Machine learning (ML) model monitoring using ml-based monitoring process
Patent Information
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- QUALCOMM INC
- Filing Date
- 2023-08-15
- Publication Date
- 2026-06-24
AI Technical Summary
Existing machine learning (ML) model monitoring techniques struggle to effectively detect concept drift in wireless communication systems, particularly when there is a mismatch between training data and inference data, leading to decreased model accuracy.
The use of an ML-based monitoring process that employs a second ML model to monitor a first ML model associated with wireless communication. This monitoring process relies on distance measures between the training dataset and inference data instances to determine the suitability of the ML model and decide whether to switch, fine-tune, or revert to a non-ML model.
This approach enhances the ability to detect concept drift and maintain the accuracy of ML models in wireless communication systems, allowing for real-time adjustments to ensure optimal performance.
Smart Images

Figure CN2023113038_20022025_PF_FP_ABST
Abstract
Description
MACHINE LEARNING (ML) MODEL MONITORING USING ML-BASED MONITORING PROCESS
[0001] FIELD OF THE DISCLOSURE
[0002] Aspects of the present disclosure generally relate to machine learning (ML) model monitoring using an ML-based monitoring process.BACKGROUND
[0003] Wireless communications systems are widely deployed to provide various telecommunications services such as telephony, video, data, messaging, and broadcasts. Typical wireless communications systems may employ multiple-access technologies capable of supporting communications with multiple users by sharing available system resources (e.g., bandwidth, transmit power, and / or the like) . Examples of such multiple-access technologies include code division multiple access (CDMA) systems, time division multiple access (TDMA) systems, frequency-division multiple access (FDMA) systems, orthogonal frequency-division multiple access (OFDMA) systems, single-carrier frequency-division multiple access (SC-FDMA) systems, time division synchronous code division multiple access (TD-SCDMA) systems, and long term evolution (LTE) . LTE / LTE-Advanced is a set of enhancements to the universal mobile telecommunications system (UMTS) mobile standard promulgated by the Third Generation Partnership Project (3GPP) . Narrowband (NB) -Internet of things (IoT) and enhanced machine-type communications (eMTC) are a set of enhancements to LTE for machine type communications.
[0004] A wireless communications network may include a number of base stations (BSs) that can support communications for a number of user equipment (UEs) . A user equipment (UE) may communicate with a base station (BS) via the downlink and uplink. The downlink (or forward link) refers to the communications link from the BS to the UE, and the uplink (or reverse link) refers to the communications link from the UE to the BS. As will be described in more detail, a BS may be referred to as a Node B, an evolved Node B (eNB) , a gNB, an access point (AP) , a radio head, a transmit and receive point (TRP) , a new radio (NR) BS, a 5G Node B, and / or the like.
[0005] The above multiple access technologies have been adopted in various telecommunications standards to provide a common protocol that enables different user equipment to communicate on a municipal, national, regional, and even global level. New Radio (NR) , which may also be referred to as 5G, is a set of enhancements to the LTE mobile standard promulgated by the Third Generation Partnership Project (3GPP) . NR is designed to better support mobile broadband Internet access by improving spectral efficiency, lowering costs, improving services, making use of new spectrum, and better integrating with other open standards using orthogonal frequency division multiplexing (OFDM) with a cyclic prefix (CP) (CP-OFDM) on the downlink (DL) , using CP-OFDM and / or SC-FDM (e.g., also known as discrete Fourier transform spread OFDM (DFT-s-OFDM) ) on the uplink (UL) , as well as supporting beamforming, multiple-input multiple-output (MIMO) antenna technology, and carrier aggregation.
[0006] Artificial neural networks may comprise interconnected groups of artificial neurons (e.g., neuron models) . The artificial neural network may be a computational device or represented as a method to be performed by a computational device. Convolutional neural networks, such as deep convolutional neural networks, are a type of feed-forward artificial neural network. Convolutional neural networks may include layers of neurons that may be configured in a tiled receptive field. It would be desirable to apply neural network processing to wireless communications to achieve greater efficiencies.
[0007] Machine learning (ML) monitoring involves monitoring mismatches between data and a ML model. The mismatches between the data and the ML model may, for example, include data and concept drift after the ML model has been deployed. The ML model may be associated with a desirable data distribution, such as the distribution of a training dataset for which the ML model was designed to operate. The relationship between the input data and the output data may change over time that, in turn, may impact the underlying mapping function. As a result, the accuracy of the ML model may decline.
[0008] Concept drift may, for instance, occur if the ML model is operating in a different environment or under different operating conditions than a specific environment and operating conditions under which the ML model had been trained. As such, the performance may be significantly degraded.SUMMARY
[0009] In some aspects of the present disclosure, a method of wireless communication performed by a user equipment (UE) includes operating, by the UE, a first machine learning (ML) model of a number of first ML models. The first ML model is associated with wireless communication. The method also includes monitoring, by the UE, the first ML model using a second ML model. The second ML model is configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.
[0010] Various aspects of the present disclosure are directed to an apparatus including means for operating, by the UE, a first machine learning (ML) model of a number of first ML models. The first ML model is associated with wireless communication. The apparatus further includes means for monitoring, by the UE, the first ML model using a second ML model. The second ML model is configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.
[0011] In some aspects of the present disclosure, a non-transitory computer-readable medium with non-transitory program code recorded thereon is disclosed. The program code is executed by a processor and includes program code to operate, by the UE, a first machine learning (ML) model of a number of first ML models. The first ML model is associated with wireless communication. The program code further includes program code to monitor, by the UE, the first ML model using a second ML model. The second ML model is configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.
[0012] Various aspects of the present disclosure are directed to an apparatus having a memory and one or more processors coupled to the memory. The processor (s) is configured to operate, by the UE, a first machine learning (ML) model of a number of first ML models. The first ML model is associated with wireless communication. The apparatus also includes means for monitor, by the UE, the first ML model using a second ML model. The second ML model is configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.
[0013] In some aspects of the present disclosure, a method of wireless communication performed by a network device includes receiving, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication. The method also includes using a second ML model to monitor the first ML model based on the distance measure information. The method further includes determining, based on the distance measure information, whether to select a different first ML model of a number of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0014] Various aspects of the present disclosure are directed to an apparatus including means for receiving, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication. The apparatus also includes means for using a second ML model to monitor the first ML model based on the distance measure information. The apparatus further includes means for determining, based on the distance measure information, whether to select a different first ML model of a number of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0015] In some aspects of the present disclosure, a non-transitory computer-readable medium with non-transitory program code recorded thereon is disclosed. The program code is executed by a processor and includes program code to receive, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication. The program code also includes program code to use a second ML model to monitor the first ML model based on the distance measure information. The program code further includes program code to determine, based on the distance measure information, whether to select a different first ML model of a number of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0016] Another aspect of the present disclosure is directed to an apparatus having at least one memory and one or more processors coupled to the at least one memory. The processor (s) is configured to receive, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication. The apparatus further includes means for use a second ML model to monitor the first ML model based on the distance measure information. The apparatus still further includes means for determine, based on the distance measure information, whether to select a different first ML model of a number of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0017] Aspects generally include a method, apparatus, system, computer program product, non-transitory computer-readable medium, user equipment, base station, wireless communication device, and processing system as substantially described with reference to and as illustrated by the accompanying drawings and specification.
[0018] The foregoing has outlined rather broadly the features and technical advantages of examples according to the disclosure in order that the detailed description that follows may be better understood. Additional features and advantages will be described. The conception and specific examples disclosed may be readily utilized as a basis for modifying or designing other structures for carrying out the same purposes of the present disclosure. Such equivalent constructions do not depart from the scope of the appended claims. Characteristics of the concepts disclosed, both their organization and method of operation, together with associated advantages will be better understood from the following description when considered in connection with the accompanying figures. Each of the figures is provided for the purposes of illustration and description, and not as a definition of the limits of the claims.BRIEF DESCRIPTION OF THE DRAWINGS
[0019] So that features of the present disclosure can be understood in detail, a particular description may be had by reference to aspects, some of which are illustrated in the appended drawings. It is to be noted, however, that the appended drawings illustrate only certain aspects of this disclosure and are therefore not to be considered limiting of its scope, for the description may admit to other equally effective aspects. The same reference numbers in different drawings may identify the same or similar elements.
[0020] FIGURE 1 is a block diagram conceptually illustrating an example of a wireless communications network, in accordance with various aspects of the present disclosure.
[0021] FIGURE 2 is a block diagram conceptually illustrating an example of a base station in communication with a user equipment (UE) in a wireless communications network, in accordance with various aspects of the present disclosure.
[0022] FIGURE 3 is a block diagram illustrating an example disaggregated base station architecture, in accordance with various aspects of the present disclosure.
[0023] FIGURE 4 illustrates an example implementation of designing a neural network using a system-on-a-chip (SOC) , including a general-purpose processor, in accordance with certain aspects of the present disclosure.
[0024] FIGURES 5A, 5B, and 5C are diagrams illustrating a neural network, in accordance with various aspects of the present disclosure.
[0025] FIGURE 5D is a diagram illustrating an exemplary deep convolutional network (DCN) , in accordance with various aspects of the present disclosure.
[0026] FIGURE 6 is a block diagram illustrating an exemplary deep convolutional network (DCN) , in accordance with various aspects of the present disclosure.
[0027] FIGURE 7 is a graph illustrating an example of machine learning (ML) -based monitoring based on K-nearest neighbors (KNN) , in accordance with various aspects of the present disclosure.
[0028] FIGURE 8 is diagram illustrating an example of ML model monitoring complexity reduction by diluting training samples, in accordance with various aspects of the present disclosure.
[0029] FIGURE 9 is a diagram illustrating an example of ML model monitoring complexity reduction by clustering training samples, in accordance with various aspects of the present disclosure.
