Reporting user equipment artificial intelligence or machine learning model representation or quantization preferences
By allowing UEs to report their AI/ML model quantization preferences to network nodes, the system optimizes AI/ML model performance in wireless communication systems by aligning model representation and quantization with the UE's hardware capabilities.
Patent Information
- Application Number
- PCT/CN2023/119598
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-09-19
- Publication Date
- 2025-03-27
AI Technical Summary
Existing wireless communication systems do not consider mechanisms for allowing network nodes to be aware of user equipment (UE) side artificial intelligence (AI) or machine learning (ML) model quantization preferences, leading to suboptimal AI/ML model performance due to mismatched quantization schemes.
The system enables UE to transmit indications of its AI/ML model representation or quantization preferences to the network node, allowing the network node to adjust the representation or quantization of AI/ML models based on these preferences, thereby optimizing the models for the UE's capabilities.
This approach allows for AI/ML models that are specifically designed based on the UE's capabilities, resulting in improved performance and efficiency in wireless communication by tailoring the models to the implementation limitations of the UE's hardware-based AI/ML accelerators.
Smart Images

Figure CN2023119598_27032025_PF_FP_ABST
Abstract
Description
REPORTING USER EQUIPMENT ARTIFICIAL INTELLIGENCE OR MACHINE LEARNING MODEL REPRESENTATION OR QUANTIZATION PREFERENCES
[0001] FIELD OF THE DISCLOSURE
[0002] Aspects of the present disclosure generally relate to wireless communication and to techniques and apparatuses for reporting user equipment (UE) artificial intelligence or machine learning (AI / ML) model representation or quantization preferences.BACKGROUND
[0003] Wireless communication systems are widely deployed to provide various telecommunication services such as telephony, video, data, messaging, and broadcasts. Typical wireless communication systems may employ multiple-access technologies capable of supporting communication with multiple users by sharing available system resources (e.g., bandwidth, transmit power, 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) .
[0004] A wireless network may include one or more network nodes that support communication for wireless communication devices, such as a user equipment (UE) or multiple UEs. A UE may communicate with a network node via downlink communications and uplink communications. “Downlink” (or “DL” ) refers to a communication link from the network node to the UE, and “uplink” (or “UL” ) refers to a communication link from the UE to the network node. Some wireless networks may support device-to-device communication, such as via a local link (e.g., a sidelink (SL) , a wireless local area network (WLAN) link, and / or a wireless personal area network (WPAN) link, among other examples) .
[0005] The above multiple access technologies have been adopted in various telecommunication standards to provide a common protocol that enables different UEs to communicate on a municipal, national, regional, and / or global level. New Radio (NR) , which may be referred to as 5G, is a set of enhancements to the LTE mobile standard promulgated by the 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, using CP-OFDM and / or single-carrier frequency division multiplexing (SC-FDM) (also known as discrete Fourier transform spread OFDM (DFT-s-OFDM) ) on the uplink, as well as supporting beamforming, multiple-input multiple-output (MIMO) antenna technology, and carrier aggregation. As the demand for mobile broadband access continues to increase, further improvements in LTE, NR, and other radio access technologies remain useful.SUMMARY
[0006] In some implementations, an apparatus for wireless communication at a user equipment (UE) includes one or more memories; and one or more processors, coupled to the one or more memories, which, individually or in any combination, are operable to cause the apparatus to: transmit an indication of UE artificial learning or machine learning (AI / ML) model representation or quantization preferences; and receive an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0007] In some implementations, an apparatus for wireless communication at a network node includes one or more memories; and one or more processors, coupled to the one or more memories, which, individually or in any combination, are operable to cause the apparatus to: receive an indication of UE AI / ML model representation or quantization preferences; and transmit an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0008] In some implementations, a method of wireless communication performed by a UE includes transmitting an indication of UE AI / ML model representation or quantization preferences; and receiving an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0009] In some implementations, a method of wireless communication performed by a network node includes receiving an indication of UE AI / ML model representation or quantization preferences; and transmitting an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0010] In some implementations, a non-transitory computer-readable medium storing a set of instructions for wireless communication includes one or more instructions that, when executed by one or more processors of a UE, cause the UE to: transmit an indication of UE AI / ML model representation or quantization preferences; and receive an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0011] In some implementations, a non-transitory computer-readable medium storing a set of instructions for wireless communication includes one or more instructions that, when executed by one or more processors of a network node, cause the network node to: receive an indication of UE AI / ML model representation or quantization preferences; and transmit an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0012] In some implementations, an apparatus for wireless communication includes means for transmitting an indication of UE AI / ML model representation or quantization preferences; and means for receiving an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0013] In some implementations, an apparatus for wireless communication includes means for receiving an indication of UE AI / ML model representation or quantization preferences; and means for transmitting an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0014] Aspects generally include a method, apparatus, system, computer program product, non-transitory computer-readable medium, user equipment, base station, network entity, network node, wireless communication device, and / or processing system as substantially described herein with reference to and as illustrated by the drawings and specification.
[0015] 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 hereinafter. 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 herein, 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.
[0016] While aspects are described in the present disclosure by illustration to some examples, those skilled in the art will understand that such aspects may be implemented in many different arrangements and scenarios. Techniques described herein may be implemented using different platform types, devices, systems, shapes, sizes, and / or packaging arrangements. For example, some aspects may be implemented via integrated chip embodiments or other non-module-component based devices (e.g., end-user devices, vehicles, communication devices, computing devices, industrial equipment, retail / purchasing devices, medical devices, and / or artificial intelligence devices) . Aspects may be implemented in chip-level components, modular components, non-modular components, non-chip-level components, device-level components, and / or system-level components. Devices incorporating described aspects and features may include additional components and features for implementation and practice of claimed and described aspects. For example, transmission and reception of wireless signals may include one or more components for analog and digital purposes (e.g., hardware components including antennas, radio frequency (RF) chains, power amplifiers, modulators, buffers, processors, interleavers, adders, and / or summers) . It is intended that aspects described herein may be practiced in a wide variety of devices, components, systems, distributed arrangements, and / or end-user devices of varying size, shape, and constitution.BRIEF DESCRIPTION OF THE DRAWINGS
[0017] So that the above-recited features of the present disclosure can be understood in detail, a more particular description, briefly summarized above, 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 typical 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.
[0018] Fig. 1 is a diagram illustrating an example of a wireless network, in accordance with the present disclosure.
[0019] Fig. 2 is a diagram illustrating an example of a network node in communication with a user equipment (UE) in a wireless network, in accordance with the present disclosure.
[0020] Fig. 3 is a diagram illustrating an example disaggregated base station architecture, in accordance with the present disclosure.
[0021] Fig. 4 is a diagram illustrating an example of beam management, in accordance with the present disclosure.
[0022] Fig. 5 is a diagram illustrating examples of artificial intelligence (AI) and / or machine learning (ML) (AI / ML) models for beam prediction, in accordance with the present disclosure.
[0023] Figs. 6-9 are diagrams illustrating examples associated with reporting UE AI / ML model representation or quantization preferences, in accordance with the present disclosure.
[0024] Figs. 10-11 are diagrams illustrating example processes associated with reporting UE AI / ML model representation or quantization preferences, in accordance with the present disclosure.
[0025] Figs. 12-13 are diagrams of example apparatuses for wireless communication, in accordance with the present disclosure.DETAILED DESCRIPTION
[0026] A network node may train an artificial intelligence (AI) and / or machine learning (ML) (AI / ML) model for beam prediction based at least in part on data collected via user equipment (UE) reported layer 1 (L1) reference signal received power (RSRP) (L1-RSRP) measurements. The UE reported L1-RSRP measurements may be historical measurements for certain areas, scenarios, and / or codebook configurations. When a certain UE enters such an area, scenario, or codebook configuration, the network node may transfer the AI / ML model to that UE. The network node may transfer the AI / ML model to the UE via an AI / ML air interface. The UE may locally run a model inference using the AI / ML model received from the network node. In this case, the network node may train the AI / ML model, as opposed to the UE autonomously training an AI / ML model. The network node training the AI / ML model, as opposed to the UE training the AI / ML model, may be beneficial because the network node may have a better understanding on codebooks, antenna setups, tilting directions, and / or parent-child relationships of its beams as compared to the UE.
[0027] An AI / ML model compilation, an AI / ML model representation, and / or an AI / ML model quantization may be carried out by the network node or the UE. When labeled training data is used for AI / ML model quantization, the AI / ML model quantization may occur at the network node. The use of labeled training data may result in improved AI / ML model inference accuracy. Otherwise, the UE may quantize the AI / ML model by itself, for example without labeled training data (due to fewer resources at the UE as compared to the network node) , which may lead to a degraded AI / ML model inference accuracy.
