Model management for channel state estimation and feedback

Abstract

Certain aspects of the present disclosure provide techniques for wireless communication. One aspect provides a method of wireless communication by a user equipment (UE), the method comprising: receiving a reference signal from a network entity; processing the reference signal with a machine learning model to generate a machine learning model output; and determining an action to take based on the machine learning model output and a model monitoring configuration.

Description

Background Art Technical Field

[0001] Aspects of the present disclosure relate to wireless communications and, more particularly, to techniques for managing models for channel state estimation and feedback.

[0002] Related technologies

[0003] Wireless communication systems are widely deployed to provide various telecommunication services such as telephony, video, data, messaging, broadcasts or other similar types of services. These wireless communication systems may employ multiple-access technologies capable of supporting communication with multiple users by sharing available wireless communication system resources with the multiple users.

[0004] Although wireless communication systems have made tremendous technical progress over the years, challenges still exist. For example, complex and dynamic environments can still attenuate or block signals between wireless transmitters and wireless receivers. Therefore, there is a continued desire to improve the technical performance of wireless communication systems, including, for example: improving the speed and data carrying capacity of communications, improving the efficiency of using shared communication media, reducing the power used by transmitters and receivers when performing communications, improving the reliability of wireless communications, avoiding redundant transmission and / or reception and related processing, improving the coverage area of ​​wireless communications, increasing the number and types of devices that can access wireless communication systems, increasing the ability of different types of devices to communicate with each other, increasing the number and types of wireless communication media available for use, etc. Therefore, there is a need to further improve wireless communication systems to overcome the above-mentioned technical challenges and other challenges. Summary of the invention

[0005] One aspect provides a method for wireless communication by a user equipment (UE). The method includes: receiving a reference signal from a network entity; processing the reference signal with a machine learning model to generate a machine learning model output; and determining an action to be taken based on the machine learning model output and a model monitoring configuration.

[0006] Another aspect provides a method for wireless communication by a network entity. The method includes: transmitting a model monitoring configuration to a user equipment; transmitting a reference signal to the user equipment; and receiving one of a model variance indication or a model failure indication from the user equipment based on the reference signal.

[0007] Other aspects provide: an apparatus operable to, configured to, or otherwise adapted to perform any one or more of the foregoing methods and / or those described elsewhere herein; a non-transitory computer-readable medium comprising instructions that, when executed by a processor of the apparatus, cause the apparatus to perform the foregoing methods and those described elsewhere herein; a computer program product embodied on a computer-readable storage medium, the computer program product comprising code for performing the foregoing methods and those described elsewhere herein; and / or an apparatus comprising components for performing the foregoing methods and those described elsewhere herein. By way of example, an apparatus may include a processing system, a device having a processing system, or a processing system cooperating through one or more networks.

[0008] For purposes of illustration, the following description and drawings set forth certain features. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The accompanying drawings depict certain features of the various aspects described herein and should not be considered to limit the scope of the disclosure.

[0010] Figure 1 An example wireless communication network is depicted.

[0011] Figure 2 An example disaggregated base station architecture is depicted.

[0012] Figure 3 Aspects of an example base station and example user equipment are depicted.

[0013] Figure 4A , Figure 4B , Figure 4C and Figure 4D Various example aspects of data structures for a wireless communication network are described.

[0014] Figure 5 Depicts an example of monitoring the performance of a machine learning model over time.

[0015] Figure 6 Depicts an example of a model monitoring framework.

[0016] Figure 7 Depicts an example of a model monitoring configuration for inference mode.

[0017] Figure 8 Depicts an example of a model monitoring configuration for monitoring mode.

[0018] Fig. 9 Depicts examples of model monitoring configurations for inference and monitoring modes.

[0019] FIG. 10A to FIG. 10DVarious examples of model monitoring mode switching are depicted.

[0020] FIG. 11A to FIG. 11B Depicted is an example method for counting model variance events.

[0021] Fig.12 Depicted are aspects related to reporting model variance events and model failures, such as when user equipment is operating in monitoring mode.

[0022] Fig.13 Depicted is an example of using low-density CSI-RS for channel estimation based on a machine learning model.

[0023] Fig.14 An example of paired CSI-RS resources between a first (target) set of resources and a second (reference) set of resources is depicted.

[0024] Fig.15 An example of a model monitoring method for a machine learning based channel estimation model is depicted.

[0025] Fig.16 A method for wireless communication is described.

[0026] Fig.17

[0013] Another method for wireless communication is described.

[0027] Fig.18 Aspects of an example communications device are depicted.

[0028] Fig.19 Aspects of another example communications device are depicted. DETAILED DESCRIPTION

[0029] Aspects of the present disclosure provide apparatus, methods, processing systems, and computer-readable media for managing models for channel state estimation and feedback.

[0030] Understanding the channel state between devices communicating in a wireless communication system is an important aspect of improving the performance of wireless communication. Conventionally, many techniques have been used to measure channel state and report feedback so that performance can be improved. However, such conventional techniques are usually relatively slow, power-consuming, and fixed in method.

[0031] Machine learning provides the opportunity to improve many conventional techniques for measuring channel states and reporting feedback. For example, machine learning models can reduce the number of resource elements required to estimate channel states and improve estimates of values ​​used when reporting channel states. However, because wireless environments are often extremely dynamic, it is important to be able to monitor the performance of machine learning models that implement key channel state measurement and feedback procedures, and take remedial actions in situations where, for example, the model begins to perform poorly. Such remedial actions may include, for example, falling back to a baseline model (e.g., a non-machine learning model) to perform various aspects until the machine learning model can be reconfigured to maintain optimal performance.

[0032] The various aspects described herein are generally related to methods for monitoring performance and detecting failures associated with machine learning models for channel state measurement and feedback procedures. In various aspects, the network may be configured and / or the user equipment may implement various modes for monitoring model performance. In some aspects, model performance is monitored by determining output variance events and reporting such variance events to the network and / or using such variance events to determine when the model has become unreliable or "failed". In some aspects, a model variance event may be an out-of-distribution (OOD) event, which generally refers to a machine learning model generating an uncertain output based on an input different from its training data.

[0033] By actively monitoring the performance of the machine learning model and detecting failures with respect to the machine learning model, with or without network assistance, the various aspects described herein achieve the benefits of the machine learning model, such as faster, more energy-efficient, and more accurate operation, while reducing the likelihood that the performance of the machine learning model will degrade over time. Such degradation may, for example, result from exposure of the machine learning model to new environments and new conditions that were not initially considered during the training of the machine learning model. In the context of the various examples described herein, such exposure may include user equipment using the machine learning model to perform channel estimation and predict channel state information feedback in a radio environment that is different from the environment considered during the training of the model. Detection of such degradation allows reconfiguration (e.g., retraining) of the machine learning model to maintain a state of technical performance, and allows for a fallback to a baseline model at the same time.

[0034] Thus, the aspects described herein that implement the robust use of machine learning models for channel state measurement and feedback procedures enhance wireless communication performance overall, and more specifically, enhance wireless communication performance by reducing power usage, increasing battery life, improving spectral efficiency, reducing latency, and reducing network overhead, to name a few technical improvements.

[0035] Introduction to wireless communication networks

[0036] The techniques and methods described herein can be used in various wireless communication networks. Although various aspects may be described herein using terms commonly associated with 3G, 4G and / or 5G wireless technologies, various aspects of the present disclosure may also be applicable to other communication systems and standards not explicitly mentioned herein.

[0037] Figure 1 An example of a wireless communication network 100 is depicted in which various aspects described herein may be implemented.

[0038] In general, the wireless communication network 100 includes various network entities (alternatively, network elements or network nodes). A network entity is typically a communication device and / or a communication function performed by a communication device (e.g., user equipment (UE), base station (BS), component of a BS, server, etc.). For example, various functions of a network and various devices associated with and interacting with the network may be considered network entities. In addition, the wireless communication network 100 includes ground aspects, such as ground-based network entities (e.g., BS 102), and non-ground aspects, such as satellites 140 and aircraft 145, which may include airborne network entities (e.g., one or more BSs) capable of communicating with other network elements (e.g., ground BSs) and user equipment.

[0039] In the depicted example, the wireless communication network 100 includes a BS 102, a UE 104, and one or more core networks (such as an evolved packet core (EPC) 160 and a 5G core (5GC) network 190) that interoperate to provide communication services over various communication links (including wired and wireless links).

[0040] Figure 1 Various example UEs 104 are depicted, which may more generally include: a cellular phone, a smart phone, a Session Initiation Protocol (SIP) phone, a laptop, a personal digital assistant (PDA), a satellite radio, a global positioning system, a multimedia device, a video device, a digital audio player, a camera, a game console, a tablet computer, a smart device, a wearable device, a vehicle, an electric meter, a gas pump, a large or small kitchen appliance, a healthcare device, an implant, a sensor / actuator, a display, an Internet of Things (IoT) device, an Always-On (AON) device, an edge processing device, or other similar devices. UE 104 may also be more generally referred to as a mobile device, a wireless device, a wireless communication device, a station, a mobile station, a subscriber station, a mobile subscriber station, a mobile unit, a subscriber unit, a wireless unit, a remote unit, a remote device, an access terminal, a mobile terminal, a wireless terminal, a remote terminal, a handset, and others.

[0041] BS 102 wirelessly communicates with (e.g., sends signals to or receives signals from) UE 104 via communication link 120. Communication link 120 between BS 102 and UE 104 may include uplink (UL) (also referred to as a reverse link) transmissions from UE 104 to BS 102 and / or downlink (DL) (also referred to as a forward link) transmissions from BS 102 to UE 104. In various aspects, communication link 120 may use multiple-input multiple-output (MIMO) antenna technology, including spatial multiplexing, beamforming, and / or transmit diversity.

[0042] BS 102 may generally include: Node B, enhanced Node B (eNB), next generation enhanced Node B (ng-eNB), next generation Node B (gNB or gNodeB), access point, base transceiver station, radio base station, radio transceiver, transceiver function, transmission and reception point, and / or other. Each of BS 102 may provide communication coverage for a corresponding geographic coverage area 110, which may sometimes be referred to as a cell, and may overlap in some cases (e.g., small cell 102' may have a coverage area 110' that overlaps with the coverage area 110 of a macro cell). For example, a BS may provide communication coverage for a macro cell (covering a relatively large geographic area), a pico cell (covering a relatively small geographic area, such as a stadium), a femto cell (relatively small geographic area (e.g., a home)), and / or other types of cells.

[0043] Although BS102 is depicted as a single communication device in various aspects, BS102 can be implemented in various configurations. For example, one or more components of the base station may be decomposed, including a central unit (CU), one or more distributed units (DU), one or more radio units (RU), a near real-time (near RT) RAN intelligent controller (RIC) or a non-real-time (non-RT) RIC, to name a few examples. In another example, various aspects of the base station may be virtualized. More generally, a base station (e.g., BS102) may include components located at a single physical location or components located at various physical locations. In an example in which the base station includes components located at various physical locations, the various components may each perform a function so that the various components together implement functions similar to those of a base station located at a single physical location. In some aspects, a base station including components located at various physical locations may be referred to as a decomposed radio access network architecture (such as an open RAN (O-RAN) or virtualized RAN (VRAN) architecture). Figure 2 An example decomposed base station architecture is depicted and described.

[0044] Different BSs 102 within the wireless communication network 100 may also be configured to support different radio access technologies (such as 3G, 4G, and / or 5G). For example, a BS 102 configured for 4G LTE (collectively referred to as the Evolved Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access Network (E-UTRAN)) may interface with the EPC 160 via a first backhaul link 132 (e.g., an S1 interface). A BS 102 configured for 5G (e.g., 5G NR or Next Generation RAN (NG-RAN)) may interface with the 5GC 190 via a second backhaul link 184. The BSs 102 may communicate with each other directly or indirectly (e.g., via the EPC 160 or the 5GC 190) over a third backhaul link 134 (e.g., an X2 interface), which may be wired or wireless.

[0045] The wireless communication network 100 may subdivide the electromagnetic spectrum into various categories, frequency bands, channels, or other characteristics. In some aspects, the subdivision is provided based on wavelength and frequency, where frequency may also be referred to as a carrier, subcarrier, channel, tone, or subband. For example, 3GPP currently defines frequency range 1 (FR1) as including 410 MHz-7125 MHz, which is often (interchangeably) referred to as "below 6 GHz". Similarly, 3GPP currently defines frequency range 2 (FR2) as including 24,250 MHz-52,600 MHz, which is sometimes (interchangeably) referred to as "millimeter wave" ("mmW" or "mmWave"). A base station (e.g., a mmWave base station such as BS180) configured to communicate using mmWave / near mmWave radio bands may utilize beamforming (e.g., 182) with a UE (e.g., 104) to improve path loss and range.

