pairs corresponding to the accuracy level to use for encoding the channel state feedback. That is, after receiving the indication of the accuracy level, the UE can avoid training the neural network pair based on the indicated accuracy level, thereby reducing latency.

[0013] Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for autonomously selecting the neural network pair from the set of neural network pairs for reporting the channel state feedback. Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for receiving an indication of a subset of the set of neural network pairs for training by the UE. Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for receiving data from the base station on a subband or spatial layer or according to channel taps based on reporting the channel state feedback corresponding to the indicated accuracy level. Using these techniques, based on one or more configurations used by the base station to send data to the UE, the base station can adjust the accuracy level used by the UE to report CSI feedback.

[0014] Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for performing the following steps: receiving a retransmission of the same data that the UE failed to decode based on reporting the channel state feedback corresponding to the indicated accuracy level. Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for performing the following steps: identifying a number of bits for reporting channel state feedback based on the accuracy level, wherein the number of bits may be directly related to the accuracy level; and reporting the channel state feedback corresponding to the indicated accuracy level using the identified number of bits.

[0015] Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for performing the following steps: receiving an indication of the number of bits used to report the channel state feedback based on the accuracy level. In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, receiving the indication of the accuracy level may include operations, features, components, or instructions for performing the following steps: receiving the indication of the accuracy level in Radio Resource Control (RRC) signaling or a MAC Control Element (MAC-CE). Because the UE can identify or select the number of bits used to report the channel state feedback based on the accuracy level, the overhead of reporting the channel state feedback can be minimized when appropriate.

[0016] A method for wireless communication at a base station is described. The method may include: sending an indication of an accuracy level for reporting channel state feedback to a UE to the base station; sending downlink data or a reference signal to the UE; and receiving channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE.

[0017] An apparatus for wireless communication at a base station is described. The apparatus may include a processor, a memory coupled to the processor, and instructions stored in the memory. The instructions may be executed by the processor to cause the apparatus to: send an indication of an accuracy level for reporting channel state feedback to a UE; send downlink data or a reference signal to the UE; and receive channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE.

[0018] Another apparatus for wireless communication at a base station is described. The apparatus may include means for: sending an indication of an accuracy level for reporting channel state feedback to a UE to the base station; sending downlink data or a reference signal to the UE; and receiving channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE.

[0019] A non-transitory computer-readable medium storing code for wireless communication at a base station is described. The code may include instructions executable by a processor to: send an indication of an accuracy level for reporting channel state feedback to a UE; send downlink data or a reference signal to the UE; and receive channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE.

[0020] In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, sending the indication of the accuracy level for reporting channel state feedback may include operations, features, components, or instructions for performing the following steps: sending an indication of a loss function used by the UE to train a neural network for reporting the channel state feedback. Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for performing the following steps: receiving coefficients of a neural network used at a decoder to decode the channel state feedback from the UE from the UE; and decoding the channel state feedback from the UE using the neural network at the decoder.

[0021] Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for transmitting an indication that the UE trains a set of neural network pairs based on a set of accuracy levels, each neural network pair including a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder. In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, transmitting the indication of the accuracy level may include operations, features, components, or instructions for transmitting an indication that the UE uses a neural network pair from the set of neural network pairs for reporting the channel state feedback.

[0022] Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for performing the following steps: sending an indication of a subset of the neural network pair set for training by the UE. In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, sending the indication of the accuracy level for reporting channel state feedback may include operations, features, components, or instructions for performing the following steps: sending indications of different accuracy levels for reporting channel state feedback for different subbands, spatial layers, channel taps, or in response to failure to decode different numbers of downlink transmissions including the same data. In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, the indicated accuracy level may include operations, features, components, or instructions for performing the following steps: sending an indication of a second accuracy level for reporting channel state feedback to be used to schedule a second downlink transmission, the first accuracy level being different from the second accuracy level.

[0023] Some examples of the methods, apparatuses, and non-transitory computer-readable media described herein may also include operations, features, components, or instructions for: sending an indication of a number of bits for the UE to use to report the channel state feedback based on the accuracy level, wherein the number of bits may be directly related to the accuracy level; and receiving the channel state feedback corresponding to the indicated accuracy level using the identified number of bits. In some examples of the methods, apparatuses, and non-transitory computer-readable media described herein, sending the indication of the accuracy level may include operations, features, components, or instructions for: sending the indication of the accuracy level in RRC signaling or MAC-CE. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 An example of a wireless communication system supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown.

[0025] Figure 2 An example of CSI feedback encoded by an encoder using a first neural network and decoded by a decoder using a second neural network according to aspects of the present disclosure is shown.

[0026] Figure 3 An example of a wireless communication system supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown.

[0027] Figure 4 An example of a process flow supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown.

[0028] Figure 5 and Figure 6 A block diagram of a device supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown.

[0029] Figure 7 A block diagram of a communication manager supporting configurable metrics for channel state compression and feedback is shown, according to aspects of the present disclosure.

[0030] Figure 8 A schematic diagram of a system including a device supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown.

[0031] Figure 9 and Figure 10 A block diagram of a device supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown.

[0032] Figure 11 A block diagram of a communication manager supporting configurable metrics for channel state compression and feedback is shown, according to aspects of the present disclosure.

[0033] Figure 12 A diagram of a system including devices supporting configurable metrics for channel state compression and feedback is shown according to aspects of the present disclosure.

[0034] Figure 13 and Figure 14 A flow chart of a method for supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. DETAILED DESCRIPTION

[0035] In some wireless communication systems, a user equipment (UE) may be configured to perform channel measurements on downlink signals received from a base station and report the channel measurements to the base station. The UE may report the channel measurements as channel state information (CSI) feedback. Using the CSI feedback, the base station may identify appropriate parameters for downlink transmissions to the UE to increase the likelihood that the UE will receive the downlink transmissions. The UE may use a first neural network to encode the CSI, and the base station may use a second neural network to decode the CSI. The first neural network and the second neural network may form a neural network pair, and the UE may train the neural network pair based on an accuracy level (e.g., using a loss metric or function). However, in some cases, the UE 115 may be configured to train the neural network pair based on the same accuracy level for all CSI feedback, and sending all CSI feedback at the same accuracy level may be inefficient. For example, if the accuracy level of the reported CSI feedback is unnecessarily high, the overhead of reporting the CSI may also be unnecessarily high. Alternatively, if the accuracy level or the reported CSI feedback is too low, downlink transmissions scheduled using CSI feedback may be unreliable.

[0036] As described herein, a wireless communication system may support efficient techniques that may allow a UE to report CSI to a base station with an appropriate level of accuracy. Specifically, the UE may be configured to train a neural network pair (e.g., comprising a first neural network at an encoder and a second neural network at a decoder) based on an accuracy level indicated by the base station. For example, the base station may indicate a loss metric or function corresponding to the accuracy level that the UE will use to report the CSI, and the UE may use the loss metric or function to train the neural network pair. In this way, the UE may report the CSI with the accuracy level indicated by the base station. The indicated accuracy level may depend on how the base station intends to use the CSI. For example, the base station may indicate different accuracy levels for CSI associated with different subbands, channel taps, spatial streams, feedback instances, etc. (e.g., so that delay-sensitive and reliability-sensitive transmissions are scheduled based on high-accuracy CSI feedback). Thus, when a higher accuracy level is appropriate, the UE may report the CSI with that higher accuracy level. Otherwise, the UE may report the CSI with a lower accuracy level.

[0037] The following describes aspects of the present disclosure introduced above in the context of a wireless communication system. Examples of processes and signaling exchanges supporting configurable metrics for channel state compression and feedback are then described. Aspects of the present disclosure are further illustrated and described with reference to apparatus diagrams, system diagrams, and flow charts related to configurable metrics for channel state compression and feedback.

[0038] Figure 1An example of a wireless communication system 100 that supports configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. 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 Mobile Broadband (MBB) communication, Enhanced MBB (eMBB) communication, ultra-reliable (e.g., mission-critical) communication, low-latency communication, communication with low-cost and low-complexity devices, or any combination thereof.

[0039] Base stations 105 may be dispersed throughout a geographic area to form wireless communication system 100 and may be devices of varying form factors or 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 a UE 115 and base station 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 communication of signals according to one or more radio access technologies.

[0040] 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 both, at different times. The UEs 115 may be devices of different forms or capabilities. Figure 1 1. Some example UEs 115 are shown in FIG. 1. The UEs 115 described herein are capable of communicating with various types of devices, such as other UEs 115, base stations 105, or network devices (e.g., core network nodes, relays, integrated access and backhaul (IAB) nodes, or other network devices), such as Figure 1 As shown.

[0041] The base stations 105 can communicate with the core network 130, or 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 via the backhaul links 120 (e.g., via X2, Xn, or other interfaces) directly (e.g., directly between the base stations 105) or indirectly (e.g., via the core network 130), or in both ways. In some examples, the backhaul links 120 can be or include one or more wireless links.

[0042] The one or more base stations 105 described herein may include or may be referred to by one of ordinary skill in the art as a base transceiver, a radio base station, an access point, a radio transceiver, a NodeB, an eNodeB (eNB), a next generation NodeB or giga-NodeB (any of which may be referred to as a gNB), a Home NodeB, a Home eNodeB, or other suitable terminology.

[0043] UE 115 may include or may 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 may 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.