[0028] Different UEs may have various preferences regarding the quantization of AI / ML models. Such various quantization preferences (e.g., number of bits per parameter, quantization metrics, and / or step-sizes) may be due to implementation limitations of hardware based AI / ML accelerators at the different UEs. Depending on available computational resources, different components / layers associated with the same AI / ML model may also be preferred with different quantization granularities.
[0029] For AI / ML model transfer from the network node to the UE, a UE preference on AI / ML quantization may not be known to the network node. AI / ML model quantization may depend on training / inference data, so AI / ML model quantization may be more reasonable to be carried out by the network node, as opposed to the UE, when the AI / ML model is trained by the network node and transferred to the UE. A quantization dynamic range in respective layers may depend on an original model and training / inference data, but such data may not be available at the UE, so the AI / ML model quantization may be carried out by the network node instead of the UE. However, existing designs on AI / ML model transfer from the network node to the UE do not consider mechanisms for allowing the network node to be aware of UE-side AI / ML model quantization preferences.
[0030] Various aspects relate generally to reporting UE AI / ML model representation or quantization preferences. Some aspects more specifically relate to transferring an AI / ML model from a network node to a UE based at least in part on the UE AI / ML model representation or quantization preferences. In some examples, a UE may transmit, to a network node, an indication of UE AI / ML model representation or quantization preferences. The UE AI / ML model representation or quantization preferences may indicate a number of bits to represent or quantize a single parameter, metrics associated with determining an AI / ML model parameter representation or quantization, an indication of whether the UE prefers a network node to represent or quantize an AI / ML model or whether the UE prefers to represent or quantize the AI / ML model, an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, and / or components or layer specific AI / ML model parameter representation or quantization preferences. The UE may receive, from the network node, an AI / ML model. A representation or quantization of one or more parameters associated with the AI / ML model may be based at least in part on the UE AI / ML model representation or quantization preferences. The AI / ML model may be associated with a network node side AI / ML model parameter representation or quantization, or alternatively, the AI / ML model may be associated with a UE side AI / ML model parameter representation or quantization, based at least in part on the UE AI / ML model representation or quantization preferences. The UE may perform a model inference using the AI / ML model.
[0031] 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, by reporting UE AI / ML model representation or quantization preferences, the described techniques can be used to transfer AI / ML models that are based at least in part on the UE AI / ML model representation or quantization preferences. The network node, when quantizing the AI / ML model, may consider UE preferences, which may vary depending on implementation limitations of hardware based AI / ML accelerators at the UE. Thus, the AI / ML model may be specifically designed based on a UE capability, as indicated by the UE preferences, which may result in AI / ML models that are better suited to run at the UE. Such AI / ML models may improve an overall performance at the UE because the AI / ML models may be tailored to the implementation limitations of hardware based AI / ML accelerators at the UE.
[0032] Various aspects of the disclosure are described more fully hereinafter 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. One skilled in the art should appreciate that the scope of the disclosure is intended to cover any aspect of the disclosure disclosed herein, 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 herein. 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 herein. It should be understood that any aspect of the disclosure disclosed herein may be embodied by one or more elements of a claim.
[0033] Several aspects of telecommunication 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, 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] While aspects may be described herein using terminology commonly associated with a 5G or New Radio (NR) radio access technology (RAT) , aspects of the present disclosure can be applied to other RATs, such as a 3G RAT, a 4G RAT, and / or a RAT subsequent to 5G (e.g., 6G) .
[0035] Fig. 1 is a diagram illustrating an example of a wireless network 100, in accordance with the present disclosure. The wireless network 100 may be or may include elements of a 5G (e.g., NR) network and / or a 4G (e.g., Long Term Evolution (LTE) ) network, among other examples. The wireless network 100 may include one or more network nodes 110 (shown as a network node 110a, a network node 110b, a network node 110c, and a network node 110d) , a UE 120 or multiple UEs 120 (shown as a UE 120a, a UE 120b, a UE 120c, a UE 120d, and a UE 120e) , and / or other entities. A network node 110 is a network node that communicates with UEs 120. As shown, a network node 110 may include one or more network nodes. For example, a network node 110 may be an aggregated network node, meaning that the aggregated network node is configured to utilize a radio protocol stack that is physically or logically integrated within a single radio access network (RAN) node (e.g., within a single device or unit) . As another example, a network node 110 may be a disaggregated network node (sometimes referred to as a disaggregated base station) , meaning that the network node 110 is configured to utilize a protocol stack that is physically or logically distributed among two or more nodes (such as one or more central units (CUs) , one or more distributed units (DUs) , or one or more radio units (RUs) ) .
[0036] In some examples, a network node 110 is or includes a network node that communicates with UEs 120 via a radio access link, such as an RU. In some examples, a network node 110 is or includes a network node that communicates with other network nodes 110 via a fronthaul link or a midhaul link, such as a DU. In some examples, a network node 110 is or includes a network node that communicates with other network nodes 110 via a midhaul link or a core network via a backhaul link, such as a CU. In some examples, a network node 110 (such as an aggregated network node 110 or a disaggregated network node 110) may include multiple network nodes, such as one or more RUs, one or more CUs, and / or one or more DUs. A network node 110 may include, for example, an NR base station, an LTE base station, a Node B, an eNB (e.g., in 4G) , a gNB (e.g., in 5G) , an access point, a transmission reception point (TRP) , a DU, an RU, a CU, a mobility element of a network, a core network node, a network element, a network equipment, a RAN node, or a combination thereof. In some examples, the network nodes 110 may be interconnected to one another or to one or more other network nodes 110 in the wireless network 100 through various types of fronthaul, midhaul, and / or backhaul interfaces, such as a direct physical connection, an air interface, or a virtual network, using any suitable transport network.
[0037] In some examples, a network node 110 may provide communication coverage for a particular geographic area. In the Third Generation Partnership Project (3GPP) , the term “cell” can refer to a coverage area of a network node 110 and / or a network node subsystem serving this coverage area, depending on the context in which the term is used. A network node 110 may provide communication coverage for a macro cell, a pico cell, a femto cell, and / or another type of cell. A macro cell may cover a relatively large geographic area (e.g., several kilometers in radius) and may allow unrestricted access by UEs 120 with service subscriptions. A pico cell may cover a relatively small geographic area and may allow unrestricted access by UEs 120 with service subscriptions. A femto cell may cover a relatively small geographic area (e.g., a home) and may allow restricted access by UEs 120 having association with the femto cell (e.g., UEs 120 in a closed subscriber group (CSG) ) . A network node 110 for a macro cell may be referred to as a macro network node. A network node 110 for a pico cell may be referred to as a pico network node. A network node 110 for a femto cell may be referred to as a femto network node or an in-home network node. In the example shown in Fig. 1, the network node 110a may be a macro network node for a macro cell 102a, the network node 110b may be a pico network node for a pico cell 102b, and the network node 110c may be a femto network node for a femto cell 102c. A network node may support one or multiple (e.g., three) cells. In some examples, a cell may not necessarily be stationary, and the geographic area of the cell may move according to the location of a network node 110 that is mobile (e.g., a mobile network node) .
[0038] In some aspects, the terms “base station” or “network node” may refer to an aggregated base station, a disaggregated base station, an integrated access and backhaul (IAB) node, a relay node, or one or more components thereof. For example, in some aspects, “base station” or “network node” may refer to a CU, a DU, an RU, a Near-Real Time (Near-RT) RAN Intelligent Controller (RIC) , or a Non-Real Time (Non-RT) RIC, or a combination thereof. In some aspects, the terms “base station” or “network node” may refer to one device configured to perform one or more functions, such as those described herein in connection with the network node 110. In some aspects, the terms “base station” or “network node” may refer to a plurality of devices configured to perform the one or more functions. For example, in some distributed systems, each of a quantity of different devices (which may be located in the same geographic location or in different geographic locations) may be configured to perform at least a portion of a function, or to duplicate performance of at least a portion of the function, and the terms “base station” or “network node” may refer to any one or more of those different devices. In some aspects, the terms “base station” or “network node” may refer to one or more virtual base stations or one or more virtual base station functions. For example, in some aspects, two or more base station functions may be instantiated on a single device. In some aspects, the terms “base station” or “network node” may refer to one of the base station functions and not another. In this way, a single device may include more than one base station.