[0046] The communication link 120 between the BS 102 and, for example, the UE 104 may be through one or more carriers, which may have different bandwidths (e.g., 5 MHz, 10 MHz, 15 MHz, 20 MHz, 100 MHz, 400 MHz, and / or other MHz) and may be aggregated in various aspects. The carriers may or may not be adjacent to each other. The allocation of carriers may be asymmetric with respect to DL and UL (e.g., more or fewer carriers may be allocated for DL ​​compared to UL).

[0047] Communications using higher frequency bands may have higher path loss and shorter range than lower frequency communications. Therefore, some base stations (e.g. Figure 1180) may utilize beamforming 182 with UE 104 to improve path loss and range. For example, BS 180 and UE 104 may each include multiple antennas, such as antenna elements, antenna panels, and / or antenna arrays, to facilitate beamforming. In some cases, BS 180 may send beamformed signals to UE 104 in one or more transmit directions 182'. UE 104 may receive beamformed signals from BS 180 in one or more receive directions 182". UE 104 may also send beamformed signals to BS 180 in one or more transmit directions 182". BS 180 may also receive beamformed signals from UE 104 in one or more receive directions 182'. BS 180 and UE 104 may then perform beam training to determine the best receive and transmit directions for each of BS 180 and UE 104. It is noteworthy that the transmit direction and receive direction of BS 180 may be the same or different. Similarly, the transmit direction and receive direction of UE 104 may be the same or different.

[0048] Wireless communication network 100 also includes a Wi-Fi AP 150 that communicates with a Wi-Fi station (STA) 152 via a communication link 154 in, for example, the 2.4 GHz and / or 5 GHz unlicensed spectrum.

[0049] Certain UEs 104 may communicate with each other using a device-to-device (D2D) communication link 158. The D2D communication link 158 may use one or more sidelink channels, such as a physical sidelink broadcast channel (PSBCH), a physical sidelink discovery channel (PSDCH), a physical sidelink shared channel (PSSCH), a physical sidelink control channel (PSCCH), and / or a physical sidelink feedback channel (PSFCH).

[0050] The EPC 160 may include various functional components, including: a mobility management entity (MME) 162, other MMEs 164, a serving gateway 166, a multimedia broadcast multicast service (MBMS) gateway 168, a broadcast multicast service center (BM-SC) 170, and / or a packet data network (PDN) gateway 172, such as in the depicted example. The MME 162 may communicate with a home subscriber server (HSS) 174. The MME 162 is a control node that handles signaling between the UE 104 and the EPC 160. In general, the MME 162 provides bearer and connection management.

[0051] Generally, user Internet Protocol (IP) packets are delivered through the Serving Gateway 166, which itself is connected to the PDN Gateway 172. The PDN Gateway 172 provides UE IP address allocation and other functions. The PDN Gateway 172 and the BM-SC 170 are connected to the IP Services 176, which may include, for example, the Internet, an Intranet, an IP Multimedia Subsystem (IMS), a Packet Switched (PS) streaming media service, and / or other IP services.

[0052] The BM-SC 170 may provide functionality for MBMS user service configuration and delivery. The BM-SC 170 may serve as an entry point for content provider MBMS delivery, may be used to authorize and initiate MBMS bearer services within a public land mobile network (PLMN), and / or may be used to schedule MBMS delivery. The MBMS Gateway 168 may be used to distribute MBMS services to BSs 102 belonging to a multicast broadcast single frequency network (MBSFN) area broadcasting a specific service, and / or may be responsible for session management (start / stop) and for collecting eMBMS related charging information.

[0053] 5GC 190 may include various functional components, including: access and mobility management function (AMF) 192, other AMFs 193, session management function (SMF) 194, and user plane function (UPF) 195. AMF 192 may communicate with unified data management (UDM) 196.

[0054] AMF 192 is a control node that handles signaling between UE 104 and 5GC 190. AMF 192 provides, for example, Quality of Service (QoS) flow and session management.

[0055] Internet Protocol (IP) packets are delivered through UPF 195, which is connected to IP Services 197 and provides IP address allocation for UEs and other functions for 5GC 190. IP Services 197 may include, for example, the Internet, Intranet, IMS, PS streaming services, and / or other IP services.

[0056] In various aspects, a network entity or network node may be implemented as a converged base station, a decomposed base station, a component of a base station, an integrated access and backhaul (IAB) node, a relay node, a sidelink node, to name a few examples.

[0057] Figure 2An example disaggregated base station 200 architecture is depicted. The disaggregated base station 200 architecture may include one or more central units (CUs) 210 that may communicate directly with a core network 220 via a backhaul link, or indirectly with the core network 220 through one or more disaggregated base station units, such as a near real-time (near-RT) RAN intelligent controller (RIC) 225 via an E2 link, or a non-real-time (non-RT) RIC 215 associated with a service management and orchestration (SMO) framework 205, or both. The CU 210 may communicate with one or more distributed units (DUs) 230 via respective midhaul links, such as an F1 interface. The DU 230 may communicate with one or more radio units (RUs) 240 via respective fronthaul links. The RU 240 may communicate with respective UEs 104 via one or more radio frequency (RF) access links. In some implementations, a UE 104 may be served simultaneously by multiple RUs 240.

[0058] Each of the units (e.g., CU 210, DU 230, RU 240, and near-RT RIC 225, non-RT RIC 215, and SMO framework 205) may include or be coupled to one or more interfaces configured to receive or send signals, data, or information (collectively, signals) via a wired or wireless transmission medium. Each of the units or an associated processor or controller that provides instructions to the communication interface of the unit may be configured to communicate with one or more of the other units via a transmission medium. For example, the units may include a wired interface configured to receive or send signals to one or more of the other units via a wired transmission medium. Additionally or alternatively, the unit may include a wireless interface that may include a receiver, a transmitter, or a transceiver (such as a radio frequency (RF) transceiver) configured to receive or send signals, or both, to one or more of the other units on a wireless transmission medium.

[0059] In some aspects, CU 210 may host one or more higher layer control functions. Such control functions may include radio resource control (RRC), packet data convergence protocol (PDCP), service data adaptation protocol (SDAP), etc. Each control function may be implemented using an interface configured to communicate signals with other control functions hosted by CU 210. CU 210 may be configured to handle user plane functions (e.g., central unit-user plane (CU-UP)), control plane functions (e.g., central unit-control plane (CU-CP)), or a combination thereof. In some specific implementations, CU 210 may be logically split into one or more CU-UP units and one or more CU-CP units. When implemented in an O-RAN configuration, the CU-UP unit may communicate bidirectionally with the CU-CP unit via an interface (such as an E1 interface). CU 210 may be implemented to communicate with DU 230 as needed for network control and signaling.

[0060] DU 230 may correspond to a logical unit that includes one or more base station functions for controlling the operation of one or more RUs 240. In some aspects, DU 230 may host one or more of a radio link control (RLC) layer, a medium access control (MAC) layer, and one or more high physical (PHY) layers (such as modules for forward error correction (FEC) encoding and decoding, scrambling, modulation and demodulation, etc.) at least in part according to a functional split such as those defined by the 3rd Generation Partnership Project (3GPP). In some aspects, DU 230 may also host one or more low PHY layers. Each layer (or module) may be implemented using an interface that is configured to communicate signals with other layers (and modules) hosted by DU 230 or with control functions hosted by CU 210.

[0061] The lower layer functions may be implemented by one or more RUs 240. In some deployments, the RU 240 controlled by the DU 230 may correspond to a logical node that hosts RF processing functions or low PHY layer functions (such as performing fast Fourier transform (FFT), inverse FFT (iFFT), digital beamforming, physical random access channel (PRACH) extraction and filtering, etc.), or both, based at least in part on functional splitting such as lower layer functional splitting. In such an architecture, the RU 240 may be implemented to handle over-the-air (OTA) communications with one or more UEs 104. In some specific implementations, real-time and non-real-time aspects of control plane and user plane communications with the RU 240 may be controlled by the corresponding DU 230. In some scenarios, this configuration may enable the DU 230 and the CU 210 to be implemented in a cloud-based RAN architecture such as a vRAN architecture.

[0062] The SMO framework 205 may be configured to support RAN deployment and configuration of non-virtualized network elements and virtualized network elements. For non-virtualized network elements, the SMO framework 205 may be configured to support the deployment of dedicated physical resources for RAN coverage requirements, which may be managed via an operation and maintenance interface (such as an O1 interface). For virtualized network elements, the SMO framework 205 may be configured to interact with a cloud computing platform (such as an open cloud (O-Cloud) 290) to perform network element lifecycle management (such as an example virtualized network element) via a cloud computing platform interface (such as an O2 interface). Such virtualized network elements may include, but are not limited to, CU 210, DU 230, RU 240, and near-RT RIC 225. In some specific implementations, the SMO framework 205 may communicate with hardware aspects of the 4G RAN such as an open eNB (O-eNB) 211 via the O1 interface. Additionally, in some specific implementations, the SMO framework 205 may communicate directly with one or more RUs 240 via the O1 interface. The SMO framework 205 may also include a non-RT RIC 215 configured to support the functionality of the SMO framework 205 .

[0063] The non-RT RIC 215 may be configured to include logic functions that implement non-real-time control and optimization of RAN elements and resources, artificial intelligence / machine learning (AI / ML) workflows including model training and updating, or policy-based guidance of applications / features in the near-RT RIC 225. The non-RT RIC 215 may be coupled to or in communication with the near-RT RIC 225 (such as via an A1 interface). The near-RT RIC 225 may be configured to include logic functions that implement near-real-time control and optimization of RAN elements and resources via data collection and actions through an interface (such as via an E2 interface) that connects one or more CUs 210, one or more DUs 230, or both, and the O-eNB with the near-RT RIC 225.

[0064] In some implementations, in order to generate an AI / ML model to be deployed in the near-RT RIC 225, the non-RT RIC 215 may receive parameters or external enrichment information from an external server. Such information may be utilized by the near-RT RIC 225 and may be received from a non-network data source or from a network function at the SMO framework 205 or the non-RT RIC 215. In some examples, the non-RT RIC 215 or the near-RT RIC 225 may be configured to tune RAN behavior or performance. For example, the non-RT RIC 215 may monitor long-term trends and patterns of performance and employ AI / ML models to perform corrective actions through the SMO framework 205 (such as via reconfiguration of O1) or via creation of RAN management policies (such as A1 policies).

[0065] Figure 3 Aspects of an example BS 102 and UE 104 are depicted.

[0066] In general, BS 102 includes various processors (e.g., 320, 330, 338, and 340), antennas 334a-334t (collectively 334), transceivers 332a-332t (collectively 332) including modulators and demodulators, and other aspects that enable wireless transmission of data (e.g., data source 312) and wireless reception of data (e.g., data sink 339). For example, BS 102 can transmit and receive data between BS 102 and UE 104. BS 102 includes a controller / processor 340 that can be configured to implement various functions described herein related to wireless communication.

[0067] Generally speaking, the UE 104 includes various processors (e.g., 358, 364, 366, and 380), antennas 352a-352r (collectively, 352), transceivers 354a-354r (collectively, 354) including modulators and demodulators, and other aspects that enable wireless transmission of data (e.g., retrieved from a data source 362) and wireless reception of data (e.g., provided to a data sink 360). The UE 104 includes a controller / processor 380 that can be configured to implement various functions described herein related to wireless communications.

[0068] With respect to example downlink transmissions, BS 102 includes a transmit processor 320 that can receive data from a data source 312 and control information from a controller / processor 340. The control information can be for a physical broadcast channel (PBCH), a physical control format indicator channel (PCFICH), a physical HARQ indicator channel (PHICH), a physical downlink control channel (PDCCH), a group common PDCCH (GC PDCCH), and / or others. In some examples, the data can be for a physical downlink shared channel (PDSCH).

[0069] The transmit processor 320 may process (e.g., encode and symbol map) data and control information to obtain data symbols and control symbols, respectively. The transmit processor 320 may also generate reference symbols (such as for a primary synchronization signal (PSS), a secondary synchronization signal (SSS), a PBCH demodulation reference signal (DMRS), and a channel state information reference signal (CSI-RS)).

[0070] The transmit (TX) multiple-input multiple-output (MIMO) processor 330 may perform spatial processing (e.g., pre-decoding) on ​​data symbols, control symbols, and / or reference symbols, where applicable, and may provide an output symbol stream to a modulator (MOD) in a transceiver 332a-332t. Each modulator in the transceiver 332a-332t may process a corresponding output symbol stream to obtain an output sample stream. Each modulator may further process (e.g., convert to analog, amplify, filter, and up-convert) the output sample stream to obtain a downlink signal. The downlink signals from the modulators in the transceivers 332a-332t may be transmitted via antennas 334a-334t, respectively.