[0044] The UE 115 described herein is 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 devices, including macro eNBs or gNBs, small cell eNBs or gNBs, or relay base stations, such as Figure 1 As shown.

[0045] The UE 115 and the base station 105 can wirelessly communicate with each other via one or more communication links 125 over one or more carriers. The term "carrier" may refer to a collection of radio frequency spectrum resources having a defined physical layer structure for supporting the communication link 125. For example, a carrier used for the communication link 125 may include a portion of a radio spectrum band (e.g., a bandwidth part (BWP)) operating according to one or more physical layer channels of 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 use carrier aggregation or multi-carrier operation to support communications with the UE 115. Depending on the carrier aggregation configuration, the UE 115 may be configured with multiple downlink component carriers and one or more uplink component carriers. Carrier aggregation may be used with both frequency division duplexing (FDD) component carriers and time division duplexing (TDD) component carriers.

[0046] The signal waveform transmitted on the carrier may be composed of multiple subcarriers (e.g., using multi-carrier modulation (MCM) techniques, 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 be composed of 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 coding rate of the modulation scheme, or both). Therefore, 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. 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.

[0047] 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, T s =1 / (Δf max ·N f ) seconds sampling period, where Δf max It can represent the maximum supported subcarrier spacing, and N fThe maximum supported Discrete Fourier Transform (DFT) size may be indicated. Time intervals for communication resources may be organized according to radio frames, each of which has 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).

[0048] Each frame may include multiple consecutively numbered subframes or time slots, and each subframe or time slot may have the same duration. In some examples, the frame may be divided (e.g., in the time domain) into subframes, and each subframe may be further divided into multiple 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 multiple symbol periods (e.g., depending on the length of a cyclic prefix pre-appended to each symbol period). In some wireless communication systems 100, the time slot may be further divided into multiple micro-time slots containing one or more symbols. In addition to the cyclic prefix, each symbol 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 frequency band.

[0049] 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 a burst of a shortened TTI (sTTI)).

[0050] Physical channels may be multiplexed on a carrier according to various techniques. For example, physical control channels and physical data channels may be multiplexed on a downlink carrier using one or more of a time division multiplexing (TDM) technique, a frequency division multiplexing (FDM) technique, or a hybrid TDM-FDM technique. A control region (e.g., a control resource set (CORESET)) of a physical control channel may be defined by multiple symbol periods and may extend across the system bandwidth of a carrier or a subset of the system bandwidth. One or more control regions (e.g., CORESETs) may be configured for a set of UEs 115. For example, one or more UEs 115 may monitor or search a 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. The aggregation level of a control channel candidate may refer to the number of control channel resources (e.g., control channel elements (CCEs)) associated with coded information of 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 .

[0051] In some examples, base stations 105 are mobile and therefore provide communication coverage for mobile geographic coverage areas 110. In some examples, different geographic coverage areas 110 associated with different technologies may overlap, but the different geographic coverage areas 110 may be supported by the same base station 105. In other examples, overlapping geographic coverage areas 110 associated with different technologies may be supported by different base stations 105. The wireless communication system 100 may 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.

[0052] 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 (Ultra-Reliable Low-Latency Communication, URLLC) or mission-critical communication. 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 service priority, 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 in this document.

[0053] In some examples, UE 115 can also communicate directly with other UEs 115 via 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 unable to receive transmissions from base station 105 for other reasons. In some examples, a group 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 resource scheduling for D2D communication. In other cases, D2D communication is performed between UEs 115 without involving base station 105.

[0054] 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), which may include at least one control plane entity (e.g., Mobility Management Entity (MME), Access and Mobility Management Function (AMF)) that manages access and mobility and at least one user plane entity (e.g., Serving Gateway (S-GW), Packet Data Network (PDN) Gateway (P-GW), or User Plane Function (UPF)) that routes packets or interconnections to external networks. The control plane entity may manage non-access stratum (NAS) functions of the UE 115 served by the base station 105 associated with the core network 130, such as mobility, authentication, and bearer management. User IP packets may be transferred through user plane entities, which may provide IP address allocation and other functions. The user plane entities may be connected to network operator IP services 150. Operator IP services 150 may include access to the Internet, intranet(s), IP Multimedia Subsystem (IMS), or packet-switched streaming services.

[0055] 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 the UE 115 through one or more other access network transport entities 145, which may be referred to as radio heads, smart radio heads, or transmission / reception 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).

[0056] 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). Typically, the region from 300 MHz to 3 GHz is referred to as the ultra-high frequency (UHF) region or decimeter band (because the wavelengths range from approximately 1 decimeter to 1 meter long). UHF waves may be blocked or redirected by buildings and environmental features, but these waves can penetrate structures sufficient for a macro cell to provide service to a UE 115 located indoors. Transmissions using UHF waves may be associated with smaller antennas and a shorter range (e.g., less than 100 kilometers) than transmissions using the smaller frequencies and longer waves of the high frequency (HF) or very high frequency (VHF) portions of the spectrum below 300 MHz.

[0057] The wireless communication system 100 can utilize both licensed and unlicensed radio spectrum bands. For example, the wireless communication system 100 can employ License 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 spectrum band, devices such as the base station 105 and the UE 115 can employ carrier sensing for conflict detection and avoidance. In some examples, operations in the unlicensed band can be based on a carrier aggregation configuration (e.g., LAA) in combination with component carriers operating in the licensed band. Operations in the unlicensed spectrum can include downlink transmissions, uplink transmissions, P2P transmissions, or D2D transmissions, among others.

[0058] 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 multiple rows and columns of antenna ports that the base station 105 may 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 panels may support radio frequency beamforming for signals transmitted via the antenna ports.

[0059] 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 implemented by combining signals transmitted 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 transmitted via antenna elements can include the transmitting device or the receiving device applying an amplitude shift, a phase shift, or both to the signals carried via the antenna elements associated with the 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).

[0060] In wireless communication system 100, UE 115 can be configured to perform channel measurements on downlink signals received from base station 105 and report the channel measurements to base station 105. UE 115 can report the channel measurements as CSI feedback. Using the CSI feedback, base station 105 can identify appropriate parameters for downlink transmissions to UE 115 to increase the likelihood that UE 115 will receive the downlink transmissions. In some cases, UE 115 can encode the CSI using a first neural network, and base station 105 can decode the CSI using a second neural network, where the first neural network and the second neural network form a neural network pair.

[0061] Figure 2 An example of CSI feedback 200 is shown that is encoded by an encoder 205 using a first neural network and decoded by a decoder 210 using a second neural network according to aspects of the present disclosure. Figure 2 In the example of , UE 115 can input channel realization information into encoder 205, and encoder 205 can use a first neural network to encode the channel realization information to generate CSI feedback. The channel realization information can refer to the original channel and can correspond to measurements performed on a CSI Reference Signal (CSI-RS) received on the channel. Thus, the encoder can take the original channel as input, and UE 115 can use the encoder neural network to create and feedback CSI. UE 115 can send CSI feedback to base station 105, and base station 105 can input the CSI feedback into decoder 210. Decoder 210 can use a second neural network to decode the CSI feedback to obtain channel state (e.g., base station 105 can use a decoder neural network to recover the original channel state from the CSI feedback). Base station 105 can then use the channel state to identify appropriate parameters for downlink transmission to UE 115.

[0062] In some wireless communication systems, the UE 115 may train a pair of neural networks, comprising a first neural network at an encoder and a second neural network at a decoder, based on a fixed level of accuracy. Training the neural networks may involve unsupervised learning, supervised learning, or a combination of both. For example, the UE 115 may train the encoder 205 based on one or more machine learning algorithms in the neural network. The neural networks at the encoder 205, the decoder 210, or both may include any number of machine learning layers (e.g., convolutional layers, fully connected layers, or some combination thereof). The UE 115 may implement any machine learning technique to train the neural networks at the encoder 205, the decoder 210, or both. For example, the UE 115 may implement deep learning (e.g., using a deep recurrent network), backpropagation, linear regression, a K-means model, a random forest model, or any combination of these or other machine learning techniques to train one or both neural networks.

[0063] In some machine learning examples, a network can train a machine learning model based on a training data set. The training data can be a subset of a larger data set. In some cases, training may involve determining one or more target features in the data set. The model can then learn one or more features from the training data (e.g., based on linear regression techniques, such as a linear regression algorithm) and evaluation metrics (such as mean square error (MSE), precision, accuracy, and recall). In some cases, the evaluation metric can be calculated based on a loss function.

[0064] The neural network training can be iterative, such that the UE 115 trains the neural network based on the current version of the neural network and measurements obtained since the current version of the neural network was implemented (e.g., rather than training from scratch using a full set of historical measurements). This iterative training process can reduce the processing overhead associated with training the neural network and can reduce the amount of historical measurement information stored by the UE 115 for neural network training. During the training process, the UE can apply these layers to measurement inputs or channel realizations to compress data from one or more base stations, sensors, radio access technologies (RATs), etc. The UE can feed back the compressed data through the decoder 210 to determine a plurality of decoding coefficients or parameters of the decoder neural network.