[0039] The wireless network 100 may include one or more relay stations. A relay station is a network node that can receive a transmission of data from an upstream node (e.g., a network node 110 or a UE 120) and send a transmission of the data to a downstream node (e.g., a UE 120 or a network node 110) . A relay station may be a UE 120 that can relay transmissions for other UEs 120. In the example shown in Fig. 1, the network node 110d (e.g., a relay network node) may communicate with the network node 110a (e.g., a macro network node) and the UE 120d in order to facilitate communication between the network node 110a and the UE 120d. A network node 110 that relays communications may be referred to as a relay station, a relay base station, a relay network node, a relay node, a relay, or the like.
[0040] The wireless network 100 may be a heterogeneous network that includes network nodes 110 of different types, such as macro network nodes, pico network nodes, femto network nodes, relay network nodes, or the like. These different types of network nodes 110 may have different transmit power levels, different coverage areas, and / or different impacts on interference in the wireless network 100. For example, macro network nodes may have a high transmit power level (e.g., 5 to 40 watts) whereas pico network nodes, femto network nodes, and relay network nodes may have lower transmit power levels (e.g., 0.1 to 2 watts) .
[0041] A network controller 130 may couple to or communicate with a set of network nodes 110 and may provide coordination and control for these network nodes 110. The network controller 130 may communicate with the network nodes 110 via a backhaul communication link or a midhaul communication link. The network nodes 110 may communicate with one another directly or indirectly via a wireless or wireline backhaul communication link. In some aspects, the network controller 130 may be a CU or a core network device, or may include a CU or a core network device.
[0042] The UEs 120 may be dispersed throughout the wireless network 100, and each UE 120 may be stationary or mobile. A UE 120 may include, for example, an access terminal, a terminal, a mobile station, and / or a subscriber unit. A UE 120 may be a cellular phone (e.g., a smart phone) , a personal digital assistant (PDA) , a wireless modem, a wireless communication device, a handheld device, a laptop computer, a cordless phone, a wireless local loop (WLL) station, a tablet, a camera, a gaming device, a netbook, a smartbook, an ultrabook, a medical device, a biometric device, a wearable device (e.g., a smart watch, smart clothing, smart glasses, a smart wristband, smart jewelry (e.g., a smart ring or a smart bracelet) ) , an entertainment device (e.g., a music device, a video device, and / or a satellite radio) , a vehicular component or sensor, a smart meter / sensor, industrial manufacturing equipment, a global positioning system device, a UE function of a network node, and / or any other suitable device that is configured to communicate via a wireless or wired medium.
[0043] Some UEs 120 may be considered machine-type communication (MTC) or evolved or enhanced machine-type communication (eMTC) UEs. An MTC UE and / or an eMTC UE may include, for example, a robot, an unmanned aerial vehicle, a remote device, a sensor, a meter, a monitor, and / or a location tag, that may communicate with a network node, another device (e.g., a remote device) , or some other entity. Some UEs 120 may be considered Intemet-of-Things (IoT) devices, and / or may be implemented as NB-IoT (narrowband IoT) devices. Some UEs 120 may be considered a Customer Premises Equipment. A UE 120 may be included inside a housing that houses components of the UE 120, such as processor components and / or memory components. In some examples, the processor components and the memory components may be coupled together. For example, the processor components (e.g., one or more processors) and the memory components (e.g., a memory) may be operatively coupled, communicatively coupled, electronically coupled, and / or electrically coupled.
[0044] In general, any number of wireless networks 100 may be deployed in a given geographic area. Each wireless network 100 may support a particular RAT and may operate on one or more frequencies. A RAT may be referred to as a radio technology, an air interface, or the like. A frequency may be referred to as a carrier, a frequency channel, 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.
[0045] In some examples, 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 network node 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, or a vehicle-to-pedestrian (V2P) protocol) , and / or a mesh network. In such examples, a UE 120 may perform scheduling operations, resource selection operations, and / or other operations described elsewhere herein as being performed by the network node 110.
[0046] Devices of the wireless network 100 may communicate using the electromagnetic spectrum, which may be subdivided by frequency or wavelength into various classes, bands, channels, or the like. For example, devices of the wireless network 100 may communicate using one or more operating bands. In 5G NR, two initial operating bands have been identified as frequency range designations FR1 (410 MHz -7.125 GHz) and FR2 (24.25 GHz -52.6 GHz) . It should be understood that although a portion of FR1 is greater than 6 GHz, FR1 is often referred to (interchangeably) as a “Sub-6 GHz” band in various documents and articles. A similar nomenclature issue sometimes occurs with regard to FR2, which is often referred to (interchangeably) as a “millimeter wave” band in documents and articles, despite being different from the extremely high frequency (EHF) band (30 GHz -300 GHz) which is identified by the International Telecommunications Union (ITU) as a “millimeter wave” band.
[0047] The frequencies between FR1 and FR2 are often referred to as mid-band frequencies. Recent 5G NR studies have identified an operating band for these mid-band frequencies as frequency range designation FR3 (7.125 GHz -24.25 GHz) . Frequency bands falling within FR3 may inherit FR1 characteristics and / or FR2 characteristics, and thus may effectively extend features of FR1 and / or FR2 into mid-band frequencies. In addition, higher frequency bands are currently being explored to extend 5G NR operation beyond 52.6 GHz. For example, three higher operating bands have been identified as frequency range designations FR4a or FR4-1 (52.6 GHz -71 GHz) , FR4 (52.6 GHz -114.25 GHz) , and FR5 (114.25 GHz -300 GHz) . Each of these higher frequency bands falls within the EHF band.
[0048] With the above examples in mind, unless specifically stated otherwise, it should be understood that the term “sub-6 GHz” or the like, ifused herein, may broadly represent frequencies that may be less than 6 GHz, may be within FR1, or may include mid-band frequencies. Further, unless specifically stated otherwise, it should be understood that the term “millimeter wave” or the like, ifused herein, may broadly represent frequencies that may include mid-band frequencies, may be within FR2, FR4, FR4-a or FR4-1, and / or FR5, or may be within the EHF band. It is contemplated that the frequencies included in these operating bands (e.g., FR1, FR2, FR3, FR4, FR4-a, FR4-1, and / or FR5) may be modified, and techniques described herein are applicable to those modified frequency ranges.
[0049] In some aspects, a UE (e.g., the UE 120) may include a communication manager 140. As described in more detail elsewhere herein, the communication manager 140 may transmit an indication ofUE AI / ML model representation or quantization preferences; and receive an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences. Additionally, or alternatively, the communication manager 140 may perform one or more other operations described herein.
[0050] In some aspects, a network node (e.g., the network node 110) may include a communication manager 150. As described in more detail elsewhere herein, the communication manager 150 may receive an indication ofUE AI / ML model representation or quantization preferences; and transmit an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences. Additionally, or alternatively, the communication manager 150 may perform one or more other operations described herein.
[0051] As indicated above, Fig. 1 is provided as an example. Other examples may differ from what is described with regard to Fig. 1.
[0052] Fig. 2 is a diagram illustrating an example 200 of a network node 110 in communication with a UE 120 in a wireless network 100, in accordance with the present disclosure. The network node 110 may be equipped with a set of antennas 234a through 234t, such as T antennas (T≥ 1) . The UE 120 may be equipped with a set of antennas 252a through 252r, such as R antennas (R ≥ 1) . The network node 110 of example 200 includes one or more radio frequency components, such as antennas 234 and a modem 232. In some examples, a network node 110 may include an interface, a communication component, or another component that facilitates communication with the UE 120 or another network node. Some network nodes 110 may not include radio frequency components that facilitate direct communication with the UE 120, such as one or more CUs, or one or more DUs.
[0053] At the network node 110, a transmit processor 220 may receive data, from a data source 212, intended for the UE 120 (or a set of UEs 120) . The transmit processor 220 may select one or more modulation and coding schemes (MCSs) for the UE 120 based at least in part on one or more channel quality indicators (CQIs) received from that UE 120. The network node 110 may process (e.g., encode and modulate) the data for the UE 120 based at least in part on the MCS (s) selected for the UE 120 and may provide data symbols for the UE 120. The transmit processor 220 may process system information (e.g., for semi-static resource partitioning information (SRPI) ) and control information (e.g., CQI requests, grants, and / or upper layer signaling) and provide overhead symbols and control symbols. The transmit processor 220 may generate reference symbols for reference signals (e.g., a cell-specific reference signal (CRS) or a demodulation reference signal (DMRS) ) and synchronization signals (e.g., a primary synchronization signal (PSS) or a 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 a set of output symbol streams (e.g., T output symbol streams) to a corresponding set of modems 232 (e.g., Tmodems) , shown as modems 232a through 232t. For example, each output symbol stream may be provided to a modulator component (shown as MOD) of a modem 232. Each modem 232 may use a respective modulator component to process a respective output symbol stream (e.g., for OFDM) to obtain an output sample stream. Each modem 232 may further use a respective modulator component to process (e.g., convert to analog, amplify, filter, and / or upconvert) the output sample stream to obtain a downlink signal. The modems 232a through 232t may transmit a set of downlink signals (e.g., T downlink signals) via a corresponding set of antennas 234 (e.g., T antennas) , shown as antennas 234a through 234t.