[0030] FIGURE 10 is a flow diagram illustrating an example process for ML model monitoring performed by a UE, in accordance with various aspects of the present disclosure.
[0031] FIGURE 11 is a flow diagram illustrating an example process for ML model monitoring performed by a base station, in accordance with various aspects of the present disclosure.DETAILED DESCRIPTION
[0032] Various aspects of the disclosure are described more fully below with reference to the accompanying drawings. This disclosure may, however, be embodied in many different forms and should not be construed as limited to any specific structure or function presented throughout this disclosure. Rather, these aspects are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. Based on the teachings, one skilled in the art should appreciate that the scope of the disclosure is intended to cover any aspect of the disclosure, whether implemented independently of or combined with any other aspect of the disclosure. For example, an apparatus may be implemented or a method may be practiced using any number of the aspects set forth. In addition, the scope of the disclosure is intended to cover such an apparatus or method, which is practiced using other structure, functionality, or structure and functionality in addition to or other than the various aspects of the disclosure set forth. It should be understood that any aspect of the disclosure disclosed may be embodied by one or more elements of a claim.
[0033] Several aspects of telecommunications systems will now be presented with reference to various apparatuses and techniques. These apparatuses and techniques will be described in the following detailed description and illustrated in the accompanying drawings by various blocks, modules, components, circuits, steps, processes, algorithms, and / or the like (collectively referred to as “elements” ) . These elements may be implemented using hardware, software, or combinations thereof. Whether such elements are implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system.
[0034] It should be noted that while aspects may be described using terminology commonly associated with 5G, 6G, and other related wireless technologies, aspects of the present disclosure can be applied in other generation-based communications systems, such as and including (but not limited to) 3G and / or 4G technologies.
[0035] As described, machine learning model concept drift involves mismatch between the data, such as a training data set, and the machine learning model that may develop after the machine learning model has been deployed. Thus, it may be desirable to monitor a machine learning (ML) model using certain metrics to keep the ML model in a desirable operating scenario.
[0036] When a data-model mismatch (e.g., concept drift) is detected, possible actions may include (but is not limited to) switching between different ML models (i.e., selecting a different ML model) , reverting to a non-ML model, training a global ML model that generalizes well under different environments, or online re-training or fine tuning the ML models, for example.
[0037] When an ML model is trained under a specific environment (e.g., indoor vs. outdoor scenarios, heavy vs. sparse traffic scenarios, urban vs. rural scenarios, etc. ) or configuration, the distribution of the ML model inputs and outputs, such as ground truth labels, may be affected. By observing the input data distribution or output data distribution after inference, the monitoring process may decide to take a different action, such as switching the ML model, falling back to non-ML operation, or the like.
[0038] One approach for monitoring the ML model is by measuring the distribution similarity. For example, the similarity between the inference input data distribution and the different ML models’ input distributions may be measured using a Kolmogorov-Smirnov (KS) distance, Earth mover's distance, Kullback–Leibler (KL) divergence, or other similarity measures.
[0039] The monitoring process may select the ML model (e.g., a current ML model or a different ML model) based on input data distribution similarity with the inference data distribution (e.g., having a highest input data distribution similarity with the inference data distribution) . The monitoring process may set thresholds on the distribution similarity before determining to revert to the non-ML operation. This approach beneficially offers a simple low-complexity approach for ML model monitoring. However, the distribution similarity approach relies on the availability of prior recent inference data to generate the inference data distribution and compare the inference data distribution with the training data distribution.
[0040] The ML monitoring process may compare the similarity in the input and output data distributions between the training data and the inference data. If there is a high similarity between the training data and the inference data distributions, then the monitoring process may determine that the ML model is suitable for inference. Conversely, if there is low similarity between the training data and the inference data distributions, then the monitoring process may determine that the ML model is not suitable for inference. In this case, it may be determined, by the monitoring process, to switch the ML model or revert to non-ML operation.
[0041] Still, monitoring the ML model based on the data distributions assumes the availability of inference data distributions, which may not be always available. For example, if the UE or network did not perform any recent inference, then the monitoring process may not have any inference data distribution with which to assess the applicability of a specific ML model. In such scenarios, conventional approaches may be unable to monitor the ML model using the data distributions.
[0042] The performance of data distribution-based monitoring may be further enhanced if additional processing to the input data is carried out, for example, using an ML monitoring process. By performing additional processing on the data distributions (e.g., by using a classical ML model) , the detection of data or concept drifts may be enhanced.
[0043] Accordingly, to address the distribution similarity challenges, aspects of the present disclosure are directed to using an ML-based monitoring process to monitor the ML model, which can be applied to as few as a single data instance. For example, in some aspects, the ML-based monitoring process may rely on the K-nearest neighbors (KNN) information (for example, KNN distance, KNN majority vote, K-means distance, etc. ) or the like to monitor the ML model. The proximity between the newly received inference data instance and the training data may be used to monitor the ML model applicability, select an ML model out of a group of ML models, revert to non-ML operation, or determine whether to finetune an ML model.
[0044] Particular aspects of the subject matter described in this disclosure can be implemented to realize one or more of the following potential advantages. In some examples, the described techniques reduce concept drift and increase accuracy and utility of a ML model.
[0045] FIGURE 1 is a diagram illustrating a network 100 in which aspects of the present disclosure may be practiced. The network 100 may be a 5G or NR network or some other wireless network, such as an LTE network. The wireless network 100 may include a number of BSs 110 (shown as BS 110a, BS 110b, BS 110c, and BS 110d) and other network entities. A BS is an entity that communicates with user equipment (UEs) and may also be referred to as a base station, a NR BS, a Node B, a gNB, a 5G node B, an access point, a transmit and receive point (TRP) , a network node, a network entity, and / or the like. A BS can be implemented as an aggregated base station, as a disaggregated base station, an integrated access and backhaul (IAB) node, a relay node, a sidelink node, etc. The BS can be implemented in an aggregated or monolithic base station architecture, or alternatively, in a disaggregated base station architecture, and may include one or more of a central unit (CU) , a distributed unit (DU) , a radio unit (RU) , a near-real time (near-RT) RAN intelligent controller (RIC) , or a non-real time (non-RT) RIC. Each BS may provide communications coverage for a particular geographic area. In 3GPP, the term “cell” can refer to a coverage area of a BS and / or a BS subsystem serving this coverage area, depending on the context in which the term is used.
[0046] A BS may provide communications coverage for a macro cell, a pico cell, a femto cell, and / or another type of cell. A macro cell may cover a relatively large geographic area (e.g., several kilometers in radius) and may allow unrestricted access by UEs with service subscription. A pico cell may cover a relatively small geographic area and may allow unrestricted access by UEs with service subscription. A femto cell may cover a relatively small geographic area (e.g., a home) and may allow restricted access by UEs having association with the femto cell (e.g., UEs in a closed subscriber group (CSG) ) . A BS for a macro cell may be referred to as a macro BS. A BS for a pico cell may be referred to as a pico BS. A BS for a femto cell may be referred to as a femto BS or a home BS. In the example shown in FIGURE 1, a BS 110a may be a macro BS for a macro cell 102a, a BS 110b may be a pico BS for a pico cell 102b, and a BS 110c may be a femto BS for a femto cell 102c. A BS may support one or multiple (e.g., three) cells. The terms “eNB, ” “base station, ” “NR BS, ” “gNB, ” “AP, ” “node B, ” “5G NB, ” “TRP, ” and “cell” may be used interchangeably.
[0047] In some aspects, a cell may not necessarily be stationary, and the geographic area of the cell may move according to the location of a mobile BS. In some aspects, the BSs may be interconnected to one another and / or to one or more other BSs or network nodes (not shown) in the wireless network 100 through various types of backhaul interfaces such as a direct physical connection, a virtual network, and / or the like using any suitable transport network.
[0048] The wireless network 100 may also include relay stations. A relay station is an entity that can receive a transmission of data from an upstream station (e.g., a BS or a UE) and send a transmission of the data to a downstream station (e.g., a UE or a BS) . A relay station may also be a UE that can relay transmissions for other UEs. In the example shown in FIGURE 1, a relay station 110d may communicate with macro BS 110a and a UE 120d in order to facilitate communications between the BS 110a and UE 120d. A relay station may also be referred to as a relay BS, a relay base station, a relay, and / or the like.
[0049] The wireless network 100 may be a heterogeneous network that includes BSs of different types, e.g., macro BSs, pico BSs, femto BSs, relay BSs, and / or the like. These different types of BSs may have different transmit power levels, different coverage areas, and different impact on interference in the wireless network 100. For example, macro BSs may have a high transmit power level (e.g., 5 to 40 Watts) whereas pico BSs, femto BSs, and relay BSs may have lower transmit power levels (e.g., 0.1 to 2 Watts) .
[0050] A network controller 130 may couple to a set of BSs and may provide coordination and control for these BSs. The network controller 130 may communicate with the BSs via a backhaul. The BSs may also communicate with one another (e.g., directly or indirectly) via a wireless or wireline backhaul.
[0051] UEs 120 (e.g., 120a, 120b, 120c) may be dispersed throughout the wireless network 100, and each UE may be stationary or mobile. A UE may also be referred to as an access terminal, a terminal, a mobile station, a subscriber unit, a station, and / or the like. A UE may be a cellular phone (e.g., a smart phone) , a personal digital assistant (PDA) , a wireless modem, a wireless communications device, a handheld device, a laptop computer, a cordless phone, a wireless local loop (WLL) station, a tablet, a camera, a gaming device, a netbook, a smartbook, an ultrabook, a medical device or equipment, biometric sensors / devices, wearable devices (smart watches, smart clothing, smart glasses, smart wrist bands, smart jewelry (e.g., smart ring, smart bracelet) ) , an entertainment device (e.g., a music or video device, or a satellite radio) , a vehicular component or sensor, smart meters / sensors, industrial manufacturing equipment, a global positioning system device, or any other suitable device that is configured to communicate via a wireless or wired medium.