[0065] During training, the UE 115 may update encoder weights, encoder layers, decoder weights, decoder layers, or some combination thereof based on the feedback information. For example, the UE 115 may update the encoder weights based on a performance metric for encoding. Such a performance metric may be a metric that measures the level of compression achieved by the encoder neural network (e.g., comparing the number of bits associated with the encoded CSI feedback to the number of bits associated with the input measurement or channel realization), a metric that measures the reliability of extracting the input measurement from the encoder output using the decoder neural network, a metric that measures the computational complexity involved in compression, a metric that measures system performance based on the encoder neural network, or some combination thereof. Similarly, the UE may update the decoder weights based on a performance metric for decoding. Such a performance metric may be a metric that measures the similarity between the output measurement and the input measurement, a metric that measures the computational complexity involved in decompression, a metric that measures system performance based on the encoder output, or some combination thereof.

[0066] In some cases, the UE 115 may be configured with a fixed loss metric or loss function for training the encoder and decoder neural network pair. However, the desired loss metric or loss function for training the encoder and decoder neural network pair to report CSI feedback to the base station 105 may depend on how the base station 105 intends to use the CSI feedback. For example, for Single User MIMO (SU-MIMO), the base station 105 may be primarily concerned with learning the precoding direction, while for Multi-User MIMO (MU-MIMO), the base station 105 may be concerned with learning the raw channel state. Furthermore, the desired feedback accuracy may be different on different subbands or on different feedback instances (e.g., for eMBB communication versus URLLC, or for MU-MIMO on some subbands versus other subbands). Therefore, it may be inefficient for the UE 115 to send all CSI feedback with the same accuracy level.

[0067] The wireless communication system 100 may support efficient techniques that allow the UE 115 to report CSI to the base station with an appropriate (e.g., dynamically changing or configurable) level of accuracy. For example, the UE 115 may receive an indication of a loss function or loss metric corresponding to the level of accuracy, and when training an encoder and decoder neural network pair, the UE 115 may use the loss function or loss metric to calculate the loss. The loss function may be a function used to calculate the loss (e.g., a metric of the difference between the input to the neural network pair and the output from the neural network pair), and the loss metric may correspond to the metric used in the loss function. Based on the loss calculated in one training iteration, the UE 115 may adjust the coefficients in the neural network pair to minimize the loss calculated in future training iterations. For example, the neural network pair may implement an activation function for each layer of the network (e.g., for a hidden layer between an input layer and an output layer). The neural network pair may also implement a loss function or cost function based on the difference between actual and predicted values. For each layer of the neural network pair, the cost function may be used to adjust the weight of the next input based on the loss metric. In some examples, the cost function or loss function may implement an MSE function, which may calculate the square of the difference between the actual value and the predicted value. Thus, the loss function and loss metric may be different from the loss. Additionally, the loss function may be used to minimize the difference between the input and output of a neural network pair or the difference between one aspect of the input and output of a neural network pair.

[0068] Since the loss function may correspond to an accuracy level, any neural network pair trained using the loss function may encode and decode CSI feedback at the corresponding accuracy level. Thus, reporting CSI feedback corresponding to an accuracy level may refer to reporting CSI feedback encoded and decoded using a neural network pair trained using a loss function corresponding to the accuracy level. In some cases, the UE 115 may also be configured to train multiple neural network pairs using multiple loss functions, each corresponding to an accuracy level. In this case, the UE 115 may receive an indication of which of the multiple neural network pairs the UE 115 will use to report CSI. The indication of the neural network pair may correspond to the indication of the accuracy level because the neural network pair may be trained using a loss function corresponding to the accuracy level. Thus, reporting CSI feedback corresponding to an accuracy level may refer to reporting CSI feedback encoded and decoded using a neural network pair from the multiple neural network pairs, where the neural network pair was trained using a loss function corresponding to the accuracy level.

[0069] Furthermore, because the UE 115 can train both the neural network at the encoder and the neural network at the decoder (e.g., the encoder and decoder can be located at the UE 115, or the decoder can be located at the base station 105), the UE 115 can send the coefficients of the decoder neural network to the base station 105. Consequently, signaling between the UE 115 and the base station 105 can be minimized, as the UE 115 may not have to receive the output of the decoder from the base station 105 (e.g., for each iteration of training). Once the UE 115 completes training of the neural network pair, the UE 115 can send the coefficients of the decoder neural network to the base station 105. The UE 115 can then encode the CSI feedback using the encoder and report the CSI feedback to the base station 105, which can decode the CSI feedback using the decoder based on the coefficients received from the UE 115. Using the techniques described herein, the UE 115 can report CSI with an appropriate level of accuracy to minimize unnecessary overhead while allowing the base station 105 to identify appropriate parameters for communicating with the UE 115.

[0070] Figure 3 An example of a wireless communication system 300 supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The wireless communication system 300 includes a UE 115-a, which may be a reference Figure 1 and Figure 2 The wireless communication system 300 also includes a base station 105-a, which may be a reference Figure 1 and Figure 2An example of a base station 105 is depicted. Base station 105-a can provide communication coverage for coverage area 110-a. Wireless communication system 300 can implement aspects of wireless communication system 100. For example, wireless communication system 300 can support efficient techniques that allow UE 115-a to report CSI to base station 105-a with an appropriate level of accuracy.

[0071] exist Figure 3 In an example, the base station 105-a may send an indication of an accuracy level 305 that the UE 115-a will use to report the CSI feedback 310 to the base station 105-a, and the UE 115-a may report the CSI feedback 310 to the base station 105-a based on the indicated accuracy level 305. For example, the base station 105-a may send an indication of a loss function to the UE 115-a, and the UE 115-a may use the loss function to train an encoder and decoder neural network pair for encoding and decoding the CSI feedback 310. That is, the base station 105-a may configure the UE 115-a to use a particular loss function for encoder and decoder neural network training to report the CSI feedback 310. The UE 115-a may use real data (e.g., CSI based on actual channel measurements) or other data (e.g., other data provided by the base station 105-a) to train the encoder and decoder neural network pair. To train the neural network pair, UE 115-a may input CSI into the neural network pair, and UE 115-a may compare the CSI output from the neural network pair with the CSI input into the neural network pair. UE 115-a may then adjust coefficients (e.g., weights) in the neural network pair to minimize a loss (e.g., minimize the difference between the CSI input into the neural network pair and the CSI output from the neural network pair).

[0072] Since the accuracy level 305 or loss function may be configured by the base station 105-a, the accuracy of the CSI feedback 310 reported by the UE 115-a may be as desired by the base station 105-a as possible. Figure 1For example, the base station 105-a may configure the UE 115-a to report more accurate (e.g., high accuracy) CSI feedback 310 for delay-sensitive or reliability-sensitive communications, while the base station 105-a may configure the UE 115-a to report less accurate CSI feedback for other communications. In some cases, different accuracy levels or loss functions may be configured across different subbands, channel taps, spatial streams, or feedback instances. Specifically, the base station 105-a may configure different accuracy levels (e.g., different (relative) accuracy targets or different (relative) weightings of the loss function) across different subbands, channel taps, spatial streams, and feedback instances. Different subbands may be used for different types of communications (e.g., eMBB communications and URLLC), if UE 115-a operates on a time domain channel, different channel taps may be compressed to different levels of accuracy, different spatial streams may correspond to transmissions on different beams (e.g., higher accuracy or greater weighting in the direction of the strongest beam), and different feedback instances may correspond to different rounds of feedback (e.g., a first round of feedback for URLLC and a second round of feedback for URLLC).

[0073] The base station 105-a may send an indication of the accuracy level 305 via a higher layer message (e.g., RRC signaling) or dynamic signaling (e.g., in a MAC control element (MAC-CE)). In addition, in some cases, if quantization is performed on the CSI feedback 310 before transmission to the base station 105-a, a larger number of bits may be used to compress some paths, subbands, channel taps, spatial streams, feedback instances, etc. For example, the UE 115-a may use a different number of bits to send the CSI feedback 310 based on the accuracy level 305 indicated by the base station 105-a. In this case, the base station 105-a may send an indication of the number of bits for the UE 115-a to use to send the CSI feedback 310, and the UE 115-a may send the CSI feedback 310 using the indicated number of bits.

[0074] The base station 105-a may also request the UE 115-a to train multiple encoder and decoder neural network pairs (e.g., N neural network pairs), and the base station 105-a may configure an accuracy level for each neural network pair (e.g., for a trade-off between feedback accuracy and CSI feedback overhead). The base station 105-a may then send an indication of one of the multiple neural network pairs for the UE 115-a to use for reporting CSI feedback 310. Alternatively, the UE 115-a may autonomously select (e.g., without signaling from the base station 105-a) one of the multiple neural network pairs to use for reporting CSI feedback 310. In some cases, the base station 105-a may send an indication of an equation for training the neural network pair rather than an indication of a loss metric or loss function for training the neural network pair. In other cases, the base station 105-a may send an indication of a set of loss metrics or loss functions for training the neural network pair. In this case, base station 105-a may send an indication to UE 115-a of a loss metric or loss function from the set for the UE to use for training the neural network pair. Alternatively, UE 115-a may autonomously select a loss metric or loss function from the set for use in training the neural network pair (e.g., based on the indicated level of accuracy).

[0075] Figure 4 An example of a process flow 400 for supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The process flow 400 illustrates aspects of a technique performed by a UE 115-b, which may be a reference Figure 1-Figure 3 The process flow 400 also illustrates aspects of the technique performed by a base station 105-b, which may be a reference Figure 1-Figure 3 The process flow 400 may implement aspects of the wireless communication system 300. For example, the process flow 400 may support efficient techniques that may allow the UE 115-b to report CSI to the base station 105-b with an appropriate level of accuracy. The UE 115-b may also support the ability to perform channel state compression and feedback differently for different subbands, channel taps, spatial streams, feedback instances, etc.