[0054] At the UE 120, a set of antennas 252 (shown as antennas 252a through 252r) may receive the downlink signals from the network node 110 and / or other network nodes 110 and may provide a set of received signals (e.g., R received signals) to a set of modems 254 (e.g., R modems) , shown as modems 254a through 254r. For example, each received signal may be provided to a demodulator component (shown as DEMOD) of a modem 254. Each modem 254 may use a respective demodulator component to condition (e.g., filter, amplify, downconvert, and / or digitize) a received signal to obtain input samples. Each modem 254 may use a demodulator component to further process the input samples (e.g., for OFDM) to obtain received symbols. A MIMO detector 256 may obtain received symbols from the modems 254, may perform MIMO detection on the received symbols if applicable, and may provide detected symbols. A receive processor 258 may process (e.g., demodulate and decode) the detected symbols, may provide decoded data for the UE 120 to a data sink 260, and may provide decoded control information and system information to a controller / processor 280. The term “controller / processor” may refer to one or more controllers, one or more processors, or a combination thereof. A channel processor may determine an RSRP parameter, a received signal strength indicator (RSSI) parameter, a reference signal received quality (RSRQ) parameter, and / or a CQI parameter, among other examples. In some examples, one or more components of the UE 120 may be included in a housing 284.
[0055] The network controller 130 may include a communication unit 294, a controller / processor 290, and a memory 292. The network controller 130 may include, for example, one or more devices in a core network. The network controller 130 may communicate with the network node 110 via the communication unit 294.
[0056] One or more antennas (e.g., antennas 234a through 234t and / or antennas 252a through 252r) may include, or may be included within, one or more antenna panels, one or more antenna groups, one or more sets of antenna elements, and / or one or more antenna arrays, among other examples. An antenna panel, an antenna group, a set of antenna elements, and / or an antenna array may include one or more antenna elements (within a single housing or multiple housings) , a set of coplanar antenna elements, a set ofnon-coplanar antenna elements, and / or one or more antenna elements coupled to one or more transmission and / or reception components, such as one or more components of Fig. 2.
[0057] 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 that include RSRP, RSSI, RSRQ, and / or CQI) from the controller / processor 280. The transmit processor 264 may 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 the modems 254 (e.g., for DFT-s-OFDM or CP-OFDM) , and transmitted to the network node 110. In some examples, the modem 254 of the UE 120 may include a modulator and a demodulator. In some examples, the UE 120 includes a transceiver. The transceiver may include any combination of the antenna (s) 252, the modem (s) 254, the MIMO detector 256, the receive processor 258, the transmit processor 264, and / or the TX MIMO processor 266. The transceiver may be used by a processor (e.g., the controller / processor 280) and the memory 282 to perform aspects of any of the methods described herein (e.g., with reference to Figs. 6-13) .
[0058] At the network node 110, the uplink signals from UE 120 and / or other UEs may be received by the antennas 234, processed by the modem 232 (e.g., a demodulator component, shown as DEMOD, of the modem 232) , 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 provide the decoded control information to the controller / processor 240. The network node 110 may include a communication unit 244 and may communicate with the network controller 130 via the communication unit 244. The network node 110 may include a scheduler 246 to schedule one or more UEs 120 for downlink and / or uplink communications. In some examples, the modem 232 of the network node 110 may include a modulator and a demodulator. In some examples, the network node 110 includes a transceiver. The transceiver may include any combination of the antenna (s) 234, the modem (s) 232, the MIMO detector 236, the receive processor 238, the transmit processor 220, and / or the TX MIMO processor 230. The transceiver may be used by a processor (e.g., the controller / processor 240) and the memory 242 to perform aspects of any of the methods described herein (e.g., with reference to Figs. 6-13) .
[0059] The controller / processor 240 of the network node 110, the controller / processor 280 of the UE 120, and / or any other component (s) of Fig. 2 may perform one or more techniques associated with reporting UE AI / ML model representation or quantization preferences, as described in more detail elsewhere herein. For example, the controller / processor 240 of the network node 110, the controller / processor 280 of the UE 120, and / or any other component (s) of Fig. 2 may perform or direct operations of, for example, process 1000 of Fig. 10, process 1100 of Fig. 11, and / or other processes as described herein. The memory 242 and the memory 282 may store data and program codes for the network node 110 and the UE 120, respectively. In some examples, the memory 242 and / or the memory 282 may include a non-transitory computer-readable medium storing one or more instructions (e.g., code and / or program code) for wireless communication. For example, the one or more instructions, when executed (e.g., directly, or after compiling, converting, and / or interpreting) by one or more processors of the network node 110 and / or the UE 120, may cause the one or more processors, the UE 120, and / or the network node 110 to perform or direct operations of, for example, process 1000 of Fig. 10, process 1100 of Fig. 11, and / or other processes as described herein. In some examples, executing instructions may include running the instructions, converting the instructions, compiling the instructions, and / or interpreting the instructions, among other examples.
[0060] In some aspects, a UE (e.g., the UE 120) includes means for transmitting an indication ofUE AI / ML model representation or quantization preferences; and / or means for receiving an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences. The means for the UE to perform operations described herein may include, for example, one or more of communication manager 140, antenna 252, modem 254, MIMO detector 256, receive processor 258, transmit processor 264, TX MIMO processor 266, controller / processor 280, or memory 282.
[0061] In some aspects, a network node (e.g., the network node 110) includes means for receiving an indication of UE AI / ML model representation or quantization preferences; and / or means for transmitting an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences. The means for the network node to perform operations described herein may include, for example, one or more of communication manager 150, transmit processor 220, TX MIMO processor 230, modem 232, antenna 234, MIMO detector 236, receive processor 238, controller / processor 240, memory 242, or scheduler 246.
[0062] In some aspects, an individual processor may perform all of the functions described as being performed by the one or more processors. In some aspects, one or more processors may collectively perform a set of functions. For example, a first set of (one or more) processors of the one or more processors may perform a first function described as being performed by the one or more processors, and a second set of (one or more) processors of the one or more processors may perform a second function described as being performed by the one or more processors. The first set of processors and the second set of processors may be the same set of processors or may be different sets of processors. Reference to “one or more processors” should be understood to refer to any one or more of the processors described in connection with Fig. 2. Reference to “one or more memories” should be understood to refer to any one or more memories of a corresponding device, such as the memory described in connection with Fig. 2. For example, functions described as being performed by one or more memories can be performed by the same subset of the one or more memories or different subsets of the one or more memories.
[0063] While blocks in Fig. 2 are illustrated as distinct components, the functions described above with respect to the blocks may be implemented in a single hardware, software, or combination component or in various combinations of components. For example, the functions described with respect to the transmit processor 264, the receive processor 258, and / or the TX MIMO processor 266 may be performed by or under the control of the controller / processor 280.
[0064] As indicated above, Fig. 2 is provided as an example. Other examples may differ from what is described with regard to Fig. 2.
[0065] Deployment of communication systems, such as 5G 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 RAN node, a core network node, a network element, a base station, or a network equipment may be implemented in an aggregated or disaggregated architecture. For example, a base station (such as a Node B (NB) , an evolved NB (eNB) , an NR base station, a 5G NB, an access point (AP) , a TRP, or a cell, among other examples) , or one or more units (or one or more components) performing base station functionality, may be implemented as an aggregated base station (also known as a standalone base station or a monolithic base station) or a disaggregated base station. “Network entity” or “network node” may refer to a disaggregated base station, or to one or more units of a disaggregated base station (such as one or more CUs, one or more DUs, one or more RUs, or a combination thereof) .
[0066] An aggregated base station (e.g., an aggregated network node) may be configured to utilize a radio protocol stack that is physically or logically integrated within a single RAN node (e.g., within a single device or unit) . A disaggregated base station (e.g., a disaggregated network node) may be configured to utilize a protocol stack that is physically or logically distributed among two or more units (such as one or more CUs, one or more DUs, or one or more RUs) . In some examples, a CU may be implemented within a network node, and one or more DUs may be co-located with the CU, or alternatively, may be geographically or virtually distributed throughout one or multiple other network 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, such as a virtual central unit (VCU) , a virtual distributed unit (VDU) , or a virtual radio unit (VRU) , among other examples.