[0052] Some UEs may be considered machine-type communications (MTC) or evolved or enhanced machine-type communications (eMTC) UEs. MTC and eMTC UEs include, for example, robots, drones, remote devices, sensors, meters, monitors, location tags, and / or the like, that may communicate with a base station, another device (e.g., remote device) , or some other entity. A wireless node may provide, for example, connectivity for or to a network (e.g., a wide area network such as Internet or a cellular network) via a wired or wireless communications link. Some UEs may be considered Internet-of-Things (IoT) devices, and / or may be implemented as NB-IoT (narrowband internet of things) devices. Some UEs may be considered a customer premises equipment (CPE) . UE 120 may be included inside a housing that houses components of UE 120, such as processor components, memory components, and / or the like.
[0053] In general, any number of wireless networks may be deployed in a given geographic area. Each wireless network may support a particular radio access technology (RAT) and may operate on one or more frequencies. A RAT may also be referred to as a radio technology, an air interface, and / or the like. A frequency may also be referred to as a carrier, a frequency channel, and / or the like. Each frequency may support a single RAT in a given geographic area in order to avoid interference between wireless networks of different RATs. In some cases, NR or 5G RAT networks may be deployed.
[0054] In some aspects, two or more UEs 120 (e.g., shown as UE 120a and UE 120e) may communicate directly using one or more sidelink channels (e.g., without using a base station 110 as an intermediary to communicate with one another) . For example, the UEs 120 may communicate using peer-to-peer (P2P) communications, device-to-device (D2D) communications, a vehicle-to-everything (V2X) protocol (e.g., which may include a vehicle-to-vehicle (V2V) protocol, a vehicle-to-infrastructure (V2I) protocol, and / or the like) , a mesh network, and / or the like. In this case, the UE 120 may perform scheduling operations, resource selection operations, and / or other operations described elsewhere as being performed by the base station 110. For example, the base station 110 may configure a UE 120 via downlink control information (DCI) , radio resource control (RRC) signaling, a media access control-control element (MAC-CE) or via system information (e.g., a system information block (SIB) .
[0055] The UEs 120 may include a machine learning (ML) monitoring module 140. For brevity, only one UE 120d is shown as including the ML monitoring module 140. The ML monitoring module 140 may monitor, by the UE, a first machine learning model to detect concept drift using a second machine learning model, the second machine learning model being configured to monitor the first machine learning model based on one or more distance measures between a first dataset for training the first machine learning model and one or more inference data instances generated by the first machine learning model.
[0056] The network controller 130 or the base stations 110 or any other network device (e.g., as seen in FIGURE 3) may include an ML monitoring module 138 for receiving, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication, using a second ML model to monitor the first ML model based on the distance measure information and determining, based on the distance measure information, whether to select a different first ML model of a plurality of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0057] As indicated above, FIGURE 1 is provided merely as an example. Other examples may differ from what is described with regard to FIGURE 1.
[0058] FIGURE 2 shows a block diagram of a design 200 of the base station 110 and UE 120, which may be one of the base stations and one of the UEs in FIGURE 1. The base station 110 may be equipped with T antennas 234a through 234t, and UE 120 may be equipped with R antennas 252a through 252r, where in general T ≥ 1 and R ≥ 1.
[0059] At the base station 110, a transmit processor 220 may receive data from a data source 212 for one or more UEs, select one or more modulation and coding schemes (MCS) for each UE based at least in part on channel quality indicators (CQIs) received from the UE, process (e.g., encode and modulate) the data for each UE based at least in part on the MCS (s) selected for the UE, and provide data symbols for all UEs. Decreasing the MCS lowers throughput but increases reliability of the transmission.
[0060] The transmit processor 220 may also process system information (e.g., for semi-static resource partitioning information (SRPI) and / or the like) and control information (e.g., CQI requests, grants, upper layer signaling, and / or the like) and provide overhead symbols and control symbols. The transmit processor 220 may also generate reference symbols for reference signals (e.g., the cell-specific reference signal (CRS) ) and synchronization signals (e.g., the primary synchronization signal (PSS) and secondary synchronization signal (SSS) ) . A transmit (TX) multiple-input multiple-output (MIMO) processor 230 may perform spatial processing (e.g., precoding) on the data symbols, the control symbols, the overhead symbols, and / or the reference symbols, if applicable, and may provide T output symbol streams to T modulators (MODs) 232a through 232t.
[0061] Each modulator 232 may process a respective output symbol stream (e.g., for orthogonal frequency division multiplexing (OFDM) and / or the like) to obtain an output sample stream. Each modulator 232 may further process (e.g., convert to analog, amplify, filter, and upconvert) the output sample stream to obtain a downlink signal. T downlink signals from modulators 232a through 232t may be transmitted via T antennas 234a through 234t, respectively. According to various aspects described in more detail below, the synchronization signals can be generated with location encoding to convey additional information.
[0062] At the UE 120, antennas 252a through 252r may receive the downlink signals from the base station 110 and / or other base stations and may provide received signals to demodulators (DEMODs) 254a through 254r, respectively. Each demodulator 254 may condition (e.g., filter, amplify, downconvert, and digitize) a received signal to obtain input samples. Each demodulator 254 may further process the input samples (e.g., for OFDM and / or the like) to obtain received symbols. A MIMO detector 256 may obtain received symbols from all R demodulators 254a through 254r, perform MIMO detection on the received symbols if applicable, and provide detected symbols. A receive processor 258 may process (e.g., demodulate and decode) the detected symbols, provide decoded data for the UE 120 to a data sink 260, and provide decoded control information and system information to a controller / processor 280. A channel processor may determine reference signal received power (RSRP) , received signal strength indicator (RSSI) , reference signal received quality (RSRQ) , channel quality indicator (CQI) , and / or the like. In some aspects, one or more components of the UE 120 may be included in a housing.
[0063] On the uplink, at the UE 120, a transmit processor 264 may receive and process data from a data source 262 and control information (e.g., for reports comprising RSRP, RSSI, RSRQ, CQI, and / or the like) from the controller / processor 280. Transmit processor 264 may also generate reference symbols for one or more reference signals. The symbols from the transmit processor 264 may be precoded by a TX MIMO processor 266 if applicable, further processed by modulators 254a through 254r (e.g., for discrete Fourier transform spread OFDM (DFT-s-OFDM) , CP-OFDM, and / or the like) , and transmitted to the base station 110. At the base station 110, the uplink signals from the UE 120 and other UEs may be received by the antennas 234, processed by the demodulators 254, detected by a MIMO detector 236 if applicable, and further processed by a receive processor 238 to obtain decoded data and control information sent by the UE 120. The receive processor 238 may provide the decoded data to a data sink 239 and the decoded control information to a controller / processor 240. The base station 110 may include communications unit 244 and communicate to the network controller 130 via the communications unit 244. The network controller 130 may include a communications unit 294, a controller / processor 290, and a memory 292.
[0064] The controller / processor 240 of the base station 110, the controller / processor 280 of the UE 120, and / or any other component (s) of FIGURE 2 may perform one or more techniques associated with determining membership for zone-based federated learning as described in more detail elsewhere. For example, the controller / processor 240 of the base station 110, the controller / processor 280 of the UE 120, and / or any other component (s) of FIGURE 2 may perform or direct operations of, for example, the processes of FIGURES 9-10 and / or other processes as described. Memories 242 and 282 may store data and program codes for the base station 110 and UE 120, respectively. A scheduler 246 may schedule UEs for data transmission on the downlink and / or uplink.
[0065] In some aspects, the UE 120 and / or base station 110 may include means for operating, means for monitoring, means for receiving, and means for using a second ML model and / or means for determining. Such means may include one or more components of the UE 120 or base station 110 described in connection with FIGURE 2.
[0066] As indicated above, FIGURE 2 is provided merely as an example. Other examples may differ from what is described with regard to FIGURE 2.
[0067] In some cases, different types of devices supporting different types of applications and / or services may coexist in a cell. Examples of different types of devices include UE handsets, customer premises equipment (CPEs) , vehicles, Internet of Things (IoT) devices, and / or the like. Examples of different types of applications include ultra-reliable low-latency communications (URLLC) applications, massive machine-type communications (mMTC) applications, enhanced mobile broadband (eMBB) applications, vehicle-to-anything (V2X) applications, and / or the like. Furthermore, in some cases, a single device may support different applications or services simultaneously.
[0068] Deployment of communication systems, such as 5G new radio (NR) systems, may be arranged in multiple manners with various components or constituent parts. In a 5G NR system, or network, a network node, a network entity, a mobility element of a network, a radio access network (RAN) node, a core network node, a network element, or a network equipment, such as a base station (BS) , or one or more units (or one or more components) performing base station functionality, may be implemented in an aggregated or disaggregated architecture. For example, a BS (such as a Node B (NB) , an evolved NB (eNB) , an NR BS, 5G NB, an access point (AP) , a transmit and receive point (TRP) , or a cell, etc. ) may be implemented as an aggregated base station (also known as a standalone BS or a monolithic BS) or a disaggregated base station.
[0069] An aggregated base station may be configured to utilize a radio protocol stack that is physically or logically integrated within a single RAN node. A disaggregated base station may be configured to utilize a protocol stack that is physically or logically distributed among two or more units (such as one or more central or centralized units (CUs) , one or more distributed units (DUs) , or one or more radio units (RUs) ) . In some aspects, a CU may be implemented within a RAN node, and one or more DUs may be co-located with the CU, or alternatively, may be geographically or virtually distributed throughout one or multiple other RAN nodes. The DUs may be implemented to communicate with one or more RUs. Each of the CU, DU, and RU also can be implemented as virtual units, i.e., a virtual central unit (VCU) , a virtual distributed unit (VDU) , or a virtual radio unit (VRU) .