[0071] To receive downlink transmissions, UE 104 includes antennas 352a-352r that can receive downlink signals from BS 102 and can provide received signals to demodulators (DEMODs) in transceivers 354a-354r, respectively. Each demodulator in transceivers 354a-354r can condition (e.g., filter, amplify, downconvert, and digitize) a corresponding received signal to obtain input samples. Each demodulator can further process the input samples to obtain received symbols.

[0072] A MIMO detector 356 may obtain received symbols from all demodulators in transceivers 354a-354r, perform MIMO detection on the received symbols if applicable, and provide detected symbols. A receive processor 358 may process (e.g., demodulate, deinterleave, and decode) the detected symbols, provide decoded data for UE 104 to a data sink 360, and provide decoded control information to a controller / processor 380.

[0073] With respect to example uplink transmissions, the UE 104 also includes a transmit processor 364 that may receive and process data from a data source 362 (e.g., for a PUSCH) and control information from a controller / processor 380 (e.g., for a physical uplink control channel (PUCCH)). The transmit processor 364 may also generate reference symbols for reference signals (e.g., a sounding reference signal (SRS)). The symbols from the transmit processor 364 may be pre-decoded by a TX MIMO processor 366, if applicable, further processed by a modulator in the transceivers 354a-354r (e.g., for SC-FDM), and transmitted to the BS 102.

[0074] At BS 102, uplink signals from UE 104 may be received by antennas 334 a-334 t, processed by demodulators in transceivers 332 a-332 t, detected by MIMO detector 336 if applicable, and further processed by receive processor 338 to obtain decoded data and control information transmitted by UE 104. Receive processor 338 may provide the decoded data to a data sink 339 and the decoded control information to a controller / processor 340.

[0075] Memory 342 and memory 382 may store data and program codes for BS 102 and UE 104, respectively.

[0076] Scheduler 344 may schedule UEs for data transmission on the downlink and / or uplink.

[0077] In various aspects, BS 102 may be described as sending and receiving various types of data associated with the methods described herein. In these contexts, "sending" may refer to various mechanisms for outputting data, such as outputting data from a data source 312, a scheduler 344, a memory 342, a transmit processor 320, a controller / processor 340, a TX MIMO processor 330, a transceiver 332a-332t, an antenna 334a-334t, and / or other aspects described herein. Similarly, "receiving" may refer to various mechanisms for obtaining data, such as obtaining data from an antenna 334a-334t, a transceiver 332a-332t, a RX MIMO detector 336, a controller / processor 340, a receive processor 338, a scheduler 344, a memory 342, and / or other aspects described herein.

[0078] In various aspects, the UE 104 may also be described as sending and receiving various types of data associated with the methods described herein. In these contexts, "sending" may refer to various mechanisms for outputting data, such as outputting data from a data source 362, a memory 382, ​​a transmit processor 364, a controller / processor 380, a TX MIMO processor 366, a transceiver 354a-354t, an antenna 352a-352t, and / or other aspects described herein. Similarly, "receiving" may refer to various mechanisms for obtaining data, such as obtaining data from an antenna 352a-352t, a transceiver 354a-354t, a RX MIMO detector 356, a controller / processor 380, a receive processor 358, a memory 382, ​​and / or other aspects described herein.

[0079] In some aspects, the processor may be configured to perform various operations (such as those associated with the methods described herein) and send (output) data to or receive (obtain) data from another interface configured to send or receive data, respectively.

[0080] Figure 4A , Figure 4B , Figure 4C and Figure 4D Describes a method for use in a wireless communication network such as Figure 1 Various aspects of the data structure of the wireless communication network 100).

[0081] Specifically, Figure 4A is a diagram 400 illustrating an example of a first subframe within a 5G (e.g., 5G NR) frame structure, Figure 4B is a diagram 430 illustrating an example of a DL channel within a 5G subframe, Figure 4C is a diagram 450 illustrating an example of a second subframe within a 5G frame structure, and Figure 4D FIG480 is a diagram illustrating an example of UL channels within a 5G subframe.

[0082] Wireless communication networks can utilize Orthogonal Frequency Division Multiplexing (OFDM) with a Cyclic Prefix (CP) on both the uplink and downlink. Such systems can also support half-duplex operation using Time Division Duplex (TDD). OFDM and Single Carrier Frequency Division Multiplexing (SC-FDM) can be used to maximize the system bandwidth (e.g., Figure 4B and Figure 4D The modulation symbols may be transmitted in the frequency domain using OFDM and / or in the time domain using SC-FDM.

[0083] The wireless communication frame structure may be frequency division duplex (FDD), where for a particular set of subcarriers, subframes within the set of subcarriers are dedicated to either DL or UL. The wireless communication frame structure may also be time division duplex (TDD), where for a particular set of subcarriers, subframes within the set of subcarriers are dedicated to both DL and UL.

[0084] exist Figure 4A and Figure 4CIn the wireless communication frame structure, TDD is used, where D is DL, U is UL, and X is used flexibly between DL / UL. The UE can be configured with a time slot format (dynamically configured by DL control information (DCI) or semi-statically / statically configured by radio resource control (RRC) signaling) through the received time slot format indicator (SFI). In the depicted example, the 10ms frame is divided into 10 equally sized 1ms subframes. Each subframe may include one or more time slots. In some examples, each time slot may include 7 or 14 symbols, depending on the time slot format. The subframe may also include micro-time slots, which typically have fewer symbols than the entire time slot. Other wireless communication technologies may have different frame structures and / or different channels.

[0085] In certain aspects, the number of time slots within a subframe is based on the time slot configuration and the parameter set. For example, for time slot configuration 0, different parameter sets (μ) 0 to 5 allow for 1, 2, 4, 8, 16, and 32 time slots per subframe, respectively. For time slot configuration 1, different parameter sets 0 to 2 allow for 2, 4, and 8 time slots per subframe, respectively. Thus, for time slot configuration 0 and parameter set μ, there are 14 symbols / time slot and 2μ time slots / subframe. The subcarrier spacing and symbol length / duration are a function of the parameter set. The subcarrier spacing may be equal to 2 μ × 15kHz, where μ is parameter set 0 to 5. Thus, parameter set μ=0 has a subcarrier spacing of 15kHz, while parameter set μ=5 has a subcarrier spacing of 480kHz. The symbol length / duration is inversely related to the subcarrier spacing. Figure 4A , Figure 4B , Figure 4C and Figure 4D An example is provided for slot configuration 0 with 14 symbols per slot and parameter set μ=2 with 4 slots per subframe. The slot duration is 0.25 ms, the subcarrier spacing is 60 kHz, and the symbol duration is approximately 16.67 μs.

[0086] like Figure 4A , Figure 4B , Figure 4C and Figure 4D As depicted in , a resource grid can be used to represent a frame structure. Each time slot includes a resource block (RB) (also called a physical RB (PRB)) extending, for example, 12 consecutive subcarriers. The resource grid is divided into a plurality of resource elements (REs). The number of bits carried by each RE depends on the modulation scheme.

[0087] like Figure 4A As illustrated, some REs carry data for UEs (e.g. Figure 1 and Figure 3The RS may include a demodulation RS (DMRS) and / or a channel state information reference signal (CSI-RS) for channel estimation at the UE. The RS may also include a beam measurement RS (BRS), a beam refinement RS (BRRS), and / or a phase tracking RS (PT-RS).

[0088] Figure 4B Examples of various DL channels within a subframe of a frame are illustrated.The Physical Downlink Control Channel (PDCCH) carries DCI within one or more Control Channel Elements (CCEs), each CCE comprising, for example, nine RE Groups (REGs), each REG comprising, for example, four consecutive REs in an OFDM symbol.

[0089] The primary synchronization signal (PSS) may be in symbol 2 of a particular subframe of a frame. The PSS is transmitted by a UE (e.g., Figure 1 and Figure 3 104) is used to determine subframe / symbol timing and physical layer identification.

[0090] A Secondary Synchronization Signal (SSS) may be within symbol 4 of a particular subframe of a frame. The SSS is used by the UE to determine the physical layer cell identity group number and radio frame timing.

[0091] Based on the physical layer identifier and the physical layer cell identifier group number, the UE can determine the physical cell identifier (PCI). Based on the PCI, the UE can determine the location of the aforementioned DMRS. The physical broadcast channel (PBCH) carrying the master information block (MIB) can be logically grouped with the PSS and SSS to form a synchronization signal (SS) / PBCH block. The MIB provides the number of RBs in the system bandwidth and the system frame number (SFN). The physical downlink shared channel (PDSCH) carries user data, broadcast system information (such as system information blocks (SIBs)) that is not sent via the PBCH, and / or paging messages.

[0092] like Figure 4CAs illustrated, some REs carry DMRS for channel estimation at the base station (indicated as R for a particular configuration, but other DMRS configurations are also possible). The UE may send DMRS for PUCCH and DMRS for PUSCH. The PUSCH DMRS may be sent, for example, in the first one or two symbols of the PUSCH. The PUCCH DMRS may be sent in different configurations depending on whether a short PUCCH or a long PUCCH is sent and depending on the specific PUCCH format used. UE 104 may send a sounding reference signal (SRS). The SRS may be sent, for example, in the last symbol of a subframe. The SRS may have a comb structure, and the UE may send the SRS on one of the comb structures in the comb structure. The SRS may be used by the base station for channel quality estimation to achieve frequency-dependent scheduling of the UL.

[0093] Figure 4D Examples of various UL channels within a subframe of a frame are illustrated. The PUCCH may be located at a position as indicated in one configuration. The PUCCH carries uplink control information (UCI), such as scheduling requests, channel quality indicator (CQI), precoding matrix indicator (PMI), rank indicator (RI), and HARQ ACK / NACK feedback. The PUSCH carries data and may additionally be used to carry a buffer status report (BSR), a power headroom report (PHR), and / or UCI.

[0094] Aspects related to managing models for channel state estimation and feedback

[0095] Conventionally, several techniques have been used to help determine the channel state between wireless communication devices so that those devices can optimize their wireless communication configurations (e.g., select the best beam for sending and receiving data). For example, a channel state information reference signal (CSI-RS) can be sent by one device and measured by another device in order to estimate the channel state and provide channel state information (CSI) feedback useful for optimizing wireless communication between the two devices.

[0096] However, due to the increasing complexity and capabilities of wireless communication devices, such as those capable of transmitting and receiving via multiple input and output antenna ports (e.g., implementing multiple-input multiple-output (MIMO) technology), conventional techniques may require a large amount of processing power and time, which reduces the performance of both the device and the entire wireless communication network. These technical problems are exacerbated by the typical use cases and environments of wireless communications, which are often dynamic. In other words, because the channel state changes frequently, the channel estimation and feedback procedures are often performed frequently, resulting in high power usage and large network overhead (e.g., in terms of time and frequency resources dedicated to channel estimation) for the wireless communication system. One way to alleviate such problems is to implement a machine learning model that can perform various functions related to channel state estimation and feedback more accurately and efficiently.

[0097] For example, conventional wireless communication systems may use, for example, time division multiplexing (TDM), code division multiplexing (CDM), and / or frequency division multiplexing (FDM) to allocate N packets per resource block. t Multiplex N resource elements t Such a system may generally achieve a resource block density between 0.5 and 1, such that a resource element is transmitted in every other or every single resource block. In contrast, a machine learning model deployed by a transmitting device (e.g., a base station) may be trained to perform N on L resource elements per resource block. t Multiplexing of ports, where L <N t , which therefore reduces the number of resource elements required for channel estimation, leaving more resource elements available for data transmission. In addition to reducing the number of resource elements required for channel estimation (which reduces power and enhances resource utilization), such models can also operate at reduced resource block densities (e.g., below 0.5) and can also implement non-uniform resource block patterns, which further improves the aforementioned benefits. On the receiving device (e.g., user equipment) side, a machine learning-based channel estimator can be trained to recover the complete channel over all resource blocks (e.g., N t ports), while receiving a reduced number of resources L. In various aspects, the CSI-RS multiplexing models at the transmitter side and the receiver side may be trained jointly or sequentially.

[0098] As another example, a conventional CSI reporting configuration may rely on a precoding matrix indicator (PMI) search algorithm and a PMI codebook to determine the best PMI codeword (e.g., CSI feedback) and report it to the network. However, machine learning-based models (such as encoders and decoders) may be trained to directly generate CSI feedback, which eliminates the need for a PMI search algorithm (replaced by the encoder) and a PMI codebook (replaced by the decoder). In various aspects described herein, a CSI encoder at the user equipment side may be trained to compress a channel estimate to a few bits that are then reported to a network entity (e.g., a base station), while a CSI decoder at the network entity side is trained to use the reported bits to recover the channel or precoding matrix.