[0076] The accuracy level may refer to the difference between the original CSI (or measured actual channel conditions) input to the encoder at UE 115-b and the CSI generated by the decoder at base station 105-b. For example, a high accuracy level may indicate little or no difference between the original CSI (or measured actual channel conditions) input to the encoder and the CSI generated by the decoder, while a low accuracy level may indicate a large difference between the original CSI (or measured actual channel conditions) input to the encoder and the CSI generated by the decoder. In other words, the higher the accuracy, the lower the compression, and vice versa. In addition, in some cases, when base station 105-b signals the accuracy level, base station 105-b may signal one or more aspects of CSI that UE 115-b should focus on when measuring and reporting CSI. Then, when training the neural network pair, the loss function may prioritize or apply greater weight to one or more aspects of CSI indicated by base station 105-b. In this way, UE 115-b is able to generate and report CSI feedback consistent with how base station 105-b intends to use CSI feedback.In this case, the accuracy level may refer to the priority or weight applied to those aspects of CSI desired by base station 105.

[0077] At 405, UE 115-b may identify a neural network pair that includes a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder. For example, base station 105-b may indicate the number of neural network pairs to be trained by UE 115-b. In some cases, UE 115-b may select a neural network pair based on the indication. At 410, UE 115-b may receive an indication of an accuracy level for reporting CSI feedback to base station 105-b (e.g., in an RRC or MAC-CE message). That is, base station 105-b and UE 115-b may support different accuracy levels for CSI feedback. As an example, UE 115-b may receive an indication of a loss metric or loss function from base station 105-b, and at 415, UE 115-b may use the loss metric or loss function to train the neural network pair (e.g., where the loss metric or loss function implicitly indicates the accuracy level). At 420 , UE 115 - b may send coefficients of a second neural network at a decoder to base station 105 - b based on the trained neural network pair.

[0078] In some cases, the indicated accuracy level (or loss metric) may be based on one or more of the subband, spatial layer, or channel tap to which the CSI feedback corresponds. That is, the base station 105-b may indicate to the UE 115-b a loss metric or loss function and weights for training a neural network pair for different subbands, channel taps, spatial streams, or feedback instances. In this case, the UE 115-b may receive data from the base station 105-b on the subband or spatial layer or according to the channel tap based on reporting CSI feedback corresponding to the indicated accuracy level. Additionally or alternatively, the indicated accuracy level may be based on the number of downlink transmissions (e.g., feedback instances or feedback rounds) that included the same data that the UE 115-b failed to decode. In this case, the UE 115-b may receive retransmissions of the same data that the UE failed to decode based on reporting CSI feedback corresponding to the indicated accuracy level. For example, if the UE 115-b fails to decode multiple retransmissions of the same data, the UE 115-b may report CSI feedback with high accuracy. Therefore, the base station 105-b can retransmit the same data based on the highly accurate CSI feedback, and the chance that the UE 115-b can successfully decode the retransmission of the same data may be high.

[0079] In some cases, UE 115-b may receive an indication to train multiple neural network pairs based on multiple loss metrics or loss functions (e.g., multiple accuracy levels), and UE 115-b may train each of the multiple neural network pairs based on a corresponding accuracy level in the multiple accuracy levels. In this case, base station 105-b may send, and UE 115-b may receive, an indication of which neural network pair UE 115-b is to use to report CSI feedback to base station 105-b. Alternatively, UE 115-b may autonomously select one of the multiple neural network pairs to report CSI feedback to base station 105-b. In addition, UE 115-b may receive an indication of a subset of the multiple neural network pairs that UE 115-b is to train. That is, base station 105-b may send the number of neural network pairs to be trained at UE 115-b and which neural network pairs UE 115-b is to use for reporting CSI feedback for a particular subband, channel tap, spatial stream, or feedback instance.

[0080] At 425, the base station 105-b may send downlink data or a reference signal (e.g., a CSI-RS) to the UE 115-b, and the UE 115-b may perform channel measurements based on the downlink data or reference signal received from the base station 105-b to generate CSI feedback. At 430, the UE 115-b may encode the channel state feedback using the first neural network of the neural network pair, and at 435, the UE 115-b may report the CSI feedback (e.g., CSI feedback corresponding to the indicated level of accuracy) to the base station 105-b. In some cases, the UE 115-b may identify a number of bits for reporting the CSI feedback based on the indicated level of accuracy, and the UE 115-b may report the CSI feedback corresponding to the indicated level of accuracy using the identified number of bits. The UE 115-b may receive an indication of the number of bits for reporting the CSI feedback based on the level of accuracy. That is, the base station 105-b may indicate to the UE 115-b the number of quantized bits of CSI feedback to use for different subbands, channel taps, spatial streams, or feedback instances. At 440, the base station 105-b may use a second neural network to decode the channel state feedback received from the UE 115-b (e.g., based on the decoder coefficients received at 420).

[0081] Figure 5 A block diagram 500 of a device 505 supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The device 505 can be an example of aspects of the UE 115 as described herein. The device 505 can include a receiver 510, a communication manager 515, and a transmitter 520. The device 505 can also include a processor. Each of these components can communicate with each other (e.g., via one or more buses).

[0082] The receiver 510 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 configurable metrics for channel state compression and feedback, etc.). The information may be passed to other components of the device 505. The receiver 510 may be a reference Figure 8 Examples of aspects of the transceiver 820 are described. The receiver 510 may utilize a single antenna or a collection of antennas.

[0083] The communication manager 515 may receive, from a base station, an indication of an accuracy level for reporting channel state feedback to the base station; receive downlink data or a reference signal from the base station; and report, based at least in part on the downlink data or the reference signal, channel state feedback corresponding to the indicated accuracy level to the base station. The communication manager 515 may be an example of aspects of the communication manager 810 described herein.

[0084] The communication manager 515 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 515 or its subcomponents may be performed by a general-purpose processor, a DSP, an application-specific integrated circuit (ASIC), 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 in this disclosure.

[0085] The communication manager 515 or its subcomponents can be physically located in a variety of locations, including being distributed so that portions of functionality are performed by one or more physical components at different physical locations. In some examples, according to aspects of the present disclosure, the communication manager 515 or its subcomponents can be separate and distinct components. In some examples, according to aspects of the present disclosure, the communication manager 515 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.

[0086] The transmitter 520 can transmit signals generated by other components of the device 505. In some examples, the transmitter 520 can be co-located with the receiver 510 in a transceiver module. For example, the transmitter 520 can be a reference Figure 8 Examples of aspects of the depicted transceiver 820. The transmitter 520 may utilize a single antenna or a collection of antennas.

[0087] Figure 6 A block diagram 600 of a device 605 supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The device 605 can be an example of aspects of the device 505 or UE 115 as described herein. The device 605 can include a receiver 610, a communication manager 615, and a transmitter 640. The device 605 can also include a processor. Each of these components can communicate with each other (e.g., via one or more buses).

[0088] The receiver 610 may receive information associated with various information channels (e.g., control channels, data channels, and information related to configurable metrics for channel state compression and feedback, etc.), such as packets, user data, or control information. The information may be passed to other components of the device 605. The receiver 610 may be a reference Figure 8 Examples of aspects of the transceiver 820 are described. The receiver 610 may utilize a single antenna or a collection of antennas.

[0089] The communication manager 615 may be an example of the communication manager 515 described herein. The communication manager 615 may include a CSI accuracy manager 620, a downlink manager 625, and a CSI reporter 630. The communication manager 615 may be an example of aspects of the communication manager 810 described herein.

[0090] The CSI accuracy manager 620 may receive an indication of an accuracy level for reporting channel state feedback to the base station from a base station. The downlink manager 625 may receive downlink data or a reference signal from the base station. The CSI reporter 630 may report the channel state feedback corresponding to the indicated accuracy level to the base station based on the downlink data or the reference signal.

[0091] The transmitter 640 can transmit signals generated by other components of the device 605. In some examples, the transmitter 640 can be co-located with the receiver 610 in a transceiver module. For example, the transmitter 640 can be a reference Figure 8 Examples of aspects of the transceiver 820 are described. The transmitter 640 may utilize a single antenna or a collection of antennas.

[0092] Figure 7 A block diagram 700 of a communication manager 705 supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The communication manager 705 can be an example of aspects of the communication manager 515, the communication manager 615, or the communication manager 810 described herein. The communication manager 705 can include a CSI accuracy manager 710, a downlink manager 715, a CSI manager 720, a CSI reporter 725, a neural network manager 730, an encoder 735, and a loss function manager 740. Each of these modules can communicate with each other directly or indirectly (e.g., via one or more buses).

[0093] The CSI accuracy manager 710 may receive an indication of an accuracy level for reporting channel state feedback to the base station from a base station. The downlink manager 715 may receive downlink data or a reference signal from the base station. The CSI reporter 725 may report the channel state feedback corresponding to the indicated accuracy level to the base station based on the downlink data or the reference signal.

[0094] The loss function manager 740 may receive an indication of a loss function corresponding to the accuracy level for use in training a neural network pair, the neural network pair comprising a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder. The neural network manager 730 may use the loss function to train the neural network pair. The neural network manager 730 may iteratively input a channel state feedback input into the neural network pair and identify a channel state feedback output from the neural network pair. The neural network manager 730 may then use the loss function to determine, for each iteration, a difference between the channel state feedback input and the channel state feedback output, wherein the difference comprises a loss, and the neural network manager 730 may adjust coefficients of the neural network pair for each iteration based on the determination to minimize the difference between the channel state feedback input and the channel state feedback output.