[0070] Base station-type operation or network design may consider aggregation characteristics of base station functionality. For example, disaggregated base stations may be utilized in an integrated access backhaul (IAB) network, an open radio access network (O-RAN (such as the network configuration sponsored by the O-RAN Alliance) ) , or a virtualized radio access network (vRAN, also known as a cloud radio access network (C-RAN) ) . Disaggregation may include distributing functionality across two or more units at various physical locations, as well as distributing functionality for at least one unit virtually, which can enable flexibility in network design. The various units of the disaggregated base station, or disaggregated RAN architecture, can be configured for wired or wireless communication with at least one other unit.
[0071] FIGURE 3 shows a diagram illustrating an example disaggregated base station 300 architecture. The disaggregated base station 300 architecture may include one or more central units (CUs) 310 that can communicate directly with a core network 320 via a backhaul link, or indirectly with the core network 320 through one or more disaggregated base station units (such as a near-real time (near-RT) RAN intelligent controller (RIC) 325 via an E2 link, or a non-real time (non-RT) RIC 315 associated with a service management and orchestration (SMO) framework 305, or both) . A CU 310 may communicate with one or more distributed units (DUs) 330 via respective midhaul links, such as an F1 interface. The DUs 330 may communicate with one or more radio units (RUs) 340 via respective fronthaul links. The RUs 340 may communicate with respective UEs 120 via one or more radio frequency (RF) access links. In some implementations, the UE 120 may be simultaneously served by multiple RUs 340.
[0072] Each of the units (e.g., the CUs 310, the DUs 330, the RUs 340, as well as the near-RT RICs 325, the non-RT RICs 315, and the SMO framework 305) may include one or more interfaces or be coupled to one or more interfaces configured to receive or transmit signals, data, or information (collectively, signals) via a wired or wireless transmission medium. Each of the units, or an associated processor or controller providing instructions to the communication interfaces of the units, can be configured to communicate with one or more of the other units via the transmission medium. For example, the units can include a wired interface configured to receive or transmit signals over a wired transmission medium to one or more of the other units. Additionally, the units can include a wireless interface, which may include a receiver, a transmitter or transceiver (such as a radio frequency (RF) transceiver) , configured to receive or transmit signals, or both, over a wireless transmission medium to one or more of the other units.
[0073] In some aspects, the CU 310 may host one or more higher layer control functions. Such control functions can include radio resource control (RRC) , packet data convergence protocol (PDCP) , service data adaptation protocol (SDAP) , or the like. Each control function can be implemented with an interface configured to communicate signals with other control functions hosted by the CU 310. The CU 310 may be configured to handle user plane functionality (i.e., Central Unit –User Plane (CU-UP) ) , control plane functionality (i.e., Central Unit –Control Plane (CU-CP) ) , or a combination thereof. In some implementations, the CU 310 can be logically split into one or more CU-UP units and one or more CU-CP units. The CU-UP unit can communicate bi-directionally with the CU-CP unit via an interface, such as the E1 interface when implemented in an O-RAN configuration. The CU 310 can be implemented to communicate with the DU 330, as necessary, for network control and signaling.
[0074] The DU 330 may correspond to a logical unit that includes one or more base station functions to control the operation of one or more RUs 340. In some aspects, the DU 330 may host one or more of a radio link control (RLC) layer, a 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, or the like) depending, at least in part, on a functional split, such as those defined by the Third Generation Partnership Project (3GPP) . In some aspects, the DU 330 may further host one or more low PHY layers. Each layer (or module) can be implemented with an interface configured to communicate signals with other layers (and modules) hosted by the DU 330, or with the control functions hosted by the CU 310.
[0075] Lower-layer functionality can be implemented by one or more RUs 340. In some deployments, an RU 340, controlled by a DU 330, 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, or the like) , or both, based at least in part on the functional split, such as a lower layer functional split. In such an architecture, the RU (s) 340 can be implemented to handle over the air (OTA) communication with one or more UEs 120. In some implementations, real-time and non-real-time aspects of control and user plane communication with the RU (s) 340 can be controlled by the corresponding DU 330. In some scenarios, this configuration can enable the DU (s) 330 and the CU 310 to be implemented in a cloud-based RAN architecture, such as a vRAN architecture.
[0076] The SMO Framework 305 may be configured to support RAN deployment and provisioning of non-virtualized and virtualized network elements. For non-virtualized network elements, the SMO Framework 305 may be configured to support the deployment of dedicated physical resources for RAN coverage requirements, which may be managed via an operations and maintenance interface (such as an O1 interface) . For virtualized network elements, the SMO Framework 305 may be configured to interact with a cloud computing platform (such as an open cloud (O-Cloud) 390) to perform network element life cycle management (such as to instantiate virtualized network elements) via a cloud computing platform interface (such as an O2 interface) . Such virtualized network elements can include, but are not limited to, CUs 310, DUs 330, RUs 340, and near-RT RICs 325. In some implementations, the SMO Framework 305 can communicate with a hardware aspect of a 4G RAN, such as an open eNB (O-eNB) X11, via an O1 interface. Additionally, in some implementations, the SMO Framework 305 can communicate directly with one or more RUs 340 via an O1 interface. The SMO Framework 305 also may include a non-RT RIC 315 configured to support functionality of the SMO Framework 305.
[0077] The non-RT RIC 315 may be configured to include a logical function that enables non-real-time control and optimization of RAN elements and resources, a machine learning (ML) workflows including model training and updates, or policy-based guidance of applications / features in the near-RT RIC 325. The non-RT RIC 315 may be coupled to or communicate with (such as via an A1 interface) the near-RT RIC 325. The near-RT RIC 325 may be configured to include a logical function that enables near-real-time control and optimization of RAN elements and resources via data collection and actions over an interface (such as via an E2 interface) connecting one or more CUs 310, one or more DUs 330, or both, as well as an O-eNB, with the near-RT RIC 325.
[0078] In some implementations, to generate ML models to be deployed in the near-RT RIC 325, the non-RT RIC 315 may receive parameters or external enrichment information from external servers. Such information may be utilized by the near-RT RIC 325 and may be received at the SMO Framework 305 or the non-RT RIC 315 from non-network data sources or from network functions. In some examples, the non-RT RIC 315 or the near-RT RIC 325 may be configured to tune RAN behavior or performance. For example, the non-RT RIC 315 may monitor long-term trends and patterns for performance and employ ML models to perform corrective actions through the SMO Framework 305 (such as reconfiguration via O1) or via creation of RAN management policies (such as A1 policies) .
[0079] FIGURE 4 illustrates an example implementation of a system-on-a-chip (SOC) 400, which may include a central processing unit (CPU) 402 or a multi-core CPU configured for generating gradients for neural network training, in accordance with certain aspects of the present disclosure. The SOC 400 may be included in the base station 110 or UE 120. Variables (e.g., neural signals and synaptic weights) , system parameters associated with a computational device (e.g., neural network with weights) , delays, frequency bin information, and task information may be stored in a memory block associated with a neural processing unit (NPU) 408, in a memory block associated with a CPU 402, in a memory block associated with a graphics processing unit (GPU) 404, in a memory block associated with a digital signal processor (DSP) 406, in a memory block 418, or may be distributed across multiple blocks. Instructions executed at the CPU 402 may be loaded from a program memory associated with the CPU 402 or may be loaded from a memory block 418.
[0080] The SOC 400 may also include additional processing blocks tailored to specific functions, such as a GPU 404, a DSP 406, a connectivity block 410, which may include fifth generation (5G) connectivity, fourth generation long term evolution (4G LTE) connectivity, Wi-Fi connectivity, USB connectivity, Bluetooth connectivity, and the like, and a multimedia processor 412 that may, for example, detect and recognize gestures. In one implementation, the NPU is implemented in the CPU, DSP, and / or GPU. The SOC 400 may also include a sensor processor 414, image signal processors (ISPs) 416, and / or navigation module 420, which may include a global positioning system.
[0081] The SOC 400 may be based on an ARM instruction set. In an aspect of the present disclosure, the instructions loaded into the general-purpose processor 402 may comprise code to operate, by the UE, a first machine learning (ML) model of a plurality of first ML models. The first ML model is associated with wireless communication. The general-purpose processor 402 may also comprise code to monitor, by the UE, the first ML model using a second ML model. The second ML model is configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operating of the first ML model.
[0082] In various aspects of the present disclosure, the instructions loaded into the general-purpose processor 402 may comprise code to receive, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication. The general-purpose processor 402 may also comprise code to using a second ML model to monitor the first ML model based on the distance measure information. The general-purpose processor 402 may further comprise code to determine, based on the distance measure information, whether to select a different first ML model of a plurality of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0083] Deep learning architectures may perform an object recognition task by learning to represent inputs at successively higher levels of abstraction in each layer, thereby building up a useful feature representation of the input data. In this way, deep learning addresses a major bottleneck of traditional machine learning. Prior to the advent of deep learning, a machine learning approach to an object recognition problem may have relied heavily on human engineered features, perhaps in combination with a shallow classifier. A shallow classifier may be a two-class linear classifier, for example, in which a weighted sum of the feature vector components may be compared with a threshold to predict to which class the input belongs. Human engineered features may be templates or kernels tailored to a specific problem domain by engineers with domain expertise. Deep learning architectures, in contrast, may learn to represent features that are similar to what a human engineer might design, but through training. Furthermore, a deep network may learn to represent and recognize new types of features that a human might not have considered.
[0084] A deep learning architecture may learn a hierarchy of features. If presented with visual data, for example, the first layer may learn to recognize relatively simple features, such as edges, in the input stream. In another example, if presented with auditory data, the first layer may learn to recognize spectral power in specific frequencies. The second layer, taking the output of the first layer as input, may learn to recognize combinations of features, such as simple shapes for visual data or combinations of sounds for auditory data. For instance, higher layers may learn to represent complex shapes in visual data or words in auditory data. Still higher layers may learn to recognize common visual objects or spoken phrases.