[0099] Thus, in general, machine learning models can be trained to perform many functions related to channel estimation and feedback, and such models can often be more accurate, faster, more energy efficient, and better able to maintain performance in very dynamic radio environments. However, it is still important to monitor the performance of such machine learning models to ensure robust performance over time.

[0100] Figure 5 Depicted is an example 500 of monitoring machine learning model performance over time.

[0101] As depicted, during time interval 502, model output 504 closely tracks actual value 506 (eg, of channel estimate). The model may be deployed by user equipment, such as with respect to Figure 1 and Figure 3 5. During interval 502, there are two reporting events including two out-of-distribution (OOD) events in which the device deploying the model reports the performance of the model. Notably, OOD events depict situations where the model output (e.g., based on a threshold) deviates significantly from the actual value 506. In these situations, the trained model may be processing input data that is significantly different from the training data used to train the model, and therefore the model output becomes unreliable. Based on these OOD events, network entities (such as those regarding Figures 1 to 3 The described base station 102) can determine to transmit a model update.

[0102] The second time interval 504 illustrates various possible outcomes. Without the model update, the original model output 504 deviates significantly from the actual value 506. In contrast, the updated model output 508 also closely tracks the actual value 506. In addition, fallback methods such as conventional non-machine learning based methods are depicted to show that such methods may be better than a poorly performing machine learning model, but worse than a well performing machine learning model.

[0103] Figure 6An example of a model monitoring framework 600 is depicted.

[0104] In the framework 600, a network entity 602 (e.g., Figure 1 and Figure 3 Base station 102 as depicted and described herein or with respect to Figure 2 The depicted and described decomposed base station) and user equipment (e.g., Figure 1 and Figure 3 The depicted and described user equipment 104) communicates.

[0105] At step 606, network entity 602 transmits a model monitoring configuration to user equipment 604. The model monitoring configuration may define, for example, a number of modes to be employed by the user equipment and, in some cases, an indication of which mode to employ.

[0106] In some aspects, the model monitoring configuration may include an inference mode (or task mode) in which the user equipment 604 employs a machine learning model to perform a task and relies on the output of the model for the task. For example, the user equipment 604 may use a machine learning model for channel estimation and / or channel state information (CSI) feedback. In particular, the user equipment 604 may use a machine learning model to generate a channel estimate based on a reduced set of CSI reference signals (CSI-RS). In addition, the user equipment 604 may use a machine learning model to generate CSI feedback, which is trained to generate such feedback based on channel estimation (using the aforementioned machine learning model or other methods). Figure 7 Depicts an example of a model monitoring configuration for inference mode.

[0107] In some aspects, the model monitoring configuration may include a monitoring mode in which the user equipment 604 monitors the output of the machine learning model to look for model variance events, such as OOD events (e.g., as described above and with respect to Figure 5 As described above). For example, the user equipment 604 may monitor channel estimates and / or CSI feedback generated by the machine learning model for model variance events. The monitoring mode may be useful when the machine learning model is first deployed to determine its performance, such as for validating the model after training or updating. In addition, the monitoring mode may be useful when comparing the machine learning model performance to a baseline model (e.g., a conventional technique for performing a task) in order to determine which model (machine learning or baseline) to enable for the task. Figure 8 Depicts an example of a model monitoring configuration for monitoring mode.

[0108] In some aspects, the model monitoring configuration may include an inference and monitoring mode in which the user equipment 604 performs both inference and monitoring as described above. In particular, when performing inference and monitoring, it is possible for the user equipment to determine whether a given model output (e.g., inference) is changing (e.g., an OOD event) and then select a task output (e.g., for channel estimation and / or CSI feedback) based on the model variance determination. For example, in the case where the machine learning model output is changing, the user equipment 604 may use a fallback method (e.g., a baseline model) to obtain the task output, and in the case where the machine learning model is not changing, the user equipment 604 may use the machine learning model output to obtain the task output. Thus, in the inference and monitoring mode, the user equipment 604 may “trust, but verify” the machine learning model and observe performance changes over time (such as over Figure 5 In the case where the model deteriorates during the time interval 512), it is chosen to fall back to the baseline model. Fig. 9 Depicts an example of a model monitoring configuration for both inference and monitoring modes.

[0109] At step 608, the network entity 602 transmits a reference signal (eg, a measurement signal or resource) to the user equipment 604. For example, the reference signal may be a CSI-RS for the user equipment 604 to perform channel estimation and generate CSI feedback.

[0110] At step 610, the user equipment 604 performs a model variance determination (eg, an OOD event determination).For example, the user equipment 604 may operate in a monitoring or inference and monitoring mode, as described above.

[0111] In general, determining the model variance may be performed in a variety of ways. For example, to determine the model variance relative to the machine learning based CSF model, the model variance may be determined using statistics of the implicit output of the CSI encoder or statistics for the inner layers of the CSI encoder. As another example, another model may be trained to obtain the output of the CSI encoder and classify the output as changing or not changing. Note that, herein, the output from the machine learning based CSI encoder may be referred to as Type III CSI.

[0112] As another example, determining the model variance for the machine learning based CSF model may be based on comparing the output of the CSI encoder to a baseline model, such as a baseline codebook (e.g., Type I / II or (F)e Type IICSI). In such an example, a baseline codebook and a machine learning based CSI encoder are configured. In some aspects, the difference between the CSI encoder and the baseline model may be compared to a threshold, above which the model output is considered to be varied and below which the model output is considered to be normal.

[0113] As another example, determining the model variance for the machine-learning based CSF model may be based on an error metric associated with the machine-learning based CSF model (e.g., normalized mean square error (NMSE)). In some cases, the model-based error metric may be compared to the channel estimation error metric, and in the event that the difference is above a threshold, the CSF model output may be considered to be changing, and in the event that the difference is below the threshold, the CSF model output may be considered to be normal.

[0114] As another example, determining the model variance for a machine learning based CSF model can be based on PDSCH decoding performance such that the model output can be considered to vary when the PDSCH block error rate (BLER) is below a threshold (e.g., <<10%) or above a threshold (e.g., >>10%).

[0115] Similarly, determining the model variance relative to the machine learning based CSI-RS model can be performed in many different ways. For example, the model variance can be determined using statistics of the implicit output of the CSI-RS model or statistics for the inner layers of the CSI-RS model. As another example, another model can be trained to obtain the output of the CSI-RS model and classify the output as changing or not changing.

[0116] As another example, determining the model variance relative to a machine learning based CSI-RS model can be based on comparing the NMSE of the CSI-RS optimized by the CSI-RS model with a baseline model and a threshold, such that in the event that the NMSE is higher than the NMSE of the baseline model and lower than the threshold, the CSI-RS model output is considered to be changing.

[0117] As another example, determining the model variance relative to the machine learning based CSI-RS model can be based on a channel quality metric. In one aspect, in the event that the CQI / spectral efficiency result of the CSI-RS optimized by the CSI-RS model is worse than the baseline model and below a threshold, the CSI-RS model output is considered to be changed. As described above, the baseline model may include a baseline codebook (e.g., type I / II or (F)e type IICSI).

[0118] Based on the model variance determination, the user equipment 604 optionally proceeds to a fallback mode at step 612. For example, in the fallback mode, the user equipment 604 may implement a baseline model for tasks that it previously performed with the machine learning model, such as generating CSI feedback using a PMI search algorithm instead of a machine learning model.

[0119] Block 613 depicts different methods for making a model failure determination.

[0120] In a first example, at step 614, the user equipment 604 transmits a status report including a model variance determination to the network entity 602. Note that the status report may include a number of model variance determinations (e.g., a count of model variance determinations within a monitoring interval). Based on the status report 614, the network entity 602 performs a model failure determination at step 616. For example, the model failure determination may be based on the number of model variance events within the monitoring interval.

[0121] In a second example, at step 618, the user equipment 604 performs a model failure determination. For example, the model failure determination may be based on the number of model variance events within the monitoring interval, as described above. Then, at step 620, the user equipment 604 transmits a model failure indication to the network entity 602.

[0122] At step 622, the network entity 602 transmits a model failure information query to the user equipment 604, and the user equipment 604 responds with a model failure report at step 624. In some aspects, the model failure report may include, for example, information about the model that has failed (e.g., version, time the model has been deployed, etc.) and input values ​​and / or output values ​​associated with one or more model variance events that led to the model failure determination. Such values ​​may be used to update the machine learning model.

[0123] Finally, at step 626, the network entity 602 transmits a model update (e.g., for reconfiguring a machine learning model that has failed) to the user equipment 604. In the case of a model update, the user equipment 604 may update the model and improve task performance (e.g., as described in relation to Figure 5 508 in FIG.

[0124] It should be noted that Figure 6 This is merely one example of a method, and other methods including fewer, additional, or alternative steps are possible consistent with the present disclosure.

[0125] Figure 7 An example 700 of a model monitoring configuration 702 for inference mode is depicted.

[0126] In particular, the CSI reporting configuration 702 includes a machine learning model configuration 704, which can be used by the user equipment 704 to configure a machine learning model for some tasks such as channel estimation (e.g., via CSI reporting) or CSI feedback. For example, the machine learning model configuration 704 can configure a machine learning model for Type III CSI.

[0127] The CSI reporting configuration 702 also includes a baseline model configuration 706, which is used to configure a conventional model or technique for some tasks such as channel estimation or CSI feedback. For example, the baseline model configuration 704 may configure a baseline model for Type I, II, or (F)e Type IICSI.

[0128] The CSI reporting configuration 702 also includes an optional mode flag 708, which in this example is set to indicate an inference mode. Note that in other aspects, the CSI reporting configuration 702 may be specific to the inference mode, rather than having a flag (or other indicator) that may indicate one of many model monitoring modes. In such cases, the mode flag 708 may be omitted, and the user equipment may be configured using the reporting amount 709, such as described below with respect to Figure 8 Further description.

[0129] In this example, CSI reporting configuration 702 is provided to user equipment 704 (eg, by a network entity) and configures user equipment 704 for specific channel estimation and / or feedback-related tasks.

[0130] In this example, the user equipment 704 includes a machine learning task model 710 (e.g., a channel estimation or CSI feedback model), a baseline task model 712, a model monitoring setting 714 (which in this example is set for inference mode), a model variance detector 716, and an output selector 718.

[0131] Because in this example, the user equipment 704 is configured in the inference mode by the CSI reporting configuration 702, the machine learning task model 710 is used to generate a task output 722, which can be, for example, a channel estimate, CSI feedback, or other type of output.

[0132] Figure 8 An example 800 of a model monitoring configuration 802 for a monitoring mode is depicted.

[0133] In particular, the CSI reporting configuration 802 includes a machine learning model configuration 804, a baseline model configuration 806, and an optional mode flag 808, which in this example is set to indicate a monitoring mode.

[0134] In this example, CSI reporting configuration 802 is provided to user equipment 804 (eg, by a network entity) and configures user equipment 804 for specific channel estimation and / or feedback-related tasks.

[0135] As in Figure 7In the example of , here, the user equipment 804 includes a machine learning task model 810, a baseline task model 812, a model monitoring setting 814 (which in this example is set for monitoring mode), a model variance detector 816, and an output selector 818.

[0136] Because in this example, the user equipment 804 is configured in monitoring mode by the CSI reporting configuration 802, the baseline task model 812 is used to generate a task output 822, which may be, for example, a channel estimate, CSI feedback, or other type of output.

[0137] Additionally, the machine learning task model 810 generates outputs for monitoring model variance events (e.g., OOD events), which, when detected, can be used (e.g., in status reports such as Figure 6 614) transmits a model variance indication and / or can be used to determine model failure.

[0138] As mentioned above about Figure 6 There are many ways in which the model variance detector 816 can detect a model variance event (e.g., an OOD event), as described in step 610. In one example, the output of the machine learning task model 810 and the baseline task model 812 can be compared, and in the event that the output of the machine learning task model 810 is worse than the baseline task model 812 (e.g., subject to a threshold), the output of the machine learning task model 810 can be considered to be varied and reported in the model variance indication 820.

[0139] As described above, there are other methods that do not require the use of the output of the baseline task model 812 to determine the model variance event (hence the dashed arrow between the baseline task model 812 and the model variance detector 816). For example, implicit statistics and / or error metrics associated with the machine learning task model 810 may be considered, or a separate classification model (e.g., a neural network model) may be used to classify the output as changing or not changing, as described above with respect to Figure 6 Step 610 is discussed.