[0095] The encoder 735 may encode the channel state feedback at the encoder using the first neural network based on the training. In some examples, the CSI reporter 725 may report the encoded channel state feedback. In some examples, the neural network manager 730 may send coefficients of the second neural network to the base station, the second neural network being used to decode the channel state feedback based on the training. In some examples, the neural network manager 730 may receive an indication from the base station to train a set of neural network pairs based on a set of accuracy levels, the set of neural network pairs including the neural network pair. In some examples, the neural network manager 730 may train each neural network pair in the set of neural network pairs based on a corresponding accuracy level of the set of accuracy levels. In some examples, the neural network manager 730 may receive an indication to use the neural network pair in the set of neural network pairs for reporting the channel state feedback. In some examples, the neural network manager 730 may autonomously select the neural network pair in the set of neural network pairs for reporting the channel state feedback. In some examples, the neural network manager 730 may receive an indication of a subset of the set of neural network pairs for the UE to train.

[0096] In some examples, the indicated accuracy level is based on one or more of a subband, spatial layer, or channel tap to which the channel state feedback corresponds. In some examples, the downlink manager 715 may receive data from the base station on the subband or spatial layer or according to the channel tap based on reporting the channel state feedback corresponding to the indicated accuracy level. In some examples, the indicated accuracy level is based on the number of downlink transmissions including the same data that the UE failed to decode. In some examples, the downlink manager 715 may receive retransmissions of the same data that the UE failed to decode based on reporting the channel state feedback corresponding to the indicated accuracy level.

[0097] In some examples, the CSI manager 720 may identify a number of bits to use for reporting the channel state feedback based on the accuracy level, where the number of bits is directly related to the accuracy level. In some examples, the CSI reporter 725 may utilize the identified number of bits to report the channel state feedback corresponding to the indicated accuracy level. In some examples, the CSI manager 720 may receive an indication of the number of bits to use for reporting the channel state feedback based on the accuracy level. In some examples, the CSI accuracy manager 710 may receive the indication of the accuracy level in RRC signaling or MAC-CE.

[0098] Figure 8 A schematic diagram of a system 800 is shown that includes a device 805 that supports configurable metrics for channel state compression and feedback according to aspects of the present disclosure. The device 805 may be an example of or include components of the device 505, device 605, or UE 115 as described herein. The device 805 may include components for two-way voice and data communications, including components for sending and receiving communications, including a communication manager 810, an I / O controller 815, a transceiver 820, an antenna 825, a memory 830, and a processor 840. These components may communicate electronically via one or more buses (e.g., bus 845).

[0099] The communication manager 810 may: receive, from a base station, an indication of an accuracy level for reporting channel state feedback to the base station; receive downlink data or a reference signal from the base station; and report, to the base station based on the downlink data or the reference signal, the channel state feedback corresponding to the indicated accuracy level.

[0100] I / O controller 815 can manage input and output signals for device 805. I / O controller 815 can also manage peripheral devices that are not integrated into device 805. In some cases, I / O controller 815 can represent a physical connection or port to an external peripheral device. In some cases, I / O controller 815 can utilize an operating system, such as MS- MS- Or other known operating systems. In other cases, I / O controller 815 can represent or interact with a modem, keyboard, mouse, touch screen, or similar device. In some cases, I / O controller 815 can be implemented as part of a processor. In some cases, a user can interact with device 805 via I / O controller 815 or via hardware components controlled by I / O controller 815.

[0101] The transceiver 820 can perform bidirectional communication as described above via one or more antennas, wired or wireless links. For example, the transceiver 820 can represent a wireless transceiver and can communicate bidirectionally with another wireless transceiver. The transceiver 820 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.

[0102] In some cases, a wireless device may include a single antenna 825. However, in some cases, a device may have more than one antenna 825 that is capable of sending or receiving multiple wireless transmissions simultaneously.

[0103] The memory 830 may include RAM and ROM. The memory 830 may store computer-readable, computer-executable code 835, including instructions that, when executed, cause the processor to perform the various functions described herein. In some cases, the memory 830 may include a BIOS, etc., which may control basic hardware or software operations, such as interaction with peripheral components or devices.

[0104] The processor 840 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 840 may be configured to operate the memory array using a memory controller. In other cases, the memory controller may be integrated into the processor 840. The processor 840 may be configured to execute computer-readable instructions stored in a memory (e.g., memory 830) to cause the device 805 to perform various functions (e.g., functions or tasks that support configurable metrics for channel state compression and feedback).

[0105] The code 835 may include instructions for implementing aspects of the present disclosure, including instructions for supporting wireless communications. The code 835 may be stored in a non-transitory computer-readable medium, such as system memory or other types of memory. In some cases, the code 835 may not be directly executable by the processor 840, but may cause a computer (e.g., when compiled and executed) to perform the functions described herein.

[0106] Figure 9 A block diagram 900 of a device 905 supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The device 905 can be an example of aspects of a base station 105 as described herein. The device 905 can include a receiver 910, a communication manager 915, and a transmitter 920. The device 905 can also include a processor. Each of these components can communicate with each other (e.g., via one or more buses).

[0107] The receiver 910 may receive information associated with various information channels (e.g., control channels, data channels, and information related to configurable metrics for channel state compression and feedback, etc.), such as packets, user data, or control information. The information may be passed to other components of the device 905. The receiver 910 may be a reference Figure 12 Examples of aspects of the transceiver 1220 are described. The receiver 910 may utilize a single antenna or a collection of antennas.

[0108] The communication manager 915 may: send an indication of an accuracy level for reporting channel state feedback to the base station to the UE; send downlink data or a reference signal to the UE; and receive channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE. The communication manager 915 may be an example of aspects of the communication manager 1210 described herein.

[0109] The communication manager 915 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 915 or its subcomponents may be performed by a general-purpose processor, a DSP, an application-specific integrated circuit (ASIC), 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 in this disclosure.

[0110] The communication manager 915 or its subcomponents can be physically located in a variety of locations, including being distributed so that portions of functionality are performed by one or more physical components at different physical locations. In some examples, according to aspects of the present disclosure, the communication manager 915 or its subcomponents can be separate and distinct components. In some examples, according to aspects of the present disclosure, the communication manager 915 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.

[0111] The transmitter 920 can transmit signals generated by other components of the device 905. In some examples, the transmitter 920 can be co-located with the receiver 910 in a transceiver module. For example, the transmitter 920 can be a reference Figure 12 Examples of aspects of the transceiver 1220 are described. The transmitter 920 may utilize a single antenna or a collection of antennas.

[0112] Figure 10 A block diagram 1000 of a device 1005 supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The device 1005 can be an example of aspects of the device 905 or base station 105 as described herein. The device 1005 can include a receiver 1010, a communication manager 1015, and a transmitter 1035. The device 1005 can also include a processor. Each of these components can communicate with each other (e.g., via one or more buses).

[0113] The receiver 1010 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 configurable metrics for channel state compression and feedback, etc.). The information may be passed to other components of the device 1005. The receiver 1010 may be a reference Figure 12 Examples of aspects of the transceiver 1220 are described. The receiver 1010 may utilize a single antenna or a collection of antennas.

[0114] The communication manager 1015 may be an example of aspects of the communication manager 915 described herein. The communication manager 1015 may include a CSI accuracy manager 1020, a downlink manager 1025, and a CSI manager 1030. The communication manager 1015 may be an example of aspects of the communication manager 1210 described herein.

[0115] The CSI accuracy manager 1020 may send an indication of an accuracy level for reporting channel state feedback to the base station to the UE. The downlink manager 1025 may send downlink data or a reference signal to the UE. The CSI manager 1030 may receive channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE.

[0116] The transmitter 1035 can transmit signals generated by other components of the device 1005. In some examples, the transmitter 1035 can be co-located with the receiver 1010 in a transceiver module. For example, the transmitter 1035 can be a reference Figure 12 Examples of aspects of the transceiver 1220 are described. The transmitter 1035 may utilize a single antenna or a collection of antennas.

[0117] Figure 11 A block diagram 1100 of a communication manager 1105 supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The communication manager 1105 can be an example of aspects of the communication manager 915, the communication manager 1015, or the communication manager 1210 described herein. The communication manager 1105 can include a CSI accuracy manager 1110, a downlink manager 1115, a CSI manager 1120, a neural network manager 1125, a decoder 1130, and a loss metric manager 1135. Each of these modules can communicate with each other directly or indirectly (e.g., via one or more buses).

[0118] The CSI accuracy manager 1110 may send an indication of an accuracy level for reporting channel state feedback to the base station to the UE. The downlink manager 1115 may send downlink data or a reference signal to the UE. The CSI manager 1120 may receive channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE. The loss metric manager 1135 may send an indication of a loss metric or loss function for the UE to use in training a neural network pair for reporting channel state feedback.