[0085] Deep learning architectures may perform especially well when applied to problems that have a natural hierarchical structure. For example, the classification of motorized vehicles may benefit from first learning to recognize wheels, windshields, and other features. These features may be combined at higher layers in different ways to recognize cars, trucks, and airplanes.
[0086] Neural networks may be designed with a variety of connectivity patterns. In feed-forward networks, information is passed from lower to higher layers, with each neuron in a given layer communicating to neurons in higher layers. A hierarchical representation may be built up in successive layers of a feed-forward network, as described above. Neural networks may also have recurrent or feedback (also called top-down) connections. In a recurrent connection, the output from a neuron in a given layer may be communicated to another neuron in the same layer. A recurrent architecture may be helpful in recognizing patterns that span more than one of the input data chunks that are delivered to the neural network in a sequence. A connection from a neuron in a given layer to a neuron in a lower layer is called a feedback (or top-down) connection. A network with many feedback connections may be helpful when the recognition of a high-level concept may aid in discriminating the particular low-level features of an input.
[0087] The connections between layers of a neural network may be fully connected or locally connected. FIGURE 5A illustrates an example of a fully connected neural network 502. In a fully connected neural network 502, a neuron in a first layer may communicate its output to every neuron in a second layer, so that each neuron in the second layer will receive input from every neuron in the first layer. FIGURE 5B illustrates an example of a locally connected neural network 504. In a locally connected neural network 504, a neuron in a first layer may be connected to a limited number of neurons in the second layer. More generally, a locally connected layer of the locally connected neural network 504 may be configured so that each neuron in a layer will have the same or a similar connectivity pattern, but with connections strengths that may have different values (e.g., 510, 512, 514, and 516) . The locally connected connectivity pattern may give rise to spatially distinct receptive fields in a higher layer because the higher layer neurons in a given region may receive inputs that are tuned through training to the properties of a restricted portion of the total input to the network.
[0088] One example of a locally connected neural network is a convolutional neural network. FIGURE 5C illustrates an example of a convolutional neural network 506. The convolutional neural network 506 may be configured such that the connection strengths associated with the inputs for each neuron in the second layer are shared (e.g., 508) . Convolutional neural networks may be well suited to problems in which the spatial location of inputs is meaningful.
[0089] One type of convolutional neural network is a deep convolutional network (DCN) . FIGURE 5D illustrates a detailed example of a DCN 500 designed to recognize visual features from an image 526 input from an image capturing device 530, such as a car-mounted camera. The DCN 500 of the current example may be trained to identify traffic signs and a number provided on the traffic sign. Of course, the DCN 500 may be trained for other tasks, such as identifying lane markings or identifying traffic lights.
[0090] The DCN 500 may be trained with supervised learning. During training, the DCN 500 may be presented with an image, such as the image 526 of a speed limit sign, and a forward pass may then be computed to produce an output 522. The DCN 500 may include a feature extraction section and a classification section. Upon receiving the image 526, a convolutional layer 532 may apply convolutional kernels (not shown) to the image 526 to generate a first set of feature maps 518. As an example, the convolutional kernel for the convolutional layer 532 may be a 5x5 kernel that generates 28x28 feature maps. In the present example, because four different feature maps are generated in the first set of feature maps 518, four different convolutional kernels were applied to the image 526 at the convolutional layer 532. The convolutional kernels may also be referred to as filters or convolutional filters.
[0091] The first set of feature maps 518 may be subsampled by a max pooling layer (not shown) to generate a second set of feature maps 520. The max pooling layer reduces the size of the first set of feature maps 518. That is, a size of the second set of feature maps 520, such as 14x14, is less than the size of the first set of feature maps 518, such as 28x28. The reduced size provides similar information to a subsequent layer while reducing memory consumption. The second set of feature maps 520 may be further convolved via one or more subsequent convolutional layers (not shown) to generate one or more subsequent sets of feature maps (not shown) .
[0092] In the example of FIGURE 5D, the second set of feature maps 520 is convolved to generate a first feature vector 524. Furthermore, the first feature vector 524 is further convolved to generate a second feature vector 528. Each feature of the second feature vector 528 may include a number that corresponds to a possible feature of the image 526, such as “sign, ” “60, ” and “100. ” A softmax function (not shown) may convert the numbers in the second feature vector 528 to a probability. As such, an output 522 of the DCN 500 may be a probability of the image 526 including one or more features.
[0093] In the present example, the probabilities in the output 522 for “sign” and “60” are higher than the probabilities of the others of the output 522, such as “30, ” “40, ” “50, ” “70, ” “80, ” “90, ” and “100” . Before training, the output 522 produced by the DCN 500 may likely be incorrect. Thus, an error may be calculated between the output 522 and a target output. The target output is the ground truth of the image 526 (e.g., “sign” and “60” ) . The weights of the DCN 500 may then be adjusted so the output 522 of the DCN 500 is more closely aligned with the target output.
[0094] To adjust the weights, a learning process may compute a gradient vector for the weights. The gradient may indicate an amount that an error would increase or decrease if the weight were adjusted. At the top layer, the gradient may correspond directly to the value of a weight connecting an activated neuron in the penultimate layer and a neuron in the output layer. In lower layers, the gradient may depend on the value of the weights and on the computed error gradients of the higher layers. The weights may then be adjusted to reduce the error. This manner of adjusting the weights may be referred to as “back propagation” as it involves a “backward pass” through the neural network.
[0095] In practice, the error gradient of weights may be calculated over a small number of examples, so that the calculated gradient approximates the true error gradient. This approximation method may be referred to as stochastic gradient descent. Stochastic gradient descent may be repeated until the achievable error rate of the entire system has stopped decreasing or until the error rate has reached a target level. After learning, the DCN 500 may be presented with new images (e.g., the speed limit sign of the image 526) and a forward pass through the DCN 500 may yield an output 522 that may be considered an inference or a prediction of the DCN 500.
[0096] Deep belief networks (DBNs) are probabilistic models comprising multiple layers of hidden nodes. DBNs may be used to extract a hierarchical representation of training data sets. A DBN may be obtained by stacking up layers of Restricted Boltzmann Machines (RBMs) . An RBM is a type of artificial neural network that can learn a probability distribution over a set of inputs. Because RBMs can learn a probability distribution in the absence of information about the class to which each input should be categorized, RBMs are often used in unsupervised learning. Using a hybrid unsupervised and supervised paradigm, the bottom RBMs of a DBN may be trained in an unsupervised manner and may serve as feature extractors, and the top RBM may be trained in a supervised manner (on a joint distribution of inputs from the previous layer and target classes) and may serve as a classifier.
[0097] DCNs are networks of convolutional networks, configured with additional pooling and normalization layers. DCNs have achieved state-of-the-art performance on many tasks. DCNs can be trained using supervised learning in which both the input and output targets are known for many exemplars and are used to modify the weights of the network by use of gradient descent methods.
[0098] DCNs may be feed-forward networks. In addition, as described above, the connections from a neuron in a first layer of a DCN to a group of neurons in the next higher layer are shared across the neurons in the first layer. The feed-forward and shared connections of DCNs may be exploited for fast processing. The computational burden of a DCN may be much less, for example, than that of a similarly sized neural network that comprises recurrent or feedback connections.
[0099] The processing of each layer of a convolutional network may be considered a spatially invariant template or basis projection. If the input is first decomposed into multiple channels, such as the red, green, and blue channels of a color image, then the convolutional network trained on that input may be considered three-dimensional, with two spatial dimensions along the axes of the image and a third dimension capturing color information. The outputs of the convolutional connections may be considered to form a feature map in the subsequent layer, with each element of the feature map (e.g., 220) receiving input from a range of neurons in the previous layer (e.g., feature maps 218) and from each of the multiple channels. The values in the feature map may be further processed with a non-linearity, such as a rectification, max (0, x) . Values from adjacent neurons may be further pooled, which corresponds to down sampling, and may provide additional local invariance and dimensionality reduction. Normalization, which corresponds to whitening, may also be applied through lateral inhibition between neurons in the feature map.
[0100] FIGURE 6 is a block diagram illustrating a DCN 650. The DCN 650 may include multiple different types of layers based on connectivity and weight sharing. As shown in FIGURE 6, the DCN 650 includes the convolution blocks 654A, 654B. Each of the convolution blocks 654A, 654B may be configured with a convolution layer (CONV) 656, a normalization layer (LNorm) 658, and a max pooling layer (MAX POOL) 660.
[0101] Although only two of the convolution blocks 654A, 654B are shown, the present disclosure is not so limiting, and instead, any number of the convolution blocks 654A, 654B may be included in the DCN 650 according to design preference.
[0102] The convolution layers 656 may include one or more convolutional filters, which may be applied to the input data to generate a feature map. The normalization layer 658 may normalize the output of the convolution filters. For example, the normalization layer 658 may provide whitening or lateral inhibition. The max pooling layer 660 may provide down sampling aggregation over space for local invariance and dimensionality reduction.
[0103] The parallel filter banks, for example, of a deep convolutional network may be loaded on a CPU 402 or GPU 404 of an SOC 400 (e.g., FIGURE 4) to achieve high performance and low power consumption. In alternative embodiments, the parallel filter banks may be loaded on the DSP 406 or an ISP 416 of an SOC 400. In addition, the DCN 650 may access other processing blocks that may be present on the SOC 400, such as sensor processor 414 and navigation module 420, dedicated, respectively, to sensors and navigation.