[0140] Note that in other aspects, the CSI reporting configuration 802 may be specific to a particular monitoring mode (e.g., there are dedicated CSI reports for monitoring, for inference, etc.), and in such cases, the CSI reporting configuration 802 need not include a flag (or other indicator) for indicating one of many model monitoring modes. Rather, in such cases, the CSI reporting configuration 802 may include a reporting quantity 809 (e.g., "reportQuantity" in the 3GPP standard) that causes the user equipment 804 to report a model variance event or to initiate a model failure indication. For example, one value of the reporting quantity 809 in the CSI reporting configuration 802 may cause the user equipment 804 to perform Figure 6 Step 614 in the CSI report configuration 802 and reporting the model variance event, and another reporting amount 809 in the CSI report configuration 802 may cause the user equipment 804 to perform Figure 6 Therefore, CSI reporting configuration 802 can configure the user equipment to report model variance events or model failures periodically, semi-persistently, or aperiodically when triggered by the use of reporting volume.

[0141] Fig. 9 An example 900 of a model monitoring configuration 902 for both inference and monitoring modes is depicted.

[0142] In particular, the CSI reporting configuration 902 includes a machine learning model configuration 904, a baseline model configuration 906, and an optional mode flag 908, which in this example is set to indicate an inference and monitoring mode.

[0143] In this example, CSI reporting configuration 902 is provided to user equipment 904 (eg, by a network entity) and configures user equipment 904 for specific channel estimation and / or feedback-related tasks.

[0144] As in Figure 7 and Figure 8 In the example of , here, the user equipment 904 includes a machine learning task model 910, a baseline task model 912, a model monitoring setting 914 (which in this example is set for inference and monitoring modes), a model variance detector 916, and an output selector 918.

[0145] Because in this example, the user equipment 904 is configured in the inference and monitoring mode by the CSI reporting configuration 902, both the machine learning task model 910 and the baseline task model 812 are used to generate preliminary task outputs.

[0146] The preliminary task output of the machine learning task model 910 is provided to a model variance detector 916, which determines whether the preliminary task output is variable (e.g., OOD output). Figure 6 As described in step 610 , there are many ways that model variance detector 916 can detect a model variance event (eg, an OOD event).

[0147] In one example, the output of the machine learning task model 910 and the baseline task model 912 may be compared, and in the event that the output of the machine learning task model 910 is worse than the baseline task model 912 (e.g., subject to a threshold), the output of the machine learning task model 910 may be considered varied and reported in the model variance indication 820.

[0148] Additionally, there are other methods that do not require the use of the output of the baseline task model 912 to determine the model variance event (hence the dashed arrow between the baseline task model 912 and the model variance detector 916). For example, implicit statistics and / or error metrics associated with the machine learning task model 910 may be considered, or a separate classification model (e.g., a neural network model) may be used to classify the output as changed or not changed, as described above with respect to Figure 6 Step 610 is discussed.

[0149] In the case where the output of the machine learning task model 910 is variable, the output selector 918 selects the baseline task model 912 preliminary output as the total task output 922, which can be, for example, a channel estimate, CSI feedback, or other type of output. In addition, the model variance detector 916 can generate and transmit a model variance indication 920 and / or can be used to determine model failure.

[0150] On the other hand, in the case where the model variance detector 916 determines that the preliminary task output of the machine learning task model 910 is not changing, the output selector 918 selects the preliminary output of the machine learning task model 910 as the total task output 922.

[0151] As above Figure 8 In some aspects, the CSI reporting configuration 902 may be specific to a particular monitoring mode (e.g., there may be dedicated CSI reports for monitoring and inference modes), rather than including the mode flag 908. In such cases, the CSI reporting configuration 902 may include a reporting amount 909 that causes the user equipment 904 to report a model variance event or initiate a model failure indication (e.g., via the model variance indication 920). For example, one value of the reporting amount 909 in the CSI reporting configuration 902 may cause the user equipment 904 to perform Figure 6Step 614 in the CSI reporting configuration 802 and reporting the model variance event, and another reporting amount 909 in the CSI reporting configuration 802 may cause the user equipment 804 to perform Figure 6 and reporting model failure events.

[0152] Figures 7 to 9 Some examples of using CSI reporting configuration to confirm model monitoring settings on user equipment are depicted. However, note that these are just some examples and the user equipment may also be configured in other ways.

[0153] For example, for any CSI report configured on the user equipment, the mode flag may be included in alternative signaling including radio resource control (RRC) signaling, medium access control element (MAC-CE), downlink control information (DCI), and others.

[0154] In one aspect, for periodic CSI, a mode flag may be included in the CSI reporting configuration (e.g., Figures 7 to 9 ) for initial configuration, and mode changes may be effected via RRC reconfiguration and / or MAC-CE commands, such as in Fig. 10A as depicted in the example of .

[0155] On the other hand, for semi-persistent CSI (e.g., on the physical uplink control channel (PUCCH)), the mode flag may be provided together with the activation MAC-CE, or the activation may be provided via a separate MAC-CE, such as Fig. 10B In such aspects, a list of semi-persistent (SP) CSIs may be included in the MAC-CE, and each semi-persistent (SP) CSI is provided with a specific mode flag, or a common flag set to all these triggered SP CSI reports.

[0156] For semi-persistent and aperiodic CSI on PUSCH, the CSI request DCI can be used to indicate the mode flag, such as Fig. 10C In such aspects, a list of semi-persistent / aperiodic CSIs is provided, each with a specific mode flag, or a common flag set to all these triggered SP / A CSI reports.

[0157] Alternatively, the mode change may be determined implicitly based on, for example, predefined rules. For example, for semi-persistent / aperiodic CSI, once the CSI is triggered, the mode is set to inference; otherwise, the mode is set to monitoring, such as Fig. 10D as depicted in the example of .

[0158] Model Variance Event Count

[0159] FIG. 11A to FIG. 11B Depicted is an example method for counting model variance events (eg, OOD events).

[0160] In particular, Fig.11A An example of performing one model variance event detection per model variance event report is depicted. In this example, the counting of model variance events (e.g., OOD events) is not based on (or limited to) any number of CSI-RS opportunities in the time domain. Instead, the user equipment uses all CSI-RS opportunities before the CSI reference resource to evaluate the model variance event, and reports the evaluation results per reporting opportunity. In other words, the user equipment is configured to perform model variance event detection using a specific CSI-RS opportunity. To this end, the user equipment is free to use all CSI-RS observations, or the most recent one.

[0161] Fig. 11B Another alternative is depicted in which the counting of model variance events is based on the number of CSI-RS opportunities in the time domain. For example, in some aspects, a monitoring opportunity (or interval or window) includes a configured number (e.g., N) of CSI-RS opportunities in the time domain. Further, in some aspects, one model variance event determination is made per monitoring opportunity (e.g., as depicted by example monitoring windows 1104A-1104C). Thus, in Fig. 11B In the example of , there is a defined association between a specific CSI-RS opportunity and a specific model variance determination.

[0162] In this example, the monitoring opportunities 1104A-1104C are non-overlapping, but in other examples, one or more monitoring opportunities may be overlapping. In addition, in this example, the monitoring opportunities 1104A-1104C are counted using a CSI-RS opportunity preceding a CSI reference resource, which in this example is a DCI trigger 1106 for aperiodic CSI reporting of a model variance event.

[0163] Model Variance Events and Model Failure Reports

[0164] Fig.12 Depicted are aspects related to reporting model variance events and model failures, such as when user equipment is operating in monitoring mode.

[0165] In some aspects, a device (e.g., user equipment) may report model variance events (e.g., OOD events) via PUCCH / PUSCH messaging as configured and / or triggered by the network. In one example, the user equipment may report one model variance event per report, which is generally consistent with the above description of Fig.11A and Fig. 11BThe model variance event counting methods discussed in both are consistent. Fig. 11B In the case of the method, the user equipment may be configured to report whether the model is changing (or not changing) in the latest measurement window. In another example, the user equipment may report the number (count) of model variance events in a certain number (e.g., M) of monitoring opportunities. The number M may be configured by the network.

[0166] Furthermore, in the event that the number of model variance events exceeds a threshold during a monitoring window (eg, during M monitoring occasions), a device (eg, user equipment) may initiate a model failure indication (eg, as in Figure 6 in step 620 of FIG. 11 ).

[0167] In some cases, after reporting a model failure, the user equipment may then refrain from reporting a second model failure report until a timer expires to allow the network time to respond (e.g., by Figure 6 In some aspects, the model failure indication may be transmitted via reserved uplink resources (e.g., in PUCCH, PUSCH, or sounding reference (SR) resources). In such cases, the PUCCH, PUSCH, or SR resource periodicity and slot offset may be configured via RRC (e.g., with a one-to-one mapping with the CSI reporting configuration). Thus, the exact time slot may be determined as the nearest PUCCH, PUSCH, or SR opportunity that meets the CSI processing timeline 1202, meaning that the actual time slot of the PUCCH / PUSCH / SR is the nearest PUCCH / PUSCH / SR at least N slots / symbols after the latest CSI-RS opportunity, where N is the CSI processing timeline for model variance or model failure event detection.

[0168] Model management for CSI-RS optimization

[0169] As discussed above, machine learning models can also be used for CSI-RS optimization to reduce CSI-RS overhead, and as described above, such models are monitored to ensure their continued performance and alternative approaches are selected if the models begin to deviate from actual data (such as described above with respect to Figure 5 described) is beneficial.

[0170] Fig.13 An example 1300 of using low-density CSI-RS for channel estimation based on a machine learning model is depicted.

[0171] Generally speaking, Fig.13, resource block 1306 includes a full density CSI-RS (e.g., 1304) indicated as a darker block, while resource block 1308 includes a low density CSI-RS (e.g., 1302) pattern. Resource block 1306 may represent a conventional CSI-RS transmission in which each resource block has a CSI-RS and within each resource block, there are 32 ports orthogonally multiplexed on 32 resource elements. In contrast, in resource block 1308, the resource block level density is reduced, and the multiplexing within each resource block is determined by a machine learning model. For example, the machine learning model may multiplex 32 ports on fewer resource elements, such as 16 (as depicted) or 8 resource elements.

[0172] The low-density CSI-RS (eg, 1302) may be measured (eg, by a user equipment) and provided as input to a machine learning-based channel estimation model, which uses the measurements to estimate the channel and generate CSI feedback.

[0173] To monitor the machine learning-based channel estimation model, a full set (e.g., full density set) of CSI-RS resource elements 1304 may be measured (e.g., by a user equipment) and used as real data for evaluating the performance of a machine learning-based model deployed by a network entity to generate low-density CSI-RS transmissions and a machine learning model deployed by a user equipment to perform channel estimation based on low-density CSI-RS transmissions. For example, CSI-RS resource element 1302 may be measured by a user equipment configured in an inference mode, and CSI-RS resource elements 1302 and 1304 may be measured by a user equipment configured in a monitoring mode or inference and monitoring mode.

[0174] In general, the management of the model monitoring mode for the machine learning based channel estimation model can be as described above with respect to Figures 6 to 12 Executed as described.

[0175] For example, the monitoring mode may be enabled via a mode flag in the CSI reporting configuration (e.g., Figures 7 to 9 In this example, a pair of CSI-RS resources may be configured in the CSI report configuration, including a first resource set (which is a target optimization resource set) (e.g., 1302) and a second resource set (which is a reference resource set for all ports and all bandwidths) (e.g., 1304). Therefore, in the case where the monitoring mode is inference, the first resource set (target optimization set) is measured and used for channel estimation and generating CSI feedback, and in the case where the monitoring mode is monitoring, both the first resource set and the second resource set are measured, and the second set is used as real data. As described below with respect to Fig.14As depicted and described, there may be a one-to-one mapping between resources in a first (target) set of resources and a second (reference or real data) set of resources.

[0176] As another example, monitoring mode may be enabled via a dedicated CSI reporting configuration for a CSI-RS machine learning model. Figure 8 and Fig. 9 As described, the reporting amount in the CSI reporting configuration can be used to indicate to the user equipment, for example, whether the user equipment should report a model variance event (e.g., an OOD event), or whether the user equipment should initiate a model failure indication. Therefore, the dedicated CSI reporting configuration can configure the user equipment to report a model variance event or a model failure periodically, semi-persistently, or aperiodically when triggered through the use of the reporting amount.

[0177] As yet another alternative, the monitoring mode may be enabled via CSI-RS resource set configuration or activation. In such an example, the second resource set is configured (in the case where the second set is periodic) or activated (in the case where the second set is semi-persistent) or triggered (in the case where the second set is aperiodic) as a real data set for the first resource set, which may have already been configured or activated. In some aspects, the target resource in the first resource set may be associated with the reference resource in the second resource set via RRC configuration of the target resource or reference resource, or included in MAC-CE activation of the target resource or reference resource, such as described below with respect to Fig.15 Describe in more detail.

[0178] Fig.14 An example 1400 of paired CSI-RS resources between a first (target) set of resources 1402 and a second (reference) set of resources 1404 is depicted.