[0119] The neural network manager 1125 may receive from the UE coefficients of a neural network used at a decoder to decode the channel state feedback from the UE. The decoder 1130 may use the neural network at the decoder to decode the channel state feedback from the UE. In some examples, the neural network manager 1125 may send an indication that the UE is training a set of neural network pairs based on a set of accuracy levels, each neural network pair comprising a first neural network used at an encoder to encode the channel state feedback and a second neural network used at a decoder to decode the channel state feedback. In some examples, the neural network manager 1125 may send an indication that the UE is using a neural network pair from the set of neural network pairs to report the channel state feedback. In some examples, the neural network manager 1125 may send an indication of a subset of the set of neural network pairs for the UE to train.

[0120] In some examples, CSI accuracy manager 1110 may send indications of different accuracy levels for reporting channel state feedback for different subbands, spatial layers, channel taps, or in response to a failure to decode different numbers of downlink transmissions comprising the same data. In some examples, the indicated accuracy levels include a first accuracy level for reporting channel state feedback to be used for scheduling a first downlink data transmission, and CSI accuracy manager 1110 may send an indication of a second accuracy level for reporting channel state feedback to be used for scheduling a second downlink transmission, the first accuracy level being different from the second accuracy level. In some examples, CSI manager 1120 may send an indication of a number of bits for the UE to use for reporting the channel state feedback based on the accuracy level, where the number of bits is directly related to the accuracy level. In some examples, CSI manager 1120 may receive the channel state feedback corresponding to the indicated accuracy level using the identified number of bits. In some examples, CSI accuracy manager 1110 may send the indication of the accuracy level in RRC signaling or MAC-CE.

[0121] Figure 12A schematic diagram of a system 1200 according to aspects of the present disclosure is shown, including a device 1205 that supports configurable metrics for channel state compression and feedback. Device 1205 may be an example of, or include components of, device 905, device 1005, or base station 105 as described herein. Device 1205 may include components for two-way voice and data communications, including components for sending and receiving communications, including a communications manager 1210, a network communications manager 1215, a transceiver 1220, an antenna 1225, a memory 1230, a processor 1240, and an inter-station communications manager 1245. These components may communicate electronically via one or more buses (e.g., bus 1250).

[0122] The communication manager 1210 may: send an indication of an accuracy level for reporting channel state feedback to the base station to the UE; send downlink data or a reference signal to the UE; and receive channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE.

[0123] The network communications manager 1215 may manage communications with the core network (eg, via one or more wired backhaul links). For example, the network communications manager 1215 may manage the movement of data communications for client devices (eg, one or more UEs 115).

[0124] The transceiver 1220 can perform bidirectional communication as described above via one or more antennas, wired or wireless links. For example, the transceiver 1220 can represent a wireless transceiver and can communicate bidirectionally with another wireless transceiver. The transceiver 1220 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.

[0125] In some cases, a wireless device may include a single antenna 1225. However, in some cases, a device may have more than one antenna 1225, which may be capable of sending or receiving multiple wireless transmissions simultaneously.

[0126] Memory 1230 may include RAM, ROM, or a combination thereof. Memory 1230 may store computer-readable code 1235 including instructions that, when executed by a processor (e.g., processor 1240), cause the device to perform the various functions described herein. In some cases, memory 1230 may include BIOS, which may control basic hardware or software operations, such as interaction with peripheral components or devices.

[0127] Processor 1240 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, processor 1240 may be configured to operate a memory array using a memory controller. In some cases, the memory controller may be integrated into processor 1240. Processor 1240 may be configured to execute computer-readable instructions stored in a memory (e.g., memory 1230) to cause device 1205 to perform various functions (e.g., functions or tasks that support configurable metrics for channel state compression and feedback).

[0128] The inter-site communication manager 1245 can manage communications with other base stations 105 and can include a controller or scheduler for cooperating with other base stations 105 to control communications with the UE 115. For example, the inter-site communication manager 1245 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 1245 can provide an X2 interface in LTE / LTE-A wireless communication network technology to provide communications between base stations 105.

[0129] The code 1235 may include instructions for implementing aspects of the present disclosure, including instructions for supporting wireless communications. The code 1235 may be stored in a non-transitory computer-readable medium, such as system memory or other types of memory. In some cases, the code 1235 may not be directly executable by the processor 1240, but may cause a computer (e.g., when compiled and executed) to perform the functions described herein.

[0130] Figure 13 A flow chart illustrating a method 1300 for supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The operations of the method 1300 may be implemented by a UE 115 or components thereof as described herein. For example, the operations of the method 1300 may be implemented by a UE 115 or components thereof as described herein. Figures 5 to 8 In some examples, the UE may execute a set of instructions to control the functional elements of the UE to perform the functions described below. Additionally or alternatively, the UE may use dedicated hardware to perform aspects of the functions described below.

[0131] At 1305, the UE may receive an indication of an accuracy level for reporting channel state feedback to the base station from the base station. The operations of 1305 may be performed according to the methods described herein. In some examples, aspects of the operations of 1305 may be described with reference to Figure 6 and Figure 7 The described CSI accuracy manager is implemented.

[0132] At 1310, the UE may receive downlink data or a reference signal from the base station. The operations of 1310 may be performed according to the methods described herein. In some examples, aspects of the operations of 1310 may be performed as described in reference to Figure 6 and Figure 7 The downlink manager described here is used to perform the following operations:

[0133] At 1315, the UE may report the channel state feedback corresponding to the indicated accuracy level to the base station based on the downlink data or the reference signal. The operations of 1315 may be performed according to the methods described herein. In some examples, aspects of the operations of 1315 may be as described in reference to Figure 6 and Figure 7 The CSI reporter described is performed.

[0134] Figure 14 A flow chart illustrating a method 1400 for supporting configurable metrics for channel state compression and feedback according to aspects of the present disclosure is shown. The operations of the method 1400 may be implemented by a base station 105 or components thereof as described herein. For example, the operations of the method 1400 may be implemented by a base station 105 or components thereof as described herein. Figures 9 to 12 In some examples, the base station may execute a set of instructions to control the functional elements of the base station to perform the functions described below. Additionally or alternatively, the base station may use dedicated hardware to perform aspects of the functions described below.

[0135] At 1405, the base station may send an indication to the UE of the accuracy level of the channel state feedback reported to the base station. The operations of 1405 may be performed according to the methods described herein. In some examples, aspects of the operations of 1405 may be as described in reference to Figure 10 and Figure 11 The described CSI accuracy manager is implemented.

[0136] At 1410, the base station may send downlink data or a reference signal to the UE. The operations of 1410 may be performed according to the methods described herein. In some examples, aspects of the operations of 1410 may be performed by reference to Figure 10 and Figure 11 The downlink manager described here is used to perform the following operations:

[0137] At 1415, the base station may receive channel state feedback corresponding to the indicated accuracy level from the UE based on sending the downlink data or the reference signal to the UE. The operations of 1415 may be performed according to the methods described herein. In some examples, aspects of the operations of 1415 may be as described in reference to Figure 10 and 11The CSI manager described is executed.

[0138] The following provides an overview of examples of the present disclosure:

[0139] Example 1: A method for wireless communication at a UE, comprising: receiving an indication of an accuracy level for reporting channel state feedback to the base station from a base station; receiving downlink data or a reference signal from the base station; and reporting the channel state feedback corresponding to the indicated accuracy level to the base station based at least in part on the downlink data or the reference signal.

[0140] Example 2: The method of Example 1, wherein receiving the indication of the level of accuracy for reporting channel state feedback comprises: receiving an indication of a loss function corresponding to the level of accuracy for training a neural network pair, the neural network pair comprising a first neural network for encoding the channel state feedback at an encoder, and a second neural network for decoding the channel state feedback at a decoder, the method further comprising: training the neural network pair using the loss function.

[0141] Example 3: The method of any of Examples 1 or 2, wherein training the neural network pair using the loss function comprises: iteratively entering a channel state feedback input into the neural network pair, and identifying a channel state feedback output from the neural network pair; determining, using the loss function, a difference between the channel state feedback input and the channel state feedback output for each iteration, wherein the difference comprises a loss; and based at least in part on the determination, adjusting coefficients of the neural network pair for each iteration to minimize the difference between the channel state feedback input and the channel state feedback output.

[0142] Example 4: The method of any of Examples 1 to 3, wherein reporting the channel state feedback corresponding to the indicated level of accuracy includes: encoding the channel state feedback using the first neural network at the encoder based at least in part on the training; and reporting the encoded channel state feedback.

[0143] Example 5: The method of any one of Examples 1 to 4 further includes: sending coefficients of the second neural network to the base station for decoding the channel state feedback based at least in part on the training.

[0144] Example 6: The method of any of Examples 1 to 5, further comprising: receiving an indication from the base station to train a plurality of neural network pairs based at least in part on a plurality of accuracy levels, the plurality of neural network pairs including the neural network pair; and training each of the plurality of neural network pairs based at least in part on a corresponding level of accuracy in the plurality of accuracy levels.

[0145] Example 7: The method of any of Examples 1 to 6, wherein receiving the indication of the accuracy level comprises: receiving an indication to use the neural network pair of the plurality of neural network pairs for reporting the channel state feedback.

[0146] Example 8: The method of any one of Examples 1 to 7 further includes: autonomously selecting the neural network pair among the multiple neural network pairs for reporting the channel state feedback.

[0147] Example 9: The method of any one of Examples 1 to 8 further includes: receiving an indication of a subset of the plurality of neural network pairs for training by the UE.