[0067] 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 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) ) to facilitate scaling of communication systems by separating base station functionality into one or more units that can be individually deployed. A disaggregated base station may include functionality implemented across two or more units at various physical locations, as well as functionality implemented for at least one unit virtually, which can enable flexibility in network design. The various units of the disaggregated base station can be configured for wired or wireless communication with at least one other unit of the disaggregated base station.
[0068] Fig. 3 is a diagram illustrating an example disaggregated base station architecture 300, in accordance with the present disclosure. The disaggregated base station architecture 300 may include a CU 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 control units (such as a Near-RT RIC 325 via an E2 link, or a Non-RT RIC 315 associated with a Service Management and Orchestration (SMO) Framework 305, or both) . A CU 310 may communicate with one or more DUs 330 via respective midhaul links, such as through F1 interfaces. Each of the DUs 330 may communicate with one or more RUs 340 via respective fronthaul links. Each of the RUs 340 may communicate with one or more UEs 120 via respective radio frequency (RF) access links. In some implementations, a UE 120 may be simultaneously served by multiple RUs 340.
[0069] Each of the units, including 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 with 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 one or multiple communication interfaces of the respective unit, can be configured to communicate with one or more of the other units via the transmission medium. In some examples, each of 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, and a wireless interface, which may include a receiver, a transmitter or transceiver (such as an RF transceiver) , configured to receive or transmit signals, or both, over a wireless transmission medium to one or more of the other units.
[0070] In some aspects, the CU 310 may host one or more higher layer control functions. Such control functions can include radio resource control (RRC) functions, packet data convergence protocol (PDCP) functions, or service data adaptation protocol (SDAP) functions, among other examples. 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 (for example, Central Unit -User Plane (CU-UP) functionality) , control plane functionality (for example, Central Unit -Control Plane (CU-CP) functionality) , 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. A CU-UP unit can communicate bidirectionally with a 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 a DU 330, as necessary, for network control and signaling.
[0071] Each 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 depending, at least in part, on a functional split, such as a functional split defined by the 3GPP. In some aspects, the one or more high PHY layers may be implemented by one or more modules for forward error correction (FEC) encoding and decoding, scrambling, and modulation and demodulation, among other examples. In some aspects, the DU 330 may further host one or more low PHY layers, such as implemented by one or more modules for a fast Fourier transform (FFT) , an inverse FFT (iFFT) , digital beamforming, or physical random access channel (PRACH) extraction and filtering, among other examples. Each layer (which also may be referred to as a 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.
[0072] Each RU 340 may implement lower-layer functionality. 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 an FFT, performing an iFFT, digital beamforming, or PRACH extraction and filtering, among other examples, based on a functional split (for example, a functional split defined by the 3GPP) , such as a lower layer functional split. In such an architecture, each RU 340 can be operated 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 each DU 330 and the CU 310 to be implemented in a cloud-based RAN architecture, such as a vRAN architecture.
[0073] 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 O 1 interface) . For virtualized network elements, the SMO Framework 305 may be configured to interact with a cloud computing platform (such as an open cloud (O-Cloud) platform 390) to perform network element 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, non-RT RICs 315, 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) 311, via an O1 interface. Additionally, in some implementations, the SMO Framework 305 can communicate directly with each of one or more RUs 340 via a respective O1 interface. The SMO Framework 305 also may include a Non-RT RIC 315 configured to support functionality of the SMO Framework 305.
[0074] 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, AI / 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.
[0075] In some implementations, to generate AI / 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 AI / ML models to perform corrective actions through the SMO Framework 305 (such as reconfiguration via an O 1 interface) or via creation of RAN management policies (such as A1 interface policies) .
[0076] As indicated above, Fig. 3 is provided as an example. Other examples may differ from what is described with regard to Fig. 3.
[0077] Fig. 4 is a diagram illustrating an example 400 of beam management, in accordance with the present disclosure.
[0078] As shown by reference number 402, a UE may initially be in an RRC idle state or an RRC inactive state. As shown by reference number 404, the UE may perform an initial access. As shown by reference number 406, the UE may perform a beam management after entering an RRC connected state. The beam management may include P1, P2, and / or P3 beam management procedures. The P1 beam management procedure may be a beam selection procedure, an initial beam acquisition procedure, a beam sweeping procedure, a cell search procedure, and / or a beam search procedure. The P2 beam management procedure may be a beam refinement procedure, a network node beam refinement procedure, a TRP beam refinement procedure, and / or a transmit (Tx) beam refinement procedure. The P3 beam management procedure may be a beam refinement procedure, a UE beam refinement procedure, and / or a receive (Rx) beam refinement procedure. As shown by reference number 408, the UE may also perform beam management using an AI / ML-based approach. The beam management using the AI / ML-based approach may use an AI / ML model in a spatial domain (SD) , a time domain (TD) , and / or a frequency domain (FD) , which may reduce signaling overhead and latency and improve a beam selection accuracy. The AI / ML model may be associated with a lifecycle management, which may involve a model training, model deployment, model inference, model monitoring, and / or model updating. As shown by reference number 410, the UE may perform a beam failure detection (BFD) , which may be based at least in part on measurements obtained during the beam management after entering the RRC connected mode. As shown by reference number 412, the UE may perform a beam failure recovery (BFR) based at least in part on the BFD. As shown by reference number 414, when the BFR is not successful, the UE may declare a radio link failure (RLF) .
[0079] As indicated above, Fig. 4 is provided as an example. Other examples may differ from what is described with regard to Fig. 4.
[0080] Fig. 5 is a diagram illustrating examples 500 of AI / ML models for beam prediction, in accordance with the present disclosure.
[0081] As shown by reference number 502, an AI / ML model for beam prediction may run at a network node. A UE may feedback channel state information (CSI) with a beam report. The UE may feedback a sounding reference signal (SRS) . The network node may run the AI / ML model for beam prediction based at least in part on feedback received from the UE. The network node may determine a beam prediction result using the AI / ML model for beam prediction. The network node may perform a scheduling decision for the UE based at least in part on the beam prediction result. The AI / ML model for beam prediction may be run at the network node when the UE is power and / or computation power limited.
[0082] As shown by reference number 504, an AI / ML model for beam prediction may be run at a UE. The AI / ML model for beam prediction may be configured by a network node. The UE may run the AI / ML model for beam prediction based at least in part on local measurements and signaling (e.g., reference signals to be measured) from the network node. The UE may report, to the network node, a beam prediction result based at least in part on a configuration or triggering condition. The UE may run the AI / ML model for beam prediction because the UE may generally have more measurement results, as compared to the network node, and running the AI / ML model for beam prediction at the UE may involve less overhead for reporting. However, running the AI / ML model for beam prediction at the UE may involve a higher computation capability at the UE.
[0083] As indicated above, Fig. 5 is provided as an example. Other examples may differ from what is described with regard to Fig. 5.
[0084] A network node may train an AI / ML model for beam prediction based at least in part on data collected via UE reported L1-RSRP measurements. The UE reported L1-RSRP measurements may be historical measurements for certain areas, scenarios, and / or codebook configurations. When a certain UE enters such an area, scenario, or codebook configuration, the network node may transfer the AI / ML model to that UE. The network node may transfer the AI / ML model to the UE via an AI / ML air interface. The UE may locally run a model inference using the AI / ML model received from the network node. In this case, the network node may train the AI / ML model, as opposed to the UE autonomously training an AI / ML model. The network node training the AI / ML model, as opposed to the UE training the AI / ML model, may be beneficial because the network node may have a better understanding on codebooks, antenna setups, tilting directions, and / or parent-child relationships of its beams as compared to the UE.
[0085] An AI / ML model compilation, an AI / ML model representation, and / or an AI / ML model quantization may be carried out by the network node or the UE. When labeled training data is used for AI / ML model quantization, the AI / ML model quantization may occur at the network node. The use of labeled training data may result in improved AI / ML model inference accuracy. Otherwise, the UE may quantize the AI / ML model by itself, for example without labeled training data (due to fewer resources at the UE as compared to the network node) , which may lead to a degraded AI / ML model inference accuracy.
[0086] Different UEs may have various preferences regarding the quantization of AI / ML models. Such various quantization preferences (e.g., number of bits per parameter, quantization metrics, and / or step-sizes) may be due to implementation limitations of hardware based AI / ML accelerators at the different UEs. Depending on available computational resources, different components / layers associated with the same AI / ML model may also be preferred with different quantization granularities.