[0179] Detecting model variance events for machine learning based channel estimation models can be similar to the above discussion on Figure 6Similar to that described in step 610 of . For example, one option is to determine the model variance based on the statistics of the implicit outputs of the layers of the machine learning-based channel estimation model. Another option is to use a separate module (e.g., a separate model) to determine the model variance event based on the output of the machine learning-based channel estimation model. Yet another option is to determine an error metric (e.g., normalized mean square error) for the machine learning-based channel estimation model (e.g., based on real data measurements) and compare the error metric with a threshold. Another option is to use the machine learning-based channel estimation model to determine a channel quality indicator (CQI) and / or spectral efficiency, and compare the channel quality indicator (CQI) and / or spectral efficiency with a baseline model (e.g., a baseline codebook, such as type I / II, (F)e type II), and see if the difference exceeds a threshold. Note that there are only a few examples, and other examples are possible.

[0180] Reports model variance events and model failures for machine learning-based channel estimation models (e.g., as in Figure 6 Steps 614 and 620 of the present invention may be described above with respect to Fig.11A and Fig. 11B described.

[0181] For example, Fig.15 Depicted is an example 1500 of triggering a model monitoring mode for a machine learning based channel estimation model.

[0182] exist Fig.15 15, MAC-CE 1502 activates a monitoring mode for interval 1510 (changing from an inference mode during interval 1508) in which the user equipment monitors both the target (e.g., low density) CSI-RS resource set and the reference (e.g., full density) CSI-RS resource set. Based on the monitoring, a model failure determination is made (e.g., as in Figure 6 ), and in the event that the model has failed, reporting the model failure determination at 1504 (e.g., as in Figure 6 in step 620 of FIG. 11 ).

[0183] like Fig.15, a target CSI-RS resource (e.g., 1504) may be associated with a reference CSI-RS resource (e.g., 1506). As described above, once the reference CSI-RS resource is activated by an activation command (e.g., MAC-CE 1502), the user equipment may begin monitoring based on the target resource and the reference resource. Note that the association between the target CSI-RS resource (e.g., 1504) and the reference CSI-RS resource (e.g., 1506) may be established via dedicated signaling (such as RRC signaling). Alternatively, the MAC-CE (e.g., 1502) may indicate to the user equipment which is the target CSI-RS resource and which is the reference CSI-RS resource.

[0184] Example Operation of User Equipment

[0185] Fig.16 A method for performing a task by a user device such as Figure 1 and Figure 3 An example of a method 1600 for performing wireless communications with a UE 104).

[0186] Method 1600 begins at step 1605 by receiving a reference signal from a network entity. In some cases, the operation of this step refers to the reference signal. Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0187] Then, method 1600 proceeds to step 1610, where the reference signal is processed using the machine learning model to generate a machine learning model output. In some cases, the operation of this step refers to the reference signal. Fig.18 Circuits for processing and / or codes for processing are described, or may be executed by circuits for processing and / or codes for processing.

[0188] Then, method 1600 proceeds to step 1615 to determine an action to be taken based on the machine learning model output and the model monitoring configuration. In some cases, the operation of this step refers to the operation of Fig.18 The described circuits for determining and / or codes for determining may be performed by, or may be executed by, the circuits for determining and / or codes for determining.

[0189] In some aspects, the method 1600 also includes receiving a model monitoring configuration from a network entity. In some cases, the operation of this step refers to the operation of Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0190] In some aspects, a model monitoring configuration defines a plurality of model monitoring states.

[0191] In some aspects, the method 1600 also includes receiving a channel state information reporting configuration from a network entity, the channel state information reporting configuration being configured to cause the user equipment to enable a selected model monitoring state from a plurality of model monitoring states. In some cases, the operation of this step refers to as described in reference Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0192] In some aspects, each respective one of the plurality of model monitoring states is associated with a respective mode flag, and the channel state information reporting configuration includes a mode flag configured to cause the user equipment to enable a selected one of the plurality of model monitoring states.

[0193] In some aspects, the method 1600 also includes receiving a mode flag from a network entity via RRC messaging, the mode flag being configured to cause the user equipment to enable another model monitoring state in the plurality of model monitoring states. In some cases, the operation of this step is referred to as referring to Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0194] In some aspects, the method 1600 also includes receiving a mode flag from a network entity via one or more MAC-CEs, the mode flag being configured to cause the user equipment to enable another model monitoring state in the plurality of model monitoring states. In some cases, the operation of this step is referred to as referring to Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0195] In some aspects, the method 1600 also includes receiving a mode flag from a network entity via a DCI, the mode flag being configured to cause the user equipment to enable another model monitoring state in the plurality of model monitoring states. In some cases, the operation of this step refers to as described in reference Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0196] In some aspects, the channel state information reporting configuration configures a first set of target CSI-RS resources and a second set of reference CSI-RS resources, each target CSI-RS resource in the first set of target CSI-RS resources is paired with a reference CSI-RS resource in the second set of reference CSI-RS resources, and the channel state information reporting configuration is configured to further enable the user equipment to determine whether to measure the second set of reference CSI-RS resources based on a selected model monitoring state.

[0197] In some aspects, the channel state information reporting configuration includes a mode flag configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

[0198] In some aspects, the method 1600 further includes receiving a resource indication from a network entity, the resource indication being configured to cause the user equipment to determine whether to measure a second set of reference CSI-RS resources based on the selected model monitoring state. In some cases, the operation of this step refers to as referring to Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0199] In some aspects, method 1600 further includes activating a model monitoring state in a plurality of model monitoring states based on a predefined rule. In some cases, the operation of this step refers to the operation of Fig.18 The described circuit for activating and / or code for activating, or may be executed by the circuit for activating and / or code for activating.

[0200] In some aspects, the actions include transmitting the machine learning model output to a network entity.

[0201] In some aspects, the actions include determining a model variance event based on the machine learning model output.

[0202] In some aspects, determining a model variance event includes at least one of: determining a statistic associated with a machine learning model output; processing the machine learning model output with a variance model configured to determine a model variance event; determining that an error metric associated with the machine learning model output is above a threshold; determining that the machine learning model output differs from a baseline model output by more than a threshold; or determining that an error metric associated with decoding performance at a user equipment is above a threshold.

[0203] In some aspects, the actions further include transmitting an indication of the model variance event to a network entity.

[0204] In some aspects, method 1600 also includes receiving an indication of a model failure event associated with the machine learning model from a network entity. In some cases, the operation of this step refers to as described in reference Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0205] In some aspects, an indication of a model variance event is included in a report associated with a single model variance event.

[0206] In some aspects, an indication of a model variance event is included in a report associated with a plurality of model variance events occurring within a predetermined number of model variance event monitoring occasions.

[0207] In some aspects, the actions also include: determining a model failure event based on the model variance event; and transmitting an indication of the model failure event associated with the machine learning model to a network entity.

[0208] In some aspects, determining a model failure event includes: incrementing a model variance event counter value; and determining that the model variance event counter value exceeds a model variance event count threshold during a monitoring interval.

[0209] In some aspects, the monitoring interval comprises a model variance event reporting interval.

[0210] In some aspects, the monitoring interval includes a predetermined number of channel state information reference signal opportunities.

[0211] In some aspects, the actions include: determining whether the machine learning model output indicates a model variance event; in the case where the machine learning model output indicates a model variance event, transmitting a baseline model output to a network entity based on a received reference signal; and in the case where the machine learning model output does not indicate a model variance event, transmitting the machine learning model output to the network entity.

[0212] In some aspects, the method 1600 further includes receiving a model failure information request from the network entity. In some cases, the operation of this step refers to the operation of Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0213] In some aspects, the method 1600 further comprises transmitting a model failure report to a network entity. In some cases, the operation of this step is referred to as Fig.18 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0214] In some aspects, method 1600 also includes receiving an updated machine learning model from the network entity. In some cases, the operation of this step refers to as referring to Fig.18 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0215] In some aspects, the machine learning model comprises a channel state feedback machine learning model, and the machine learning model output comprises channel state information feedback.

[0216] In some aspects, the machine learning model comprises a channel estimation machine learning model, and the machine learning model output comprises a channel estimate.

[0217] In one aspect, method 1600 or any aspect related thereto may be performed by an apparatus such as Fig.18 The method 1600 is performed by a communication device 1800 that includes various components operable, configured, or adapted to perform the method 1600. The communication device 1800 is described in further detail below.

[0218] It should be noted that Fig.16 This is merely one example of a method, and other methods including fewer, additional, or alternative steps are possible consistent with the present disclosure.

[0219] Example Operations of Network Entities

[0220] Fig.17 A method for transmitting data to a network entity such as Figure 1 and Figure 3 BS102 or such Figure 2 An example of a method 1700 for performing wireless communications using a decomposed base station) as discussed above.

[0221] Method 1700 begins at step 1705 by transmitting a model monitoring configuration to a user device. In some cases, the operation of this step is referred to as Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0222] Then, method 1700 proceeds to step 1710, transmitting a reference signal to the user equipment. In some cases, the operation of this step refers to the reference signal. Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0223] Then, method 1700 proceeds to step 1715, receiving one of a model variance indication or a model failure indication from the user equipment based on the reference signal. In some cases, the operation of this step refers to the reference signal. Fig.19 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0224] In some aspects, a model monitoring configuration defines a plurality of model monitoring states.

[0225] In some aspects, the method 1700 also includes transmitting a channel state information reporting configuration to the user equipment, the channel state information reporting configuration being configured to cause the user equipment to enable a selected model monitoring state from among a plurality of model monitoring states. In some cases, the operation of this step is referred to as referring to Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0226] In some aspects, each respective one of the plurality of model monitoring states is associated with a respective mode flag, and the channel state information reporting configuration includes a mode flag configured to cause the user equipment to enable a selected one of the plurality of model monitoring states.

[0227] In some aspects, the method 1700 also includes transmitting a mode flag to the user equipment via RRC messaging, the mode flag being configured to cause the user equipment to enable another model monitoring state in the plurality of model monitoring states. In some cases, the operation of this step is referred to as referring to Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0228] In some aspects, the method 1700 further includes transmitting a mode flag to the user equipment via one or more MAC-CEs, the mode flag being configured to cause the user equipment to enable another model monitoring state in the plurality of model monitoring states. In some cases, the operation of this step is referred to as referring to Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0229] In some aspects, the method 1700 further includes transmitting a mode flag to the user equipment via the DCI, the mode flag being configured to cause the user equipment to enable another model monitoring state in the plurality of model monitoring states. In some cases, the operation of this step is referred to as referring to Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0230] In some aspects, the channel state information reporting configuration configures a first set of target CSI-RS resources and a second set of reference CSI-RS resources, each target CSI-RS resource in the first set of target CSI-RS resources is paired with a reference CSI-RS resource in the second set of reference CSI-RS resources, and the channel state information reporting configuration is configured to enable the user equipment to determine whether to measure the second set of reference CSI-RS resources based on a selected model monitoring state.

[0231] In some aspects, the channel state information reporting configuration includes a mode flag configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

[0232] In some aspects, the method 1700 further includes transmitting a resource indication to the user equipment, the resource indication being configured to cause the user equipment to determine whether to measure a second set of reference CSI-RS resources based on the selected model monitoring state. In some cases, the operation of this step refers to as referring to Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0233] In some aspects, receiving one of a model variance indication or a model failure indication from the user equipment comprises receiving a model variance indication.

[0234] In some aspects, method 1700 also includes transmitting an indication of a model failure event to the user equipment based at least in part on receiving the model variance indication. In some cases, the operation of this step refers to as described in reference to Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0235] In some aspects, an indication of a model variance event is included in a report associated with a single model variance event.

[0236] In some aspects, an indication of a model variance event is included in a report associated with a plurality of model variance events occurring within a predetermined number of model variance event monitoring occasions.

[0237] In some aspects, receiving one of a model variance indication or a model failure indication from the user equipment comprises receiving a model failure indication.

[0238] In some aspects, method 1700 also includes transmitting a model failure information request to the user equipment. In some cases, the operation of this step is referred to as reference Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0239] In some aspects, method 1700 also includes receiving a model failure report from the user equipment. In some cases, the operation of this step is referred to as reference Fig.19 Circuits for receiving and / or code for receiving are described, or may be performed by circuits for receiving and / or code for receiving.

[0240] In some aspects, method 1700 also includes transmitting the updated machine learning model to the user equipment. In some cases, the operation of this step refers to as referring to Fig.19 The described circuits for transmitting and / or codes for transmitting, or may be performed by the circuits for transmitting and / or codes for transmitting.

[0241] In some aspects, the model monitoring configuration is associated with a channel state feedback machine learning model.

[0242] In some aspects, the model monitoring configuration is associated with a channel estimation machine learning model.