[0148] Example 10: The method of any one of Examples 1 to 9, wherein the indicated level of accuracy is based at least in part on one or more of the subbands, spatial layers, or channel taps to which the channel state feedback corresponds, the method further comprising: receiving data from a base station on a subband or spatial layer or according to a channel tap based at least in part on reporting the channel state feedback corresponding to the indicated level of accuracy.

[0149] Example 11: The method of any of Examples 1 to 10, wherein the indicated level of accuracy is based at least in part on the number of downlink transmissions comprising the same data that the UE failed to decode, the method further comprising: receiving retransmissions of the same data that the UE failed to decode based at least in part on reporting the channel state feedback corresponding to the indicated level of accuracy.

[0150] Example 12: The method of any one of Examples 1 to 11 also includes: identifying a number of bits used to report the channel state feedback based at least in part on the accuracy level, wherein the number of bits is directly related to the accuracy level; and reporting the channel state feedback corresponding to the indicated accuracy level using the identified number of bits.

[0151] Example 13: The method of any one of Examples 1 to 12, further comprising: receiving an indication of a number of bits used to report the channel state feedback based at least in part on the accuracy level.

[0152] Example 14: The method of any one of Examples 1 to 13, wherein receiving the indication of the accuracy level comprises: receiving the indication of the accuracy level in RRC signaling or MAC-CE.

[0153] Example 15: A method for wireless communication at a base station, comprising: sending an indication to a user equipment (UE) of a level of accuracy for reporting channel state feedback to the base station; sending downlink data or a reference signal to the UE; and receiving channel state feedback from the UE corresponding to the indicated level of accuracy based at least in part on sending the downlink data or reference signal to the UE.

[0154] Example 16: The method of Example 15, wherein sending the indication of the accuracy level for reporting channel state feedback includes: sending an indication of a loss function for use by the UE to train a neural network pair for reporting the channel state feedback.

[0155] Example 17: The method of any of Examples 15 or 16 further includes: receiving from the UE coefficients of a neural network used at a decoder to decode the channel state feedback from the UE; and using the neural network at the decoder to decode the channel state feedback from the UE.

[0156] Example 18: The method of any of Examples 15 to 17, further comprising: sending an indication that the UE trains a plurality of neural network pairs based at least in part on a plurality of accuracy levels, each neural network pair comprising a first neural network at an encoder for encoding the channel state feedback and a second neural network at a decoder for decoding the channel state feedback.

[0157] Example 19: The method of any one of Examples 15 to 18 further includes: sending an indication that the UE uses a neural network pair from the multiple neural network pairs to report the channel state feedback.

[0158] Example 20: The method of any one of Examples 15 to 19, further comprising: sending an indication of a subset of the plurality of neural network pairs for training by the UE.

[0159] Example 21: The method of any of Examples 15 to 20, wherein sending the indication of the level of accuracy for reporting channel state feedback includes sending indications of different levels of accuracy for reporting channel state feedback for different subbands, spatial layers, channel taps, or in response to a failure to decode different numbers of downlinks comprising the same data.

[0160] Example 22: The method of any one of Examples 15 to 21, wherein the indicated accuracy level includes a first accuracy level for reporting channel state feedback to be used for scheduling a first downlink data transmission, and the method further includes: sending an indication of a second accuracy level for reporting channel state feedback to be used for scheduling a second downlink transmission, the first accuracy level being different from the second accuracy level.

[0161] Example 23: The method of any one of Examples 15 to 22 further includes: sending an indication of the number of bits for the UE to use for reporting the channel state feedback based at least in part on the accuracy level, wherein the number of bits is directly related to the accuracy level; and receiving the channel state feedback corresponding to the indicated accuracy level using the identified number of bits.

[0162] Example 24: The method of any one of Examples 15 to 23, wherein sending the indication of the accuracy level includes: sending the indication of the accuracy level in RRC signaling or MAC-CE.

[0163] Example 25: An apparatus for wireless communication comprising at least one component for performing the method of any one of Examples 1 to 14.

[0164] Example 26: An apparatus for wireless communication, comprising a processor and a memory coupled to the processor. The processor and the memory may be configured to cause the apparatus to perform the method of any one of Examples 1 to 14.

[0165] Example 27: A non-transitory computer-readable medium storing code for wireless communication, comprising a processor, a memory connected to the processor, and instructions stored in the memory and executable by the processor to cause the device to perform the method of any one of Examples 1 to 14.

[0166] Example 28: An apparatus for wireless communication, comprising at least one component for performing the method of any of Examples 15 to 24.

[0167] Example 29: An apparatus for wireless communication, comprising a processor and a memory coupled to the processor. The processor and the memory may be configured to cause the apparatus to perform the method of any one of Examples 15 to 24.

[0168] Example 30: A non-transitory computer-readable medium storing code for wireless communication, comprising a processor, a memory coupled to the processor, and instructions stored in the memory and executable by the processor to cause the apparatus to perform the method of any one of Examples 15 to 24.

[0169] It should be noted that the methods described herein describe possible implementations, that operations and steps may be rearranged or modified, and that other implementations are possible. Additionally, aspects of two or more methods may be combined.

[0170] Although aspects of LTE, LTE-A, LTE-A Pro, or NR systems may be described for example purposes, and the terminology of LTE, LTE-A, LTE-A Pro, or NR may be used throughout much of the description, the techniques described herein are applicable beyond LTE, LTE-A, LTE-A Pro, or NR networks. For example, the techniques described herein 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.

[0171] 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 referred to throughout the specification may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

[0172] The various illustrative blocks and components described herein may be implemented or performed using 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 alternatively, the processor may be any processor, controller, microcontroller, or state machine. A 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).

[0173] 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, the functions may be stored or transmitted as one or more instructions or code on a computer-readable medium. Other examples and implementations are 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 of these. Features that implement the functions may also be physically located in various locations, including being distributed so that portions of the functions are implemented in different physical locations.

[0174] Computer-readable media include non-transitory computer storage media and communication media, including any media that facilitates the transfer of computer programs from one place to another. Non-transitory storage media can be any available media that can be accessed by a general-purpose or special-purpose computer. As an example and not limitation, non-transitory computer-readable media may include random-access memory (RAM), read-only memory (ROM), electrically erasable programmable ROM (EEPROM), flash memory, compact disk (CD) ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other non-transitory media that can be used to carry or store desired program code components in the form of instructions or data structures and can be accessed by a general-purpose or special-purpose computer or a general-purpose or special-purpose processor. Similarly, any connection is properly referred to as a computer-readable medium. For example, if a coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL), or wireless technologies such as infrared, radio, and microwaves are used to transmit software from a website, server, or other remote source, 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. Disks and optical discs, as used herein, include CDs, laser discs, optical optical discs, digital versatile discs (DVDs), floppy disks, and Blu-ray discs, where disks typically reproduce data magnetically, while optical discs reproduce data optically with lasers. Combinations of the above are also included within the scope of computer-readable media.

[0175] As used herein, including in the claims, "or" used in a list of items (e.g., a list of items preceded by phrases such as "at least one" or "one or more") means an inclusive list, 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). In addition, as used herein, the phrase "based on" should not be interpreted as referring to a closed set of conditions. For example, an example step described as "based on condition A" can 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 interpreted in the same manner as the phrase "based at least in part on."

[0176] In the drawings, similar components or features may have the same reference label. In addition, various components of the same type may be distinguished by following the reference label with a dash and a second label, the second label being used to distinguish similar components. If only the first reference label is used in the specification, the description applies to any similar component having the same first reference label, regardless of the second reference label or any subsequent reference labels.

[0177] The descriptions set forth herein in conjunction with the accompanying drawings describe example configurations, but do not represent all examples that may be implemented or within the scope of the claims. The term "example" as used herein means "serving as an example, instance, or illustration," rather than "preferred" or "superior to other examples." The detailed description includes specific details for the purpose of providing an understanding of the described techniques. However, these techniques can be implemented without these specific details. In some cases, known structures and devices are shown in block diagram form to avoid obscuring the concepts of the described examples.

[0178] The description herein is intended to enable one of ordinary skill in the art to make or use the present disclosure. Various modifications to the present disclosure will be 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. Therefore, 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: receiving, from a network node, an indication of a level of accuracy for reporting channel state feedback to the network node, wherein receiving the indication of the level of accuracy comprises receiving an indication of a loss function corresponding to the level of accuracy for training a neural network pair, wherein the neural network pair comprises a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder; training the pair of neural networks using the loss function; sending coefficients of the second neural network to the network node for decoding the channel state feedback based at least in part on the training; receiving downlink data or a reference signal from the network node; as well as reporting the channel state feedback corresponding to the accuracy level to the network node based at least in part on the downlink data or the reference signal, The reporting of the channel state feedback corresponding to the accuracy level includes: Based at least in part on the training, encoding the channel state feedback using the first neural network at the encoder; and Report the encoded channel state feedback.

2. The method according to claim 1, wherein Training the neural network pair using the loss function includes: iteratively entering a channel state feedback input into the pair of neural networks and identifying a channel state feedback output from the pair of neural networks; determining, for each iteration, a difference between the channel state feedback input and the channel state feedback output using the loss function, wherein the difference comprises a loss; and Based at least in part on the determination, coefficients of the neural network pair are adjusted for each iteration to minimize the difference between the channel state feedback input and the channel state feedback output.

3. The method according to claim 1, further comprising: receiving, from the network node, an indication to train a plurality of neural network pairs based at least in part on a plurality of accuracy levels, the plurality of neural network pairs including the neural network pair; as well as Each neural network pair in the plurality of neural network pairs is trained based at least in part on a respective level of accuracy in the plurality of levels of accuracy.