[0104] The DCN 650 may also include one or more fully connected layers 662 (FC1 and FC2) . The DCN 650 may further include a logistic regression (LR) layer 664. Between each layer 656, 658, 660, 662, 664 of the DCN 650 are weights (not shown) that are to be updated. The output of each of the layers (e.g., 656, 658, 660, 662, 664) may serve as an input of a succeeding one of the layers (e.g., 656, 658, 660, 662, 664) in the DCN 650 to learn hierarchical feature representations from input data 652 (e.g., images, audio, video, sensor data and / or other input data) supplied at the first of the convolution blocks 654A. The output of the DCN 650 is a classification score 666 for the input data 652. The classification score 666 may be a set of probabilities, where each probability is the probability of the input data, including a feature from a set of features.
[0105] As described, aspects of the present disclosure are directed to an ML-based monitoring process to monitor the ML model. That is, an ML model may be configured to monitor an ML model for detecting and / or addressing concept drift. For example, the ML-based monitoring process may rely on K-nearest neighbors (KNN) information (e.g., KNN distance, KNN majority vote, a K-means distance, etc. ) to monitor the ML model. In some aspects, the monitoring may be conducted using a single inference data instance.
[0106] FIGURE 7 is a graph 700 illustrating an example of ML-based monitoring based on KNN, in accordance with various aspects of the present disclosure. Referring to FIGURE 7, training data of a first dataset and training data for a second dataset are shown. The first dataset and the second dataset may correspond to respective trained ML models for spatial beam estimation for a user equipment (UE) . The UE (e.g., UE 120) may for example select an ML model trained according to the first dataset or an ML model trained according to the second dataset.
[0107] A second ML model may monitor the ML model for beam estimation operated by the UE. For instance, the second ML model may monitor the selected ML model based on the proximity between a new inference data instance 702 and the training data (e.g., dataset 1 and dataset 2) . Accordingly, the second ML model may use the proximity information to monitor the ML model applicability, select a different ML model out of a group of ML models, fallback to non-ML operation, or determine whether to fine tune the ML model, for example.
[0108] In the example of FIGURE 7, the KNN distance may be used for spatial beam prediction. In this example, there may be two fixed beams (e.g., two reference signal received power (RSRP) measurements) as inputs to the ML model for model monitoring. However, this is merely for ease of illustration, and the present specification is not so limiting. Rather, any number of inputs may be employed.
[0109] Accordingly, the KNN distance for the new inference data instance 702 may be calculated. As the UE (e.g., UE 120) receives a new inference data instance 702, the UE (e.g., UE 120) may measure the KNN distance (or KNN majority vote, K-means distance or other proximity measure) between the new inference data instance 702 and the two datasets (e.g., dataset 1 and dataset 2) used for training two ML models. The UE (e.g., UE 120) or the network (e.g., the base station 110) may select the ML model having the lowest KNN distance.
[0110] To calculate the KNN distance for a single inference data instance (e.g., 702) , the nearest K neighboring instances in each dataset may be identified. The KNN distance may then be calculated, as follows:
[0111] where represents the distance measure, i represents the index for the nearest neighbors for an inference instance in the dataset j.
[0112] In FIGURE 7, the UE (e.g., UE 120) may select ML model 1 (e.g., trained with dataset 1) because the KNN distance between the new inference data instance 702 and data instances of dataset 1 is less than the KNN distance between the new inference data instance 702 and the data instances of dataset 2. By utilizing the KNN information for ML model monitoring, there may be no requirement for the availability of prior recent inference data to generate the inference data distribution.
[0113] In various aspects, the network may configure the UE (e.g., UE 120) to report the KNN information such as the KNN distance or KNN majority vote, between the new inference data instance (e.g., 702) , the K-means distance and the nearest K neighbors in a specified training dataset, for example. Doing so may enable the network to monitor the ML model operated at the UE.
[0114] In some aspects, a hybrid ML model monitoring approach may be employed. For instance, the UE (e.g., UE 120) may conduct the ML model monitoring and the network may determine whether to select a different ML model out of a group of ML models, to revert to a non-ML operation, or to fine tune the current ML model. In other examples, the UE may operate the ML model for spatial beam prediction. The network may operate the second ML model to monitor the ML model operated by the UE.
[0115] The network may receive the proximity information from the UE. In some examples, the network (e.g., via a network device such as base station 110) may configure the UE (e.g., UE 120) to report the statistical properties of the KNN distance or KNN majority vote over multiple inference data instances. The network may, in turn, determine whether to select a different ML model out of a group of ML models, to revert to a non-ML operation, or to fine tune the current ML model. The network may transmit a message to the UE to configure the UE in accordance with the determination.
[0116] In some aspects, the second ML model may monitor the ML model operated by the UE and / or determine a modification (e.g., selecting a different ML model, reverting to a non-ML model, or fine tuning the ML model) according to a distance measure (e.g., KNN distance or the like) determined based on one or more inference data instances. However, depending on the environment, variation in the KNN distance over multiple inference data instances may not be significant. Hence, it may be beneficial to report a statistical property of the KNN information to reduce the reporting overhead. For example, in some aspects, the network may configure the UE (e.g., UE 120) to report an average, a variance (e.g., 5th percentile or 95th percentile) , or other statistical property of the KNN information over multiple inference data instances.
[0117] Furthermore, the network may configure the periodicity of reporting KNN information statistics from previous data instances. For example, the periodicity may be specified in milliseconds (ms) , according to a number of time slots, or according to the time duration used in calculating the statistics of KNN information, for instance. The network may also configure uplink resources for the UE (e.g., UE 120) to report the calculated KNN distance based on the configured periodicity.
[0118] The network may also configure KNN information parameters to be used by the UE (e.g., UE 120) for monitoring the ML model. For example, the network may configure the value of K to be used by the UE (e.g., UE 120) in calculating the KNN distance / majority vote. The value of K may represent a tradeoff between complexity of the KNN distance calculation (e.g., increased computation) and the accuracy of the ML model selection. Larger K values may indicate a higher complexity of the monitoring process. Smaller K values may indicate more errors in selecting the best model.
[0119] In various aspects, the network may also configure one or more thresholds on the KNN distance. In doing so, the network may enable the UE (e.g., UE 120) to switch between different ML models or fallback to non-ML operation, according to the threshold. For example, if the KNN distance for the dataset of the currently active ML model exceeds a predefined threshold, the UE (e.g., UE 120) may switch the ML model. Otherwise, the UE (e.g., UE 120) may continue using the same ML model.
[0120] In some aspects, the network may be configured for complexity reduction of the KNN information calculations. One challenge of monitoring the ML model using the KNN information is that the ML calculates the KNN distance from a new inference data instance to every other data instance in the training dataset, sorts the distances, and averages the K nearest distances. Thus, the monitoring complexity may be high. As such, the network may configure the UE (e.g., UE 120) to measure the KNN distance on specific dimensions. For instance, the inputs to the ML model may include two input features. For the spatial beam prediction use case, the UE (e.g., UE 120) may measure five beams, for example. The spatial beams may be measured on historically four different occasions and supplied as inputs to the ML models.
[0121] In this case, the ML monitoring process calculates a twenty-dimensional KNN distance between the inference data instance and the training dataset of the different ML models. To reduce the complexity of the KNN distance calculations, the network may configure the UE (e.g., UE 120) to calculate the KNN distance based on specific dimensions / features and exclude other dimensions. For example, the KNN distance may be calculated based on the five input measurements instead of the twenty input measurements. Using the network’s knowledge of the environment and the operating conditions, the network may identify the more important features or inputs for the KNN distance calculations monitoring, and which features can be excluded from the ML monitoring to reduce the complexity.
[0122] In some aspects, the network may configure the UE (e.g., UE 120) to dilute the training samples for the KNN information calculation. Another challenge with the KNN information calculation may be that the UE (e.g., UE 120) retains all training data samples, thereby increasing storage overhead as well as complexity in calculating the KNN distance. To address this challenge, the network may configure the UE (e.g., UE 120) to retain a diluted version of the training samples for monitoring the ML model using the KNN information.
[0123] FIGURE 8 is diagram 800 illustrating an example of ML model monitoring complexity reduction by diluting training samples, in accordance with aspects of the present disclosure. Referring to FIGURE 8, the diagram 800 includes a graph 700 illustrating the example of model monitoring using the KNN distance. To reduce the complexity of ML monitoring, for instance using the KNN distance rather than considering the complete training dataset (e.g., dataset 1 or dataset 2) , a reduced set of training data may be used for calculating the KNN distance. As shown in FIGURE 8, a subset of the dataset 1 measurements and a subset of dataset 2 measurements are included in a graph 802. The subset of dataset 1 measurements (e.g., diluted dataset 1) and the subset of dataset 2 measurements (e.g., diluted dataset 2) may, in turn, be used to compute to the KNN distance relative to the new inference data instance 702. In some aspects, the diluted dataset 1 and diluted dataset 2 may be retained in memory (e.g., memory 418 of FIGURE 4) for subsequent ML monitoring tasks.
[0124] Reducing the complexity of the KNN information calculation may come at the expense of the accuracy of the ML monitoring, which may motivate the network to configure the dilution operation based on its knowledge of the environment. The network may configure various levels of details in the dilution operation, for example, factors of dilution (e.g., dilution by 1 / 2 or 1 / 4 or 1 / 8 training samples, etc. ) or a number of training samples to keep for monitoring. The dilution may comprise, for example, a random, uniform dilution on a specific dimension, and may be applied before or after a transformation, for instance. The network may also configure diluted training samples (e.g., diluted dataset 1 or diluted dataset 2) that the UE (e.g., UE 120) may retain for calculating the KNN information.
[0125] In some aspects, the network may further configure the UE (e.g., UE 120) to cluster the training samples for the KNN information calculation. The network may configure the UE (e.g., UE 120) to retain a clustered version of the training samples for monitoring the ML model using the KNN information. In various aspects, the network may configure various level of details in the dilution operation. For example, the network may configure a number of clusters (e.g., 1, 2, 4, or 10) in every dataset. The clustering scheme may, for example, include a centroid-based clustering, a density-based clustering, a distribution-based clustering, or the like.