[0243] In one aspect, method 1700 or any aspect related thereto may be performed by an apparatus such as Fig.19 The method 1700 is performed by a communication device 1900 that includes various components operable, configured, or adapted to perform the method 1700. The communication device 1900 is described in further detail below.

[0244] It should be noted that Fig.17 This is merely one example of a method, and other methods including fewer, additional, or alternative steps are possible consistent with the present disclosure.

[0245] Example Communication Device

[0246] Fig.18 Depicted are aspects of an example communication device 1800. In some aspects, the communication device 1800 is a user equipment, such as described above with respect to Figure 1 and Figure 3 UE 104 is described.

[0247] The communication device 1800 includes a processing system 1805 coupled to a transceiver 1875 (e.g., a transmitter and / or a receiver). The transceiver 1875 is configured to transmit and receive signals for the communication device 1800, such as various signals as described herein, via an antenna 1880. The processing system 1805 may be configured to perform processing functions for the communication device 1800, including processing signals received by the communication device 1800 and / or to be transmitted by the communication device.

[0248] The processing system 1805 includes one or more processors 1810. In various aspects, the one or more processors 1810 may represent Figure 3One or more of the receive processor 358, transmit processor 364, TX MIMO processor 366, and / or controller / processor 380 described herein. The one or more processors 1810 are coupled to the computer readable medium / memory 1840 via a bus 1870. In some aspects, the computer readable medium / memory 1840 is configured to store instructions (e.g., computer executable code) that, when executed by the one or more processors 1810, cause the one or more processors 1810 to perform operations related to the processing of the computer readable medium / memory 1840. Fig.16 The described method 1600 or any aspect related thereto. Note that reference to a processor performing a function of the communication device 1800 may include one or more processors 1810 performing that function of the communication device 1800 .

[0249] In the depicted example, computer readable medium / memory 1840 stores code (e.g., executable instructions), such as code for receiving 1845, code for processing 1850, code for determining 1855, code for activating 1860, and code for transmitting 1865. Processing of code for receiving 1845, code for processing 1850, code for determining 1855, code for activating 1860, and code for transmitting 1865 may cause communication device 1800 to perform operations related to Fig.16 The described method 1600 or any aspect related thereto.

[0250] The one or more processors 1810 include circuits configured to implement (e.g., execute) code stored in the computer-readable medium / memory 1840, including circuits such as circuits for receiving 1815, circuits for processing 1820, circuits for determining 1825, circuits for activating 1830, and circuits for transmitting 1835. Processing with the circuits for receiving 1815, processing 1820, determining 1825, activating 1830, and transmitting 1835 may cause the communication device 1800 to perform operations related to the communication. Fig.16 The described method 1600 or any aspect related thereto.

[0251] The various components of the communication device 1800 may provide for performing Fig.16 The components of the described method 1600 or any aspect related thereto. For example, the components for sending, transmitting, or outputting for sending may include Figure 3 The transceiver 354 and / or antenna 352 of the exemplary UE 104, and / or Fig.18 The transceiver 1875 and antenna 1880 of the communication device 1800 in FIG. 1 may include Figure 3The transceiver 354 and / or antenna 352 of the exemplary UE 104, and / or Fig.18 The transceiver 1875 and antenna 1880 of the communication device 1800.

[0252] Fig.19 Various aspects of an example communications device 1900 are depicted. In some aspects, communications device 1900 is a network entity such as Figure 1 and Figure 3 BS102 or such Figure 2 The decomposed base station in question.

[0253] The communication device 1900 includes a processing system 1905 coupled to a transceiver 1945 (e.g., a transmitter and / or a receiver) and / or a network interface 1955. The transceiver 1945 is configured to send and receive signals for the communication device 1900 via an antenna 1950, such as various signals as described herein. The network interface 1955 is configured to receive signals for the communication device 1900 via a communication link (such as, for example, a signal processor as described herein). Figure 2 The processing system 1905 may be configured to perform the processing functions of the communication device 1900, including processing signals received by the communication device 1900 and / or to be sent by the communication device 1900.

[0254] The processing system 1905 includes one or more processors 1910. In various aspects, the one or more processors 1910 may represent Figure 3 One or more of the receive processor 338, transmit processor 320, TX MIMO processor 330, and / or controller / processor 340 described herein. The one or more processors 1910 are coupled to the computer readable medium / memory 1925 via the bus 1940. In some aspects, the computer readable medium / memory 1925 is configured to store instructions (e.g., computer executable code) that, when executed by the one or more processors 1910, cause the one or more processors 1910 to perform operations related to the processing of the computer readable medium / memory 1925. Fig.17 The described method 1700 or any aspect related thereto. Note that reference to a processor of the communication device 1900 that performs a function may include one or more processors 1910 of the communication device 1900 that perform that function.

[0255] In the depicted example, computer readable medium / memory 1925 stores code (e.g., executable instructions), such as code for transmitting 1930 and code for receiving 1935. Processing of code for transmitting 1930 and code for receiving 1935 may cause communication device 1900 to perform operations related to Fig.17 The described method 1700 or any aspect related thereto.

[0256] The one or more processors 1910 include circuits configured to implement (e.g., execute) code stored in the computer-readable medium / memory 1925, including circuits such as circuits for transmitting 1915 and circuits for receiving 1920. Processing with circuits for transmitting 1915 and circuits for receiving 1920 may cause the communication device 1900 to perform as described above. Fig.17 The described method 1700 or any aspect related thereto.

[0257] The various components of the communication device 1900 may be provided for performing the following operations: Fig.17 The components of the described method 1700 or any aspect related thereto. The components for sending, transmitting or outputting for sending may include Figure 3 The transceiver 332 and / or antenna 334 of the exemplary BS 102, and / or Fig.19 The transceiver 1945 and antenna 1950 of the communication device 1900 in FIG. 1 may include Figure 3 The transceiver 332 and / or antenna 334 of the exemplary BS 102, and / or Fig.19 The transceiver 1945 and antenna 1950 of the communication device 1900.

[0258] Sample Clauses

[0259] Specific implementation examples are described in the following numbered clauses:

[0260] Clause 1: A method for wireless communication by a user equipment, comprising: receiving a reference signal from a network entity; processing the reference signal with a machine learning model to generate a machine learning model output; and determining an action to be taken based on the machine learning model output and a model monitoring configuration.

[0261] Clause 2: The method of clause 1, further comprising receiving the model monitoring configuration from the network entity.

[0262] Clause 3: A method according to any one of clauses 1 and 2, wherein the model monitoring configuration defines a plurality of model monitoring states.

[0263] Clause 4: The method of clause 3, further comprising receiving a channel state information reporting configuration from the network entity, the channel state information reporting configuration configured to cause the user equipment to enable a selected model monitoring state of the plurality of model monitoring states.

[0264] Clause 5: A method according to clause 4, wherein: each corresponding model monitoring state among the multiple model monitoring states is associated with a corresponding mode flag, and the channel state information reporting configuration includes a mode flag, and the mode flag is configured to enable the user equipment to enable the selected model monitoring state among the multiple model monitoring states.

[0265] Clause 6: The method of clause 4, further comprising receiving a mode flag from the network entity via RRC messaging, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

[0266] Clause 7: The method of clause 4, further comprising receiving a mode flag from the network entity via one or more MAC-CEs, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

[0267] Clause 8: The method of clause 4, further comprising receiving a mode flag from the network entity via a DCI, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

[0268] Clause 9: A method according to Clause 4, wherein: the channel state information reporting configuration configures a first group of target CSI-RS resources and a second group of reference CSI-RS resources, each target CSI-RS resource in the first group of target CSI-RS resources is paired with a reference CSI-RS resource in the second group of reference CSI-RS resources, and the channel state information reporting configuration is configured to further enable the user equipment to determine whether to measure the second group of reference CSI-RS resources based on the selected model monitoring state.

[0269] Clause 10: The method of clause 9, wherein the channel state information reporting configuration comprises a mode flag configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

[0270] Clause 11: The method of clause 9, further comprising receiving a resource indication from the network entity, the resource indication configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

[0271] Clause 12: The method of clause 3, further comprising activating a model monitoring state of the plurality of model monitoring states based on a predefined rule.

[0272] Clause 13: A method as described in any of clauses 1 to 12, wherein the action includes transmitting the machine learning model output to the network entity.

[0273] Clause 14: A method according to any one of clauses 1 to 13, wherein the action includes determining a model variance event based on the machine learning model output.

[0274] Clause 15: A method according to clause 14, wherein determining the model variance event comprises at least one of the following: determining a statistical value associated with the machine learning model output; processing the machine learning model output with a variance model configured to determine the model variance event; determining that an error metric associated with the machine learning model output is above a threshold; determining that the machine learning model output differs from a baseline model output by more than a threshold; or determining that an error metric associated with decoding performance at the user equipment is above a threshold.

[0275] Clause 16: The method of clause 14, wherein the actions further comprise transmitting an indication of the model variance event to the network entity.

[0276] Clause 17: The method of clause 16, further comprising receiving, from the network entity, an indication of a model failure event associated with the machine learning model.

[0277] Clause 18: The method of clause 16, wherein the indication of the model variance event is included in a report associated with a single model variance event.

[0278] Clause 19: The method of clause 16, wherein the indication of the model variance event is included in a report associated with a plurality of model variance events occurring within a predetermined number of model variance event monitoring occasions.

[0279] Clause 20: A method according to clause 14, wherein the action further includes: determining a model failure event based on the model variance event; and transmitting an indication of the model failure event associated with the machine learning model to the network entity.

[0280] Clause 21: The method of clause 20, wherein determining the model failure event comprises: incrementing a model variance event counter value; and determining that the model variance event counter value exceeds a model variance event count threshold during a monitoring interval.

[0281] Clause 22: The method of clause 21, wherein the monitoring interval comprises a model variance event reporting interval.

[0282] Clause 23: A method as described in clause 21, wherein the monitoring interval comprises a predetermined number of channel state information reference signal opportunities.

[0283] Clause 24: A method according to any one of clauses 1 to 23, wherein the action comprises: determining whether the machine learning model output indicates a model variance event; in the case where the machine learning model output indicates a model variance event, transmitting a baseline model output to the network entity based on the received reference signal; and in the case where the machine learning model output does not indicate a model variance event, transmitting the machine learning model output to the network entity.

[0284] Clause 25: The method of any one of clauses 1 to 24, further comprising receiving a model failure information request from the network entity.

[0285] Clause 26: The method of clause 25, further comprising transmitting a model failure report to the network entity.

[0286] Clause 27: The method of clause 26, further comprising receiving an updated machine learning model from the network entity.

[0287] Clause 28: A method according to any one of clauses 1 to 27, wherein: the machine learning model includes a channel state feedback machine learning model, and the machine learning model output includes channel state information feedback.

[0288] Clause 29: A method according to any one of clauses 1 to 28, wherein: the machine learning model includes a channel estimation machine learning model, and the machine learning model output includes a channel estimate.

[0289] Clause 30: A method of wireless communication by a network entity, comprising: transmitting a model monitoring configuration to a user equipment; transmitting a reference signal to the user equipment; and receiving one of a model variance indication or a model failure indication from the user equipment based on the reference signal.

[0290] Clause 31: The method of clause 30, wherein the model monitoring configuration defines a plurality of model monitoring states.

[0291] Clause 32: The method of clause 31, further comprising transmitting a channel state information reporting configuration to the user equipment, the channel state information reporting configuration being configured to cause the user equipment to enable a selected model monitoring state of the plurality of model monitoring states.

[0292] Clause 33: A method according to clause 32, wherein: each corresponding model monitoring state among the multiple model monitoring states is associated with a corresponding mode flag, and the channel state information reporting configuration includes a mode flag, and the mode flag is configured to enable the user equipment to enable the selected model monitoring state among the multiple model monitoring states.

[0293] Clause 34: The method of clause 32, further comprising transmitting a mode flag to the user equipment via RRC messaging, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

[0294] Clause 35: The method of clause 32, further comprising transmitting a mode flag to the user equipment via one or more MAC-CEs, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

[0295] Clause 36: The method of clause 32, further comprising transmitting a mode flag to the user equipment via a DCI, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

[0296] Clause 37: A method according to clause 32, wherein: the channel state information reporting configuration configures a first group of target CSI-RS resources and a second group of reference CSI-RS resources, each target CSI-RS resource in the first group of target CSI-RS resources is paired with a reference CSI-RS resource in the second group of reference CSI-RS resources, and the channel state information reporting configuration is configured so that the user equipment determines whether to measure the second group of reference CSI-RS resources based on the selected model monitoring state.

[0297] Clause 38: The method of clause 37, wherein the channel state information reporting configuration comprises a mode flag configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

[0298] Clause 39: The method of clause 32, further comprising transmitting a resource indication to the user equipment, the resource indication configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

[0299] Clause 40: The method of any of clauses 30 to 39, wherein receiving one of the model variance indication or the model failure indication from the user equipment comprises receiving the model variance indication.