4. The method according to claim 3, wherein: Receiving the indication of the level of accuracy comprises: An indication is received to use the neural network pair of the plurality of neural network pairs for reporting the channel state feedback.

5. The method according to claim 3, further comprising: The neural network pair of the plurality of neural network pairs is autonomously selected for use in reporting the channel state feedback.

6. The method according to claim 3, further comprising: Receive an indication of a subset of the plurality of neural network pairs for training by the UE.

7. The method according to claim 1, wherein The accuracy level is based at least in part on one or more of a subband, a spatial layer, or a channel tap to which the channel state feedback corresponds, the method further comprising: Data is received from the network node on the subband or spatial layer or according to the channel taps based at least in part on reporting the channel state feedback corresponding to the accuracy level.

8. The method according to claim 1, wherein The level of accuracy is based at least in part on a number of downlink transmissions comprising identical data that the UE failed to decode, the method further comprising: Based at least in part on reporting the channel state feedback corresponding to the level of accuracy, a retransmission of the same data that the UE failed to decode is received.

9. The method according to claim 1, further comprising: identifying a number of bits for reporting the channel state feedback based at least in part on the level of accuracy, wherein the number of bits is directly related to the level of accuracy; and The channel state feedback corresponding to the accuracy level is reported using the identified number of bits.

10. The method according to claim 9, further comprising: An indication of the number of bits used to report the channel state feedback is received based at least in part on the level of accuracy.

11. The method according to claim 1, wherein Receiving the indication of the level of accuracy comprises: The indication of the accuracy level is received in Radio Resource Control, RRC, signaling or a Medium Access Control, MAC, Control Element, MAC-CE.

12. A method for wireless communication at a network node, comprising: sending, to a user equipment (UE), an indication of a level of accuracy for reporting channel state feedback to the network node, wherein sending the indication of the level of accuracy comprises sending an indication of a loss function corresponding to the level of accuracy for training a neural network pair, wherein the neural network pair comprises a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder; receiving, from the UE, coefficients of the second neural network at the decoder for decoding the channel state feedback from the UE; sending downlink data or a reference signal to the UE; receiving channel state feedback corresponding to the level of accuracy from the UE based at least in part on transmitting the downlink data or reference signal to the UE; and The channel state feedback from the UE is decoded using the second neural network at the decoder.

13. An apparatus for performing wireless communication at a user equipment (UE), comprising: means for receiving, from a network node, an indication of a level of accuracy for reporting channel state feedback to the network node, wherein receiving the indication of the level of accuracy comprises receiving an indication of a loss function corresponding to the level of accuracy for training a neural network pair, wherein the neural network pair comprises a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder; means for training the pair of neural networks using the loss function; means for sending coefficients of the second neural network to the network node for use in decoding the channel state feedback based at least in part on the training; means for receiving downlink data or a reference signal from the network node; and means for reporting the channel state feedback corresponding to the accuracy level to the network node based at least in part on the downlink data or the reference signal, The component for reporting the channel state feedback corresponding to the accuracy level includes: means for encoding the channel state feedback using the first neural network at the encoder based at least in part on the training; and Means for reporting encoded channel state feedback.

14. An apparatus for performing wireless communication at a user equipment (UE), comprising: processor, a memory coupled to the processor; as well as instructions stored in the memory and executable by the processor to cause the apparatus to: receiving, from a network node, an indication of a level of accuracy for reporting channel state feedback to the network node, wherein receiving the indication of the level of accuracy comprises receiving an indication of a loss function corresponding to the level of accuracy for training a neural network pair, wherein the neural network pair comprises a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder; training the pair of neural networks using the loss function; sending coefficients of the second neural network to the network node for decoding the channel state feedback based at least in part on the training; receiving downlink data or a reference signal from the network node; and reporting the channel state feedback corresponding to the accuracy level to the network node based at least in part on the downlink data or the reference signal, The instruction to report the channel state feedback corresponding to the accuracy level may be executed by the processor to cause the apparatus to: encoding the channel state feedback at the encoder using the first neural network based at least in part on the training of the neural network pair; and Report the encoded channel state feedback.

15. The device according to claim 14, wherein The instructions to train the neural network pair using the loss function are executable by the processor to cause the apparatus to: iteratively entering a channel state feedback input into the pair of neural networks and identifying a channel state feedback output from the pair of neural networks; determining, for each iteration, a difference between the channel state feedback input and the channel state feedback output using the loss function, wherein the difference comprises a loss; and Based at least in part on the determination of the difference, coefficients of the neural network pair are adjusted for each iteration to minimize the difference between the channel state feedback input and the channel state feedback output.

16. The device according to claim 14, wherein The instructions are further executable by the processor to cause the apparatus to: receiving, from the network node, an indication to train a plurality of neural network pairs based at least in part on a plurality of accuracy levels, the plurality of neural network pairs including the neural network pair; and Each neural network pair in the plurality of neural network pairs is trained based at least in part on a respective level of accuracy in the plurality of levels of accuracy.

17. The device according to claim 16, wherein The instructions to receive the indication of the level of accuracy are executable by the processor to cause the apparatus to: An indication is received to use the neural network pair of the plurality of neural network pairs for reporting the channel state feedback.

18. The device according to claim 16, wherein The instructions are further executable by the processor to cause the apparatus to: The neural network pair of the plurality of neural network pairs is autonomously selected for use in reporting the channel state feedback.

19. The device according to claim 16, wherein The instructions are further executable by the processor to cause the apparatus to: Receive an indication of a subset of the plurality of neural network pairs for training by the UE.

20. The apparatus according to claim 14, wherein The accuracy level is based at least in part on one or more of a subband, a spatial layer, or a channel tap to which the channel state feedback corresponds, and wherein the instructions are further executable by the processor to cause the apparatus to: Data is received from the network node on the subband or spatial layer or according to the channel taps based at least in part on the reporting of the channel state feedback corresponding to the accuracy level.

21. The apparatus according to claim 14, wherein The level of accuracy is based at least in part on a number of downlink transmissions including identical data that the UE failed to decode, and wherein the instructions are further executable by the processor to cause the apparatus to: Based at least in part on the report of the channel state feedback corresponding to the level of accuracy, a retransmission of the same data that the UE failed to decode is received.

22. The apparatus according to claim 14, wherein The instructions are further executable by the processor to cause the apparatus to: identifying a number of bits for reporting the channel state feedback based at least in part on the level of accuracy, wherein the number of bits is directly related to the level of accuracy; and The channel state feedback corresponding to the accuracy level is reported using the identified number of bits.

23. The device according to claim 22, wherein The instructions are further executable by the processor to cause the apparatus to: An indication of the number of bits used to report the channel state feedback is received based at least in part on the level of accuracy.

24. The apparatus according to claim 14, wherein The instructions to receive the indication of the level of accuracy are executable by the processor to cause the apparatus to: Said indication of said level of accuracy is received in Radio Resource Control, RRC, signalling or in a Medium Access Control, MAC, Control Element, MAC-CE.

25. An apparatus for wireless communication at a network node, comprising: processor, a memory, coupled to the processor; as well as instructions stored in the memory and executable by the processor to cause the apparatus to: sending, to a user equipment (UE), an indication of a level of accuracy for reporting channel state feedback to the network node, wherein sending the indication of the level of accuracy comprises sending an indication of a loss function corresponding to the level of accuracy for training a neural network pair, wherein the neural network pair comprises a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder; receiving, from the UE, coefficients of the second neural network at the decoder for decoding the channel state feedback from the UE; sending downlink data or a reference signal to the UE; receiving, from the UE, channel state feedback corresponding to the level of accuracy based at least in part on the transmission of the downlink data or reference signal to the UE; and The channel state feedback from the UE is decoded using the second neural network at the decoder.

26. The device according to claim 25, wherein The instructions are further executable by the processor to cause the apparatus to: An indication is sent that the UE trains a plurality of neural network pairs based at least in part on a plurality of accuracy levels, each neural network pair comprising a first neural network for encoding the channel state feedback at an encoder and a second neural network for decoding the channel state feedback at a decoder.

27. The apparatus according to claim 25, wherein The instructions to send the indication of the accuracy level for reporting channel state feedback are executable by the processor to cause the apparatus to: Indications of different accuracy levels are sent for reporting channel state feedback for different subbands, spatial layers, channel taps, or in response to failure to decode different numbers of downlink transmissions comprising the same data.

28. The apparatus according to claim 25, wherein The accuracy level comprises a first accuracy level for reporting channel state feedback to be used for scheduling a first downlink data transmission, and wherein the instructions are further executable by the processor to cause the apparatus to: An indication of a second level of accuracy is transmitted for reporting channel state feedback to be used to schedule a second downlink transmission, the first level of accuracy being different from the second level of accuracy.

29. The apparatus according to claim 25, wherein The instructions are further executable by the processor to cause the apparatus to: transmitting an indication of a number of bits for the UE to use for reporting the channel state feedback based at least in part on the accuracy level, wherein the number of bits is directly related to the accuracy level; and The channel state feedback corresponding to the accuracy level using the identified number of bits is received.

30. A computer readable medium having stored thereon instructions that, when executed by a processor, cause the processor to perform the method of any one of claims 1-12.

31. A computer program product comprising computer instructions which, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 12.

Images