[0126] FIGURE 9 is a diagram 900 illustrating an example of ML model monitoring complexity reduction by clustering training samples, in accordance with aspects of the present disclosure. Referring to FIGURE 9, the diagram 900 includes the graph 700 illustrating the example of model monitoring using the KNN distance. Rather than considering the complete training dataset (e.g., dataset 1 or dataset 2 of the graph 700) , the dataset 1 and dataset 2 may be respectively represented as one or more clusters. For instance, as shown in the graph 902, the dataset 1 measurements may be represented as a centroid 904 and the dataset 2 measurements may be represented as a centroid 906. Then, a distance measure (e.g., KNN distance) may be computed using the centroid 904 and the centroid 906 relative to the new inference data instance 702. In some aspects, the clusters (e.g., 904, 906) may be retained in memory (e.g., the memory 418 of FIGURE 4) for subsequent ML monitoring tasks.
[0127] The UE (e.g., UE 120) may also be configured to report its capability to monitor the ML models using the KNN information (for example, KNN distance, KNN majority vote, or K-means distance) . The UE (e.g., UE 120) may report capability / recommendations for the K value to use in KNN information calculation. The UE (e.g., UE 120) may report the capability / recommendations on reducing the complexity of the monitoring using dimension reduction or training data dilution or training data clustering, for instance.
[0128] The UE (e.g., UE 120) may report recommendations for thresholds on the KNN information for switching the ML models. In some aspects, the UE capabilities may be reported using radio resource control (RRC) signaling. In other aspects, the UE recommendations may be reported using RRC signaling, a medium access control-control element (MAC-CE) , or uplink control information (UCI) . On the other hand, the network configurations may be static through RRC signaling, semi-static through MAC-CEs, or dynamically activated through the downlink control information (DCI) .
[0129] In some aspects, the UE (e.g., UE 120) may use a different ML model for ML model monitoring. For example, the UE (e.g., UE 120) may use a support vector machine (SVM) , a random forest, a decision tree, logistic regression, a neural network, and / or K-means for ML monitoring.
[0130] As indicated above, FIGURES 4-9 are provided as examples. Other examples may differ from what is described with respect to FIGURES 4-9.
[0131] FIGURE 10 is a flow diagram illustrating an example process 1000 for machine learning (ML) model monitoring performed by a UE, in accordance with various aspects of the present disclosure. The UE (e.g., UE 120) may include one or more processors such as the CPU 402, GPU 404 and / or NPU 408 for example, which may execute instructions associated with the ML monitoring module 140 to implement the example process 1000. The UE (e.g., UE 120) may also include one or more memory units such as memory 418, for instance.
[0132] At block 1002, the UE operates a first machine learning (ML) model of a plurality of first ML models. The first ML model is associated with wireless communication. For instance, the UE may select and operate one of a set of ML models for spatial beam estimation, or other wireless communication tasks.
[0133] At block 1004, the UE monitors the first ML model using a second ML model. The second ML model is configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances. As described with reference to FIGURE 7, the second ML model may monitor the ML model for beam estimation operated by the UE. For instance, the second ML model may monitor the ML model based on the proximity between a new inference data instance 702 and the training data (e.g., dataset 1 and dataset 2) .
[0134] In some aspects, at block 1006, the UE may optionally determine, based on the one or more distance measures, whether to select a different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to a non-machine learning model for the wireless communication. For instance, as described, with reference to FIGURE 7, the second ML model may use the proximity information (e.g., KNN distance) to monitor the ML model applicability, select a different ML model out of a group of ML models, fallback to non-ML operation, or determine whether to fine tune the ML model, for example.
[0135] FIGURE 11 is a flow diagram illustrating an example process 1100 for machine learning (ML) model monitoring performed by a network device, in accordance with various aspects of the present disclosure. The network device (e.g., base station 110) may include one or more processors such as the CPU 402, GPU 404 and / or NPU 408 for example, which may execute instructions associated with the ML monitoring module 138 to implement the example process 1100. The network device (e.g., base station 110) may also include one or more memory units such as memory 418, for instance.
[0136] At block 1102, the network device receives, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication. As described with respect to FIGURE 7, the network (e.g., via a network device such as base station 110) may receive the proximity information from the UE. In some examples, the network may configure the UE (e.g., UE 120) to report the statistical properties of the KNN distance or KNN majority vote over multiple inference data instances.
[0137] At block 1104, the network device uses a second ML model to monitor the first ML model based on the distance measure information. For instance, the network device may compare the received distance measure to a predefined threshold. In some aspects, the network device may determine that a different ML model should be used by the UE responsive to the distance measure exceeding the predefined threshold.
[0138] In some aspects, multiple different thresholds may be used to trigger different modifications. For instance, in response to the distance measure exceeding a first threshold, the second ML model may determine to fine tune the first ML model. In response to the distance measure exceeding a second threshold, the second ML model may determine to revert to a non-ML model for the wireless communication task (e.g., spatial beam estimation) .
[0139] At block 1106, the network device determining, based on the distance measure information, whether to select a different first ML model of a plurality of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model. The network device may determine a modification based on the monitoring. For instance, as described with reference to FIGURE 7, the network may determine whether to select a different ML model out of a group of ML models, to revert to a non-ML operation, or to fine tune the current ML model based on the distance measure. The network may transmit a message to the UE to configure the UE in accordance with the determination.
[0140] Implementation examples are provided in the following numbered clauses.
[0141] 1. A method of wireless communication performed by a user equipment (UE) , the method comprising:
[0142] operating, by the UE, a first machine learning (ML) model of a plurality of first ML models, the first ML model associated with wireless communication; and
[0143] monitoring, by the UE, the first ML model using a second ML model, the second ML model being configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.
[0144] 2. The method of clause 1, further comprising determining, by the UE, based on the one or more distance measures, whether to select a different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to a non-machine learning model for the wireless communication.
[0145] 3. The method of clause 1 or 2, in which the determination of whether to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model are respectively based on one or more configurable thresholds.
[0146] 4. The method of any of clauses 1-3, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.
[0147] 5. The method of any of clauses 1-4, further comprising receiving, from a network, a first message including one or more configuration parameters for the second ML model.
[0148] 6. The method of any of clauses 1-5, in which the one or more configuration parameters includes a value of the K integer parameter or the one or more configurable thresholds.
[0149] 7. The method of any of clauses 1-6, further comprising retaining a subset of the first dataset for training the first ML model as a diluted training dataset, the second ML model computing the one or more distance measures according to the diluted training dataset.
[0150] 8. The method of any of clauses 1-7, further comprising transmitting, to a network, a second message indicating a monitoring capability, the second message comprising a radio resource control (RRC) message, a medium access control-control element (MAC-CE) message, or uplink control information (UCI) .
[0151] 9. The method of any of clauses 1-8, further comprising performing a clustering process on the first dataset to reduce a complexity for computing the one or more distance measures.
[0152] 10. A method of wireless communication performed by a network device, the method comprising:
[0153] receiving, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication;
[0154] using a second ML model to monitor the first ML model based on the distance measure information; and
[0155] determining, based on the distance measure information, whether to select a different first ML model of a plurality of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0156] 11. The method of clause 10, further comprising:
[0157] receiving, from the UE, a first message indicating a UE capability for monitoring the first ML model associated with the wireless communication; and
[0158] transmitting, to the UE, a set of configuration parameters for the second ML model, the set of configuration parameters being used by the UE to configure the second ML model for monitoring the first ML model based on one or more distance measures between a first dataset for training the first ML model and one or more inference data instances generated by the first ML model.
[0159] 12. The method of clause 10 or 11, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.
[0160] 13. The method of any of clauses 10-12, in which the set of configuration parameters includes a value of the K integer parameter and one or more configurable threshold.
[0161] 14. The method of any of clauses 10-13, in which the set of configuration parameters includes a model complexity parameter, the model complexity parameter configuring the UE to use a diluted set of the first dataset or one or more clusters of the first dataset for computing the distance measure information.
[0162] 15. The method of any of clauses 10-14, further comprising transmitting, to the UE, a second message configuring the UE, based on the determining, to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model for the wireless communication.
[0163] 16. An apparatus for wireless communication performed by a user equipment (UE) , comprising:
[0164] at least one memory; and
[0165] at least one processor coupled to the at least one memory, the at least one processor configured to:
[0166] operate, by the UE, a first machine learning (ML) model of a plurality of first ML models, the first ML model associated with the wireless communication; and
[0167] monitor, by the UE, the first ML model using a second ML model, the second ML model being configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.
[0168] 17. The apparatus of clause 16, in which the at least one processor is further configured to determine, by the UE, based on the one or more distance measures, whether to select a different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to a non-machine learning model for the wireless communication.
[0169] 18. The apparatus of clause 16 or 17, in which the determination of whether to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model is respectively based on one or more configurable thresholds.
[0170] 19. The apparatus of any of clauses 16-18, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.
[0171] 20. The apparatus of any of clauses 16-19, in which the at least one processor is further configured to receive, from a network, a first message including one or more configuration parameters for the second ML model.
[0172] 21. The apparatus of any of clauses 16-20, in which the one or more configuration parameters includes a value of the K integer parameter or the one or more configurable thresholds.
[0173] 22. The apparatus of any of clauses 16-21, in which the at least one processor is further configured to retain a subset of the first dataset for training the first ML model as a diluted training dataset, the second ML model computing the one or more distance measures according to the diluted training dataset.
[0174] 23. The apparatus of any of clauses 16-22 in which the at least one processor is further configured to transmit, to a network, a second message indicating a monitoring capability, the second message comprising a radio resource control (RRC) message, a medium access control-control element (MAC-CE) message, or uplink control information (UCI) .