[0300] Clause 41: The method of clause 40, further comprising transmitting an indication of a model failure event to the user equipment based at least in part on receiving the model variance indication.

[0301] Clause 42: The method of clause 41, wherein the indication of the model variance event is included in a report associated with a single model variance event.

[0302] Clause 43: The method of clause 41, wherein the indication of the model variance event is included in a report associated with a plurality of model variance events occurring within a predetermined number of model variance event monitoring occasions.

[0303] Clause 44: The method of any of clauses 30 to 43, wherein receiving one of the model variance indication or the model failure indication from the user equipment comprises receiving the model failure indication.

[0304] Clause 45: The method of any of clauses 30 to 44, further comprising transmitting a model failure information request to the user equipment.

[0305] Clause 46: The method of clause 45, further comprising receiving a model failure report from the user equipment.

[0306] Clause 47: The method according to clause 46 also includes transmitting an updated machine learning model to the user equipment.

[0307] Clause 48: A method according to any one of clauses 30 to 47, wherein the model monitoring configuration is associated with a channel state feedback machine learning model.

[0308] Clause 49: A method according to any one of clauses 30 to 48, wherein the model monitoring configuration is associated with a channel estimation machine learning model.

[0309] Clause 50: An apparatus comprising: a memory including executable instructions; and a processor configured to execute the executable instructions and cause the apparatus to perform the method of any one of clauses 1 to 49.

[0310] Clause 51: An apparatus comprising means for performing the method according to any one of clauses 1 to 49.

[0311] Clause 52: A non-transitory computer readable medium comprising executable instructions which, when executed by a processor of an apparatus, cause the apparatus to perform the method of any one of clauses 1 to 49.

[0312] Clause 53: A computer program product embodied on a computer readable storage medium, the computer program product comprising code for performing the method according to any of clauses 1 to 49.

[0313] Additional considerations

[0314] The foregoing description is provided to enable any person skilled in the art to practice the various aspects described herein. The examples discussed herein do not limit the scope, applicability, or aspects set forth in the claims. Various modifications to these aspects will be apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects. For example, without departing from the scope of the present disclosure, the functions and arrangements of the elements discussed may be changed. Various examples may omit, replace, or add various procedures or components as appropriate. For example, the described methods may be performed in a different order than the described order, and various actions may be added, omitted, or combined. In addition, the features described with respect to some examples may be combined in some other examples. For example, any number of aspects set forth herein may be used to implement a device or practice method. In addition, the scope of the present disclosure is intended to cover such devices or methods practiced using other structures, functions, or structures and functions other than the various aspects of the present disclosure set forth herein or different from the various aspects of the present disclosure set forth herein. It should be understood that any aspect of the disclosure disclosed herein may be embodied by one or more elements of the claims.

[0315] The various illustrative logical blocks, modules, and circuits described in conjunction with the present disclosure may be implemented or performed using a general purpose processor, a digital signal processor (DSP), an ASIC, a field programmable gate array (FPGA) or other programmable logic device (PLD), discrete gate or transistor logic, discrete hardware components, or any combination thereof, designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in an alternative, the processor may be any commercially available processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, a system on a chip (SoC), or any other such configuration.

[0316] As used herein, a phrase referring to "at least one of" a list of items refers to any combination of those items (including single members). For example, "at least one of a, b, or c" is intended to cover a, b, c, ab, ac, bc, and abc, as well as any combination with multiple identical elements (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or any other ordering of a, b, and c).

[0317] As used herein, the term "determining" encompasses a wide variety of actions. For example, "determining" may include calculating, computing, processing, deriving, investigating, searching (e.g., searching in a table, a database, or another data structure), ascertaining, and the like. Additionally, "determining" may include receiving (e.g., receiving information), accessing (e.g., accessing data in a memory), and the like. Additionally, "determining" may include resolving, selecting, choosing, establishing, and the like.

[0318] The method disclosed herein includes one or more actions for implementing the method. The method actions may be interchangeable with each other without departing from the scope of the claims. In other words, unless a specific order of actions is specified, the order and / or use of specific actions may be modified without departing from the scope of the claims. In addition, the various operations of the method described above may be performed by any suitable component capable of performing the corresponding function. The component may include various hardware and / or software components and / or modules, including but not limited to circuits, application specific integrated circuits (ASICs) or processors.

[0319] The following claims are not intended to be limited to the various aspects shown herein, but should be given the full scope consistent with the language of the claims. Within the claims, unless specifically stated, reference to an element in the singular form is not intended to mean "one and only one", but "one or more". Unless otherwise specifically stated, the term "some" refers to one or more. Any claim element is not interpreted according to the provisions of 35 U.SC § 112 (f) unless the element is explicitly stated using the phrase "parts for...". All structural and functional equivalents of the elements of the various aspects described throughout the present disclosure that are known or will be known later to a person of ordinary skill in the art are expressly incorporated herein by reference and are intended to be covered by the claims. In addition, regardless of whether such disclosure is explicitly stated in the claims, anything disclosed herein is not intended to be disclosed to the public.

Claims

1. A method for wireless communication by a user equipment, comprising: receiving a reference signal from a network entity; processing the reference signal with a machine learning model to generate a machine learning model output; as well as Determine actions to take based on the machine learning model output and the model monitoring configuration. 2 . The method of claim 1 , further comprising receiving the model monitoring configuration from the network entity. The method of claim 1 , wherein the model monitoring configuration defines a plurality of model monitoring states.

4. The method according to claim 3, further comprising receiving a channel state information reporting configuration from the network entity, the channel state information reporting configuration being configured to cause the user equipment to enable a selected model monitoring state among the multiple model monitoring states.

5. The method according to claim 4, wherein: Each respective model monitoring state of the plurality of model monitoring states is associated with a respective mode flag, and The channel state information reporting configuration includes a mode flag configured to cause the user equipment to enable the selected model monitoring state among the plurality of model monitoring states.

6. The method of claim 1, wherein the action comprises transmitting the machine learning model output to the network entity.

7. The method of claim 1, wherein the action comprises determining a model variance event based on the machine learning model output.

8. The method of claim 7, wherein determining the model variance event comprises at least one of: determining a statistical value associated with the machine learning model output; processing the machine learning model output with a variance model configured to determine variance events for the model; determining that an error metric associated with the machine learning model output is above a threshold; Determining that the machine learning model output differs from the baseline model output by more than a threshold; or An error metric associated with decoding performance at the user equipment is determined to be above a threshold.

9. The method of claim 7, wherein the actions further comprise transmitting an indication of the model variance event to the network entity.

10. The method according to claim 9, further comprising: An indication of a model failure event associated with the machine learning model is received from the network entity.

11. The method of claim 9, wherein the indication of the model variance event is included in a report associated with a single model variance event. 12 . The method of claim 9 , wherein the indication of the model variance event is included in a report associated with a plurality of model variance events occurring within a predetermined number of model variance event monitoring occasions.

13. The method of claim 7, wherein the actions further comprise: determining a model failure event based on the model variance event; as well as Transmitting an indication of the model failure event associated with the machine learning model to the network entity.

14. The method of claim 13, wherein determining the model failure event comprises: Increment the model variance event counter value; as well as It is determined that the model variance event counter value exceeds a model variance event count threshold during a monitoring interval.

15. The method of claim 14, wherein the monitoring interval comprises a model variance event reporting interval.

16. The method of claim 14, wherein the monitoring interval comprises a predetermined number of channel state information reference signal opportunities.

17. The method of claim 1, wherein the actions comprise: determining whether the machine learning model output indicates a model variance event; In a situation where the machine learning model output indicates a model variance event, transmitting a baseline model output to the network entity based on the received reference signal; as well as In the event that the machine learning model output does not indicate a model variance event, the machine learning model output is transmitted to the network entity.

18. The method of claim 4, further comprising receiving a mode flag from the network entity via radio resource control (RRC) messaging, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

19. The method of claim 4, further comprising receiving a mode flag from the network entity via one or more medium access control elements (MAC-CEs), the mode flag being configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

20. The method of claim 4, further comprising receiving a mode flag from the network entity via downlink control information (DCI), the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

21. The method of claim 3, further comprising activating a model monitoring state of the plurality of model monitoring states based on a predefined rule.

22. The method of claim 4, wherein: The channel state information reporting configuration configures a first set of target channel state information reference signal (CSI-RS) resources and a second set of reference CSI-RS resources, Each target CSI-RS resource in the first set of target CSI-RS resources is paired with a reference CSI-RS resource in the second set of reference CSI-RS resources, and The channel state information reporting configuration is configured to further cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

23. The method of claim 22, wherein the channel state information reporting configuration comprises a mode flag configured to enable the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

24. The method of claim 22, further comprising receiving a resource indication from the network entity, the resource indication configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

25. The method of claim 1, further comprising receiving a model failure information request from the network entity.

26. The method of claim 25, further comprising transmitting a model failure report to the network entity.

27. The method of claim 26, further comprising receiving an updated machine learning model from the network entity.

28. The method of claim 1, wherein: The machine learning model includes a channel state feedback machine learning model, and The machine learning model output includes channel state information feedback.

29. The method of claim 1, wherein: The machine learning model includes a channel estimation machine learning model, and The machine learning model output includes channel estimates.

30. A method for wireless communication by a network entity, comprising: transmitting a model monitoring configuration to a user equipment; transmitting a reference signal to the user equipment; as well as Based on the reference signal, one of a model variance indication or a model failure indication is received from the user equipment.

31. The method of claim 30, wherein the model monitoring configuration defines a plurality of model monitoring states.

32. The method of claim 31 , further comprising transmitting a channel state information reporting configuration to the user equipment, the channel state information reporting configuration being configured to cause the user equipment to enable a selected model monitoring state of the plurality of model monitoring states.

33. The method of claim 32, wherein: Each respective model monitoring state of the plurality of model monitoring states is associated with a respective mode flag, and The channel state information reporting configuration includes a mode flag configured to cause the user equipment to enable the selected model monitoring state among the plurality of model monitoring states.

34. The method of claim 30, wherein receiving one of the model variance indication or the model failure indication from the user equipment comprises receiving the model variance indication.

35. The method of claim 34, further comprising transmitting an indication of a model failure event to the user equipment based at least in part on receiving the model variance indication.

36. The method of claim 35, wherein the indication of the model variance event is included in a report associated with a single model variance event.

37. The method of claim 35, wherein the indication of the model variance event is included in a report associated with a plurality of model variance events occurring within a predetermined number of model variance event monitoring occasions.

38. The method of claim 30, wherein receiving one of the model variance indication or the model failure indication from the user equipment comprises receiving the model failure indication.

39. The method of claim 32, further comprising transmitting a mode flag to the user equipment via radio resource control (RRC) messaging, the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

40. The method of claim 32, further comprising transmitting a mode flag to the user equipment via one or more medium access control elements (MAC-CEs), the mode flag being configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

41. The method of claim 32, further comprising transmitting a mode flag to the user equipment via downlink control information (DCI), the mode flag configured to cause the user equipment to enable another model monitoring state of the plurality of model monitoring states.

42. The method of claim 32, wherein: The channel state information reporting configuration configures a first set of target channel state information reference signal (CSI-RS) resources and a second set of reference CSI-RS resources, Each target CSI-RS resource in the first set of target CSI-RS resources is paired with a reference CSI-RS resource in the second set of reference CSI-RS resources, and The channel state information reporting configuration is configured to enable the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

43. The method of claim 42, wherein the channel state information reporting configuration comprises a mode flag configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

44. The method of claim 32, further comprising transmitting a resource indication to the user equipment, the resource indication configured to cause the user equipment to determine whether to measure the second set of reference CSI-RS resources based on the selected model monitoring state.

45. The method of claim 30, further comprising transmitting a model failure information request to the user equipment.

46. ​​The method of claim 45, further comprising receiving a model failure report from the user equipment.

47. The method of claim 46, further comprising transmitting an updated machine learning model to the user equipment.

48. The method of claim 30, wherein the model monitoring configuration is associated with a channel state feedback machine learning model.

49. The method of claim 30, wherein the model monitoring configuration is associated with a channel estimation machine learning model.

50. An apparatus comprising: a memory including executable instructions; and a processor configured to execute the executable instructions and cause the apparatus to perform the method according to any one of claims 1 to 49.

51. An apparatus comprising means for performing the method according to any one of claims 1 to 49.

52. A non-transitory computer readable medium comprising executable instructions which, when executed by a processor of a device, cause the device to perform the method of any one of claims 1 to 49.

53. A computer program product embodied on a computer readable storage medium, the computer program product comprising code for performing the method according to any one of claims 1 to 49.

Images