[0087] For AI / ML model transfer fiom the network node to the UE, a UE preference on AI / ML quantization may not be known to the network node. AI / ML model quantization may depend on training / inference data, so AI / ML model quantization may be more reasonable to be carried out by the network node, as opposed to the UE, when the AI / ML model is trained by the network node and transferred to the UE. A quantization dynamic range in respective layers may depend on an original model and training / inference data, but such data may not be available at the UE, so the AI / ML model quantization may be carried out by the network node instead of the UE. However, existing designs on AI / ML model transfer from the network node to the UE do not consider mechanisms for allowing the network node to be aware of UE-side AI / ML model quantization preferences.
[0088] For example, the network node may collect data based at least in part on UE feedback or uplink measurements. The network node may use such data to train AI / ML models suitable for certain network configurations (e.g., beam codebooks associated with beam prediction) . The network node may transfer such AI / ML models to the UE, which may allow the UE to locally run such AI / ML models based at least in part on the UE's own observations.
[0089] In various aspects of techniques and apparatuses described herein, a UE may transmit, to a network node, an indication ofUE AI / ML model representation or quantization preferences. The UE AI / ML model representation or quantization preferences may indicate a number of bits to represent or quantize a single parameter, metrics associated with determining an AI / ML model parameter representation or quantization, an indication of whether the UE prefers a network node to represent or quantize an AI / ML model or whether the UE prefers to represent or quantize the AI / ML model, an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, and / or components or layer specific AI / ML model parameter representation or quantization preferences. The UE may receive, from the network node, an AI / ML model. A representation or quantization of one or more parameters associated with the AI / ML model may be based at least in part on the UE AI / ML model representation or quantization preferences. The AI / ML model may be associated with a network node side AI / ML model parameter representation or quantization, or alternatively, the AI / ML model may be associated with a UE side AI / ML model parameter representation or quantization, based at least in part on the UE AI / ML model representation or quantization preferences. The UE may perform a model inference using the AI / ML model.
[0090] In some aspects, the UE may feedback, to the network node, AI / ML model quantization preferences for AI / ML models transferred from the network node to the UE. The UE may feedback the AI / ML model quantization preferences based at least in part on a signaling enhancement, which may result in the AI / ML model transfer from the network node to the UE being more practical and efficiently deployed. The AI / ML model quantization may be coordinated based at least in part on UE preferences regarding AI / ML model quantization.
[0091] In some aspects, by reporting UE AI / ML model representation or quantization preferences, the described techniques can be used to transfer AI / ML models that are based at least in part on the UE AI / ML model representation or quantization preferences. The network node, when quantizing the AI / ML model, may consider UE preferences, which may vary depending on implementation limitations of hardware based AI / ML accelerators at the UE. Thus, the AI / ML model may be specifically designed based on a UE capability, as indicated by the UE preferences, which may result in AI / ML models that are better suited to run at the UE. Such AI / ML models may improve an overall performance at the UE because the AI / ML models may be tailored to the implementation limitations of hardware based AI / ML accelerators at the UE.
[0092] Fig. 6 is a diagram illustrating an example 600 associated with reporting UE AI / ML model representation or quantization preferences, in accordance with the present disclosure. As shown in Fig. 6, example 600 includes communication between a UE (e.g., UE 120) and a network node (e.g., network node 110) . In some aspects, the UE and the network node may be included in a wireless network, such as wireless network 100.
[0093] As shown by reference number 602, the UE may transmit, to the network node, an indication of UE AI / ML model representation or quantization preferences. In some aspects, the UE AI / ML model representation or quantization preferences may be based at least in part on a standard definition or a network node configuration. The UE AI / ML model representation or quantization preferences may indicate a number of bits to represent or quantize a single parameter. The UE AI / ML model representation or quantization preferences may indicate metrics associated with determining a AI / ML model parameter representation or quantization. The UE AI / ML model representation or quantization preferences may indicate an indication of whether the UE prefers a network node to represent or quantize an AI / ML model or whether the UE prefers to represent or quantize the AI / ML model. The UE AI / ML model representation or quantization preferences may indicate an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation. The UE AI / ML model representation or quantization preferences may indicate components or layer specific AI / ML model parameter representation or quantization preferences. The UE may transmit, to the network node, the indication of the UE AI / ML model representation or quantization preferences based at least in part on a user plane protocol, RRC signaling, a MAC control element (MAC-CE) , uplink control information (UCI) , and / or an application layer protocol. The UE may transmit the indication of the UE AI / ML model representation or quantization preferences prior to a training of the AI / ML model.
[0094] As shown by reference number 604, the UE may receive, from the network node, the AI / ML model, where a representation or quantization of one or more parameters associated with the AI / ML model may be based at least in part on the UE AI / ML model representation or quantization preferences. The UE may perform a model inference using the AI / ML model. The UE may indicate model inference results to the network node.
[0095] In some aspects, the network node may quantize the AI / ML model, and the network node may transmit the quantized AI / ML model to the UE (e.g., as shown in Fig. 7) . A network node side AI / ML model parameter representation or quantization of the AI / ML model may be based at least in part on labeled training data and the UE AI / ML model representation or quantization preferences. The AI / ML model may be the quantized AI / ML model and may be based at least in part on the UE AI / ML model representation or quantization preferences.
[0096] In some aspects, the UE may quantize the AI / ML model instead of the network node (e.g., as shown in Fig. 9) . The UE may transmit, to the network node, a request for labeled training data. The UE may receive, from the network node, the labeled training data based at least in part on the request. A UE side AI / ML model parameter representation or quantization of the AI / ML model may be based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences. Alternatively, the UE may receive, from the network node, a reject message based at least in part on the request. A UE side AI / ML model parameter representation or quantization of the AI / ML model may be based at least in part on the UE AI / ML model representation or quantization preferences and without the labeled training data. In some aspects, the UE may transmit, to the network node, an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, and the UE AI / ML model representation or quantization scheme may be used for AI / ML functionality or model life cycle management.
[0097] In some aspects, the UE may receive, from the network node and prior to the training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model (e.g., as shown in Fig. 9) . Alternatively, the UE may transmit, to the network node and prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model. A data collection, an AI / ML model training, and / or an AI / ML model quantization may be based at least in part on the UE AI / ML model representation or quantization preferences.
[0098] In some aspects, the UE may receive, from the network node, an indication in system information regarding AI / ML model representation or quantization techniques supported by the network node. The indication may be applicable to a cell or a cell group associated with the system information, or the indication may be applicable to a whole network. In some aspects, the indication may be separate for different types of AI / ML models. In some aspects, the UE may receive, from the network node, a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, and the UE may transmit the indication ofUE AI / ML model representation or quantization preferences based at least in part on the request.
[0099] As indicated above, Fig. 6 is provided as an example. Other examples may differ from what is described with regard to Fig. 6.
[0100] Fig. 7 is a diagram illustrating an example 700 associated with reporting UE AI / ML model representation or quantization preferences, in accordance with the present disclosure.
[0101] In some aspects, a UE AI / ML model quantization preference report may be employed for AI / ML model transfer from a network node to a UE. The LUE may report to the network node regarding its preference on techniques for representing or quantizing parameters associated with one or more AI / ML models. The network node may further transfer the one or more AI / ML models to the UE, where a representation or quantization of parameters associated with the one or more AI / ML models may be based at least in part on the UE reporting.
[0102] As shown by reference number 702, at the network node (or network entity) , labeled training data may be used for AI / ML model training of the AI / ML models. The AI / ML models may be associated with a floating point representation of parameters. The labeled training data may be used as input for an AI / ML model parameter representation or quantization (e.g., fixed point or integer) . The UE may report, to the network node, a UE preference for carrying out the AI / ML model parameter representation or quantization. As shown by reference number 704, the network node may perform the AI / ML model parameter representation or quantization based at least in part on the UE reporting. The network node may perform the AI / ML model parameter representation or quantization via fixed point or integer representations, or via a floating point model when the UE prefers to quantize by itself. As shown by reference number 706, the network node may then transfer the AI / ML models to the UE.
[0103] In some aspects, representation or quantization preferences may be standard predefined or configured by the network node, and such representation or quantization preferences may be reported by the UE to the network node. The representation or quantization preferences may be predefined or configured as candidate representation or quantization preferences. The representation or quantization preferences may include a number of bits to represent or quantize a single parameter, which may include an X-bits fixed-point / integer representation or quantization per parameter, with X being an integer greater than one. A range of X may be based at least in part on a standard predefinition, or the range of X may be indicated by the network node to the UE. The representation or quantization preferences may include, given a certain value of X, metrics on determining a representation or quantization from floating point to fixed point / integer. The metrics may include clustering, K-means, and / or entropy-based metrics. The metrics may include uniformed metrics or non-uniformed metrics (e.g., log, exponential, or other irregular metrics) . The representation or quantization preferences may include an indication of whether the UE prefers the network node to represent or quantize the AI / ML model, or whether the UE prefers to represent or quantize the AI / ML model itself. For a network size quantization, quantization aware training may be considered, in which access to labeled training data may be available and retraining of the AI / ML model may be needed. For UE side quantization, post training quantization may be considered, in which access to labeled training data is not available and retraining of the AI / ML model is not needed.
[0104] In some aspects, the representation or quantization preferences may include an indication of whether the same or different representation scheme or quantization scheme is applied to weights, bias, and / or activation. When different, the UE may report respective preferences on weights, bias, and / or activation. The representation or quantization preferences may include component or layer specific representation preferences or quantization preferences. The UE may report different components / layers of the AI / ML model. For example, certain layers / components may be quantized via INT8 (8-bit integer) , while other layer / components may be quantized via INT16 (16-bit integer) . The component or layer specific representation preferences or quantization preferences may be based at least in part on whether the network node has informed the UE of an AI / ML model structure. Further, the components / layers may include fully connected (FC) layer (s) , convolutional (Conv) layer (s) , pooling, and / or addition.
[0105] In some aspects, various schemes to feedback UE representation or quantization preferences may be defined. In some aspects, the UE may feedback the UE representation or quantization preferences based at least in part on user plane protocols. Such reporting may be considered to be based at least in part on a single choice or a choice combination. An AI / ML model transfer from the network node to the UE may be based at least in part on protocols in a user plane, so UE reports on the UE representation or quantization preferences may also be based at least in part on the user plane protocols. In some aspects, the UE may feedback the UE representation or quantization preferences based at least in part on RRC signaling (e.g., as UE capabilities during an initial access) . Such reporting may be considered to be based at least in part on a single choice or a choice combination. Alternatively, such reporting may be based at least in part on reporting multiple choices, or multiple choice combinations. For AI / ML models that assist with wireless communications between the network node and the UE, the UE may feedback its UE representation or quantization preferences through RRC via a UE capability report during the initial access. The network node may transfer, to the UE, variable / different AI / ML models when the UE moves from one scenario, environment, configuration, and / or area to another scenario, environment, configuration, and / or area, where different AI / ML models may be used in different scenarios, environments, configurations, and / or areas.
[0106] In some aspects, the UE may feedback the UE representation or quantization preferences based at least in part on MAC-CE signaling or UCI signaling. Such reporting may be considered to be based at least in part on a single choice or a choice combination. Alternatively, such reporting may be based at least in part on indicating choice or choice combination ID (s) that have been reported via RRC signaling. Computational resources available at the UE may change dynamically, so the UE representation or quantization preferences may also change dynamically, such that a MAC-CE / UCI based report may be needed. In some aspects, the UE may feedback the UE representation or quantization preferences based at least in part on application layer protocols. For example, the network node may be associated with a third party server, which may be used to feedback the UE representation or quantization preferences.
[0107] As indicated above, Fig. 7 is provided as an example. Other examples may differ from what is described with regard to Fig. 7.
[0108] Fig. 8 is a diagram illustrating an example 800 associated with reporting UE AI / ML model representation or quantization preferences, in accordance with the present disclosure.
[0109] In some aspects, a network node may transfer, to a UE, labeled training data for UE-side AI / ML model quantization. When the UE prefers to represent or quantize an AI / ML model itself, the UE may request the network node to transmit the labeled training data to assist with AI / ML model representation or quantization. The network node may confirm such request and send the labeled training data together with a raw representation (in floating-point) of the AI / ML model. Alternatively, the network node may reject the request and let the UE perform the AI / ML model representation or quantization without the labeled training data. In some aspects, the network node may signal, to the UE, restrictions on the AI / ML model representation or quantization carried out by UE. For example, the network node may indicate, to the UE, the number of bits to represent or quantize a single parameter, the metrics associated with determining the representation or quantization, the indication of whether the same or different representation scheme or quantization scheme is to be applied, and / or component or layer specific representation or quantization preferences. In some aspects, the UE may report, to the network node, information associated with AI / ML model representation or quantization techniques used by the UE for an associated AI / ML model. Such information may be used by the network node for AI / ML functionality / model life cycle management. For example, when a relatively poor performance of features associated with UE autonomously quantized AI / ML models is observed, the network node may request / recommend the UE to alternatively switch to an AI / model quantized by the network node and transferred from the network node to the UE.
[0110] As shown by reference number 802, at the network node (or network entity) , labeled training data may be used for AI / ML model training of AI / ML models. The AI / ML models may be associated with a floating point representation of parameters. The network node may transfer the AI / ML models to the UE. The UE may transmit, to the network node, a request for the labeled training data. As shown by reference number 804, the network node may transfer the labeled training data to the UE based at least in part on the UE request. As shown by reference number 806, the labeled training data may be used as input for an AI / ML model parameter representation or quantization (e.g., fixed point or integer) performed by the UE. As shown by reference number 808, the UE may report, to the network node, information associated with AI / ML model representation or quantization techniques used by the UE for the AI / ML models. The UE may perform the AI / ML model parameter representation or quantization using the AI / ML model representation or quantization techniques. The UE may perform the AI / ML model parameter representation or quantization via fixed point or integer representations.
[0111] As indicated above, Fig. 8 is provided as an example. Other examples may differ from what is described with regard to Fig. 8.
[0112] Fig. 9 is a diagram illustrating an example 900 associated with reporting UE AI / ML model representation or quantization preferences, in accordance with the present disclosure.
[0113] In some aspects, a UE may report, to a network node, UE representation or quantization preferences before an actual AI / ML model training. The UE may report the UE representation or quantization preferences based at least in part on a signaling scheme (e.g., user plane, RRC, or application layer protocol) . Before the UE reports the UE representation or quantization preferences, the network node may initially indicate while offline an AI / ML model structure without training the AI / ML model. The network node may indicate the AI / ML model structure based at least in part on a model identifier for a certain AI / ML functionality. The AI / ML model may not yet be trained, but its structure may already be determined. The network node may indicate multiple model structures, together with their candidate representation or quantization schemes. For example, the network node may indicate a model structure of transformer and a representation or quantization scheme of fixed point 16. The network node may indicate a model structure of long short-term memory (LSTM) and a representation or quantization scheme of INT 16. The network node may indicate a model structure of convolutional neural network (CNN) and a representation or quantization scheme of INT8.
[0114] In some aspects, model structure (s) may be identified by the UE and reported from the UE to the network node. The UE may report its UE representation or quantization preferences after the model structure (s) are identified. When multiple structures are reported from the UE to the network node, further reported representation or quantization preferences may be respectively addressed for different structures (e.g., transformer and fixed point 16, LSTM and INT16, or CNN and INT8) . Depending on the UE representation or quantization preferences reported, the network node may carry out data collection and AI / ML model training (e.g., quantization aware AI / ML model training) , and then the network node may transfer a quantized AI / ML model to the UE. Such a procedure may take a relatively long duration (e.g., in days or months) .
[0115] As shown by reference number 902, the network node may report to the UE, or the UE may report to the network node, a model identification. The model identification may be reported along with a corresponding AI / ML model structure. For example, the model identification may include a model identifier together with its structure identified for L 1 or layer 2 (L2) triggered mobility (LTM) regarding a certain city. As shown by reference number 904, the UE may transmit, to the network node, an AI / ML model parameter representation or quantization preference report. As shown by reference number 906, the network node may perform a data collection, an AI / ML model training, and / or an AI / ML model quantization. The data collection, the AI / ML model training, and / or the AI / ML model quantization may take several days or months. As shown by reference number 908, after the data collection, the AI / ML model training, and / or the AI / ML model quantization, the network node may transfer a quantized AI / ML model to the UE. As shown by reference number 910, the UE may perform an AI / ML model inference using the quantized AI / ML model. As shown by reference number 912, the UE and the network node may perform interactions based at least in part on AI / ML model inference results (e.g., reported predicted L1-RSRP measurements on non-measured synchronization signal blocks (SSBs) in LTM) .
[0116] As indicated above, Fig. 9 is provided as an example. Other examples may differ from what is described with regard to Fig. 9.
[0117] In some aspects, interactions between a network node and a UE for RRC, MAC-CE, and / or UCI based preference feedback may be defined. The UE may receive a network node indication from system information, regarding candidate AI / ML model representation or quantization techniques that are supported by the network node. The network node indication may be applied only to a cell or cell group associated with the system information, or the network node indication may be applied to a whole network. When the network node indication is only applied to the cell or cell group, the UE may receive such updated information in a cell common serving cell specific configuration, so that the UE may update such information during mobility. The network node indication may be separate for different types of AI / ML models. For example, for beam prediction based models, the network node may indicate AI / ML model representation or quantization techniques based on INT4 or INT8, while for CSI feedback (CSF) models, the network node may indicate AI / ML model representation or quantization techniques based at least in part on INT8 or INT 16. As another example, for pure temporal or pure spatial beam prediction models, the network node may indicate AI / ML model representation or quantization techniques based at least in part on INT4 or INT8, while for joint temporal and spatial beam prediction models, the network node may indicate AI / ML model representation or quantization techniques based at least in part on INT8 or INT16.
[0118] In some aspects, the network node may transmit, to the UE, a request for instantaneous UE feedback on its AI / ML model representation / quantization preferences, where the request may be associated with a certain AI / ML model or a certain AI / ML model type. For example, the network node may instantaneously determine to transfer a certain AI / ML model to the UE, where such an instantaneous request on the UE’s AI / ML model representation or quantization preferences may be needed to decide which quantized AI / ML model should be provided to the UE.
[0119] Fig. 10 is a diagram illustrating an example process 1000 performed, for example, at a UE or an apparatus of a UE, in accordance with the present disclosure. Example process 1000 is an example where the apparatus or the UE (e.g., UE 120) performs operations associated with reporting UE AI / ML model representation or quantization preferences.
[0120] As shown in Fig. 10, in some aspects, process 1000 may include transmitting an indication ofUE AI / ML model representation or quantization preferences (block 1010) . For example, the UE (e.g., using transmission component 1204 and / or communication manager 1206, depicted in Fig. 12) may transmit an indication of UE AI / ML model representation or quantization preferences, as described above.
[0121] As further shown in Fig. 10, in some aspects, process 1000 may include receiving an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences (block 1020) . For example, the UE (e.g., using reception component 1202 and / or communication manager 1206, depicted in Fig. 12) may receive an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences, as described above.
[0122] Process 1000 may include additional aspects, such as any single aspect or any combination of aspects described below and / or in connection with one or more other processes described elsewhere herein.
[0123] In a first aspect, a network node side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on labeled training data and the UE AI / ML model representation or quantization preferences.
[0124] In a second aspect, alone or in combination with the first aspect, the AI / ML model is a quantized AI / ML model and is based at least in part on the UE AI / ML model representation or quantization preferences.
[0125] In a third aspect, alone or in combination with one or more of the first and second aspects, the UE AI / ML model representation or quantization preferences are based at least in part on a standard definition or a network node configuration.
[0126] In a fourth aspect, alone or in combination with one or more of the first through third aspects, the UE AI / ML model representation or quantization preferences indicate one or more of a number of bits to represent or quantize a single parameter, metrics associated with determining an AI / ML model parameter representation or quantization, an indication of whether the UE prefers a network node to represent or quantize the AI / ML model or whether the UE prefers to represent or quantize the AI / ML model, an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, or components or layer specific AI / ML model parameter representation or quantization preferences.
[0127] In a fifth aspect, alone or in combination with one or more of the first through fourth aspects, process 1000 includes transmitting a request for labeled training data, and receiving labeled training data based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences.
[0128] In a sixth aspect, alone or in combination with one or more of the first through fifth aspects, process 1000 includes transmitting a request for labeled training data, and receiving a reject message based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences and an exclusion of the labeled training data.
[0129] In a seventh aspect, alone or in combination with one or more of the first through sixth aspects, process 1000 includes receiving an indication of a restriction associated with a UE side AI / ML model parameter representation or quantization.
[0130] In an eighth aspect, alone or in combination with one or more of the first through seventh aspects, process 1000 includes transmitting an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, wherein the UE AI / ML model representation or quantization scheme is used for AI / ML functionality or model life cycle management.
[0131] In a ninth aspect, alone or in combination with one or more of the first through eighth aspects, transmitting the indication of the UE AI / ML model representation or quantization preferences is based at least in part on a user plane protocol, RRC signaling, a MAC-CE, UCI, or an application layer protocol.
[0132] In a tenth aspect, alone or in combination with one or more of the first through ninth aspects, the indication of the UE AI / ML model representation or quantization preferences is transmitted prior to a training of the AI / ML model.
[0133] In an eleventh aspect, alone or in combination with one or more of the first through tenth aspects, process 1000 includes receiving, prior to a training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model, or transmitting, prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model.
[0134] In a twelfth aspect, alone or in combination with one or more of the first through eleventh aspects, one or more of a data collection, an AI / ML model training, and an AI / ML model quantization is based at least in part on the UE AI / ML model representation or quantization preferences.
[0135] In a thirteenth aspect, alone or in combination with one or more of the first through twelfth aspects, process 1000 includes receiving an indication in system information regarding AI / ML model representation or quantization techniques supported by a network node, wherein the indication is applicable to a cell or a cell group associated with the system information, or the indication is applicable to a whole network, or the indication is separate for different types of AI / ML models.
[0136] In a fourteenth aspect, alone or in combination with one or more of the first through thirteenth aspects, process 1000 includes receiving a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, wherein the indication is transmitted based at least in part on the request.
[0137] Although Fig. 10 shows example blocks of process 1000, in some aspects, process 1000 may include additional blocks, fewer blocks, different blocks, or differently arranged blocks than those depicted in Fig. 10. Additionally, or alternatively, two or more of the blocks of process 1000 may be performed in parallel.
[0138] Fig. 11 is a diagram illustrating an example process 1100 performed, for example, at a network node or an apparatus of a network node, in accordance with the present disclosure. Example process 1100 is an example where the apparatus or the network node (e.g., network node 110) performs operations associated with reporting UE AI / ML model representation or quantization preferences.
[0139] As shown in Fig. 11, in some aspects, process 1100 may include receiving an indication ofUE AI / ML model representation or quantization preferences (block 1110) . For example, the network node (e.g., using reception component 1302 and / or communication manager 1306, depicted in Fig. 13) may receive an indication of UE AI / ML model representation or quantization preferences, as described above.
[0140] As further shown in Fig. 11, in some aspects, process 1100 may include transmitting an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences (block 1120) . For example, the network node (e.g., using transmission component 1304 and / or communication manager 1306, depicted in Fig. 13) may transmit an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences, as described above.
[0141] Process 1100 may include additional aspects, such as any single aspect or any combination of aspects described below and / or in connection with one or more other processes described elsewhere herein.
[0142] In a first aspect, a network node side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on labeled training data and the UE AI / ML model representation or quantization preferences.
[0143] In a second aspect, alone or in combination with the first aspect, the AI / ML model is a quantized AI / ML model and is based at least in part on the UE AI / ML model representation or quantization preferences.
[0144] In a third aspect, alone or in combination with one or more of the first and second aspects, the UE AI / ML model representation or quantization preferences are based at least in part on a standard definition or a network node configuration.
[0145] In a fourth aspect, alone or in combination with one or more of the first through third aspects, the UE AI / ML model representation or quantization preferences indicate one or more of a number of bits to represent or quantize a single parameter, metrics associated with determining a AI / ML model parameter representation or quantization, an indication of whether the UE prefers a network node to represent or quantize the AI / ML model or whether the UE prefers to represent or quantize the AI / ML model, an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, or components or layer specific AI / ML model parameter representation or quantization preferences.
[0146] In a fifth aspect, alone or in combination with one or more of the first through fourth aspects, process 1100 includes receiving a request for labeled training data, and transmitting labeled training data based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences.
[0147] In a sixth aspect, alone or in combination with one or more of the first through fifth aspects, process 1100 includes receiving a request for labeled training data, and transmitting a reject message based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences and an exclusion of the labeled training data.
[0148] In a seventh aspect, alone or in combination with one or more of the first through sixth aspects, process 1100 includes transmitting an indication of a restriction associated with a UE side AI / ML model parameter representation or quantization.
[0149] In an eighth aspect, alone or in combination with one or more of the first through seventh aspects, process 1100 includes receiving an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, wherein the UE AI / ML model representation or quantization scheme is used for AI / ML functionality or model life cycle management.
[0150] In a ninth aspect, alone or in combination with one or more of the first through eighth aspects, receiving the indication of the UE AI / ML model representation or quantization preferences is based at least in part on a user plane protocol, RRC signaling, a MAC-CE, UCI, or an application layer protocol.
[0151] In a tenth aspect, alone or in combination with one or more of the first through ninth aspects, the