[0175] 24. The apparatus of any of clauses 16-23, in which the at least one processor is further configured to perform a clustering process on the first dataset to reduce a complexity for computing the one or more distance measures.
[0176] 25. An apparatus of wireless communication performed by a network device, comprising:
[0177] at least one memory; and
[0178] at least one processor coupled to the at least one memory, the at least one processor configured to:
[0179] receive, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication;
[0180] use a second ML model to monitor the first ML model based on the distance measure information; and
[0181] determine, based on the distance measure information, whether to select a different first ML model of a plurality of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.
[0182] 26. The apparatus of clause 25, in which the at least one processor is further configured to:
[0183] receive, from the UE, a first message indicating a UE capability for monitoring the first ML model associated with the wireless communication; and
[0184] transmit, to the UE, a set of configuration parameters for the second ML model, the set of configuration parameters being used by the UE to configure the second ML model for monitoring the first ML model based on one or more distance measures between a first dataset for training the first ML model and one or more inference data instances generated by the first ML model.
[0185] 27. The apparatus of clause 25 or 26, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.
[0186] 28. The method of any of clauses 25-27, in which the set of configuration parameters includes a value of the K integer parameter and one or more configurable threshold.
[0187] 29. The method of any of clauses 25-28, in which the set of configuration parameters includes a model complexity parameter, the model complexity parameter configuring the UE to use a diluted set of the first dataset or one or more clusters of the first dataset for computing the distance measure information.
[0188] 30. The apparatus of any of clauses 25-29, in which the at least one processor is further configured to transmit, to the UE, a second message configuring the UE, based on the determining, to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model for the wireless communication.
[0189] The foregoing disclosure provides illustration and description, but is not intended to be exhaustive or to limit the aspects to the precise form disclosed. Modifications and variations may be made in light of the above disclosure or may be acquired from practice of the aspects.
[0190] As used, the term “component” is intended to be broadly construed as hardware, firmware, and / or a combination of hardware and software. As used, a processor is implemented in hardware, firmware, and / or a combination of hardware and software.
[0191] Some aspects are described in connection with thresholds. As used, satisfying a threshold may, depending on the context, refer to a value being greater than the threshold, greater than or equal to the threshold, less than the threshold, less than or equal to the threshold, equal to the threshold, not equal to the threshold, and / or the like.
[0192] It will be apparent that systems and / or methods described may be implemented in different forms of hardware, firmware, and / or a combination of hardware and software. The actual specialized control hardware or software code used to implement these systems and / or methods is not limiting of the aspects. Thus, the operation and behavior of the systems and / or methods were described without reference to specific software code-it being understood that software and hardware can be designed to implement the systems and / or methods based, at least in part, on the description.
[0193] Even though particular combinations of features are recited in the claims and / or disclosed in the specification, these combinations are not intended to limit the disclosure of various aspects. In fact, many of these features may be combined in ways not specifically recited in the claims and / or disclosed in the specification. Although each dependent claim listed below may directly depend on only one claim, the disclosure of various aspects includes each dependent claim in combination with every other claim in the claim set. A phrase referring to “at least one of” a list of items refers to any combination of those items, including single members. As an example, “at least one of: a, b, or c” is intended to cover a, b, c, a-b, a-c, b-c, and a-b-c, as well as any combination with multiples of the same element (e.g., a-a, a-a-a, a-a-b, a-a-c, a-b-b, a-c-c, b-b, b-b-b, b-b-c, c-c, and c-c-c or any other ordering of a, b, and c) .
[0194] No element, act, or instruction used should be construed as critical or essential unless explicitly described as such. Also, as used, the articles “a” and “an” are intended to include one or more items, and may be used interchangeably with “one or more. ” Furthermore, as used, the terms “set” and “group” are intended to include one or more items (e.g., related items, unrelated items, a combination of related and unrelated items, and / or the like) , and may be used interchangeably with “one or more. ” Where only one item is intended, the phrase “only one” or similar language is used. Also, as used, the terms “has, ” “have, ” “having, ” and / or the like are intended to be open-ended terms. Further, the phrase “based on” is intended to mean “based, at least in part, on” unless explicitly stated otherwise.
Claims
1.A method of wireless communication performed by a user equipment (UE) , the method comprising:operating, by the UE, a first machine learning (ML) model of a plurality of first ML models, the first ML model associated with wireless communication; andmonitoring, by the UE, the first ML model using a second ML model, the second ML model being configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.2.The method of claim 1, further comprising determining, by the UE, based on the one or more distance measures, whether to select a different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to a non-machine learning model for the wireless communication.3.The method of claim 2, in which the determination of whether to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model are respectively based on one or more configurable thresholds.4.The method of claim 3, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.5.The method of claim 4, further comprising receiving, from a network, a first message including one or more configuration parameters for the second ML model.6.The method of claim 5, in which the one or more configuration parameters includes a value of the K integer parameter or the one or more configurable thresholds.7.The method of claim 1, further comprising retaining a subset of the first dataset for training the first ML model as a diluted training dataset, the second ML model computing the one or more distance measures according to the diluted training dataset.8.The method of claim 1, further comprising transmitting, to a network, a second message indicating a monitoring capability, the second message comprising a radio resource control (RRC) message, a medium access control-control element (MAC-CE) message, or uplink control information (UCI) .9.The method of claim 1, further comprising performing a clustering process on the first dataset to reduce a complexity for computing the one or more distance measures.10.A method of wireless communication performed by a network device, the method comprising:receiving, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication;using a second ML model to monitor the first ML model based on the distance measure information; anddetermining, based on the distance measure information, whether to select a different first ML model of a plurality of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.11.The method of claim 10, further comprising:receiving, from the UE, a first message indicating a UE capability for monitoring the first ML model associated with the wireless communication; andtransmitting, to the UE, a set of configuration parameters for the second ML model, the set of configuration parameters being used by the UE to configure the second ML model for monitoring the first ML model based on one or more distance measures between a first dataset for training the first ML model and one or more inference data instances generated by the first ML model.12.The method of claim 11, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.13.The method of claim 12, in which the set of configuration parameters includes a value of the K integer parameter and one or more configurable threshold.14.The method of claim 11, in which the set of configuration parameters includes a model complexity parameter, the model complexity parameter configuring the UE to use a diluted set of the first dataset or one or more clusters of the first dataset for computing the distance measure information.15.The method of claim 10, further comprising transmitting, to the UE, a second message configuring the UE, based on the determining, to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model for the wireless communication.16.An apparatus for wireless communication performed by a user equipment (UE) , comprising:at least one memory; andat least one processor coupled to the at least one memory, the at least one processor configured to:operate, by the UE, a first machine learning (ML) model of a plurality of first ML models, the first ML model associated with the wireless communication; andmonitor, by the UE, the first ML model using a second ML model, the second ML model being configured to monitor the first ML model based on one or more distance measures between at least a first dataset for training the first ML model and one or more inference data instances observed during the operation of the first ML model.17.The apparatus of claim 16, in which the at least one processor is further configured to determine, by the UE, based on the one or more distance measures, whether to select a different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to a non-machine learning model for the wireless communication.18.The apparatus of claim 17, in which the determination of whether to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model is respectively based on one or more configurable thresholds.19.The apparatus of claim 18, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.20.The apparatus of claim 19, in which the at least one processor is further configured to receive, from a network, a first message including one or more configuration parameters for the second ML model.21.The apparatus of claim 20, in which the one or more configuration parameters includes a value of the K integer parameter or the one or more configurable thresholds.22.The apparatus of claim 16, in which the at least one processor is further configured to retain a subset of the first dataset for training the first ML model as a diluted training dataset, the second ML model computing the one or more distance measures according to the diluted training dataset.23.The apparatus of claim 16, in which the at least one processor is further configured to transmit, to a network, a second message indicating a monitoring capability, the second message comprising a radio resource control (RRC) message, a medium access control-control element (MAC-CE) message, or uplink control information (UCI) .24.The apparatus of claim 16, in which the at least one processor is further configured to perform a clustering process on the first dataset to reduce a complexity for computing the one or more distance measures.25.An apparatus of wireless communication performed by a network device, comprising:at least one memory; andat least one processor coupled to the at least one memory, the at least one processor configured to:receive, from a user equipment (UE) , distance measure information related to a first machine learning (ML) model associated with wireless communication;use a second ML model to monitor the first ML model based on the distance measure information; anddetermine, based on the distance measure information, whether to select a different first ML model of a plurality of first ML models for the UE to operate, to fine tune the first ML model, or to revert to a non-machine learning model.26.The apparatus of claim 25, in which the at least one processor is further configured to:receive, from the UE, a first message indicating a UE capability for monitoring the first ML model associated with the wireless communication; andtransmit, to the UE, a set of configuration parameters for the second ML model, the set of configuration parameters being used by the UE to configure the second ML model for monitoring the first ML model based on one or more distance measures between a first dataset for training the first ML model and one or more inference data instances generated by the first ML model.27.The apparatus of claim 26, in which the one or more distance measures comprise a nearest neighbor distance, a K-nearest neighbors (KNN) distance or a K-means distance, where K is an integer parameter.28.The apparatus of claim 27, in which the set of configuration parameters includes a value of the K integer parameter and one or more configurable threshold.29.The apparatus of claim 26, in which the set of configuration parameters includes a model complexity parameter, the model complexity parameter configuring the UE to use a diluted set of the first dataset or one or more clusters of the first dataset for computing the distance measure information.30.The apparatus of claim 25, in which the at least one processor is further configured to transmit, to the UE, a second message configuring the UE, based on the determining, to select the different first ML model of the plurality of first ML models to operate, to fine tune the first ML model, or to revert to the non-machine learning model for the wireless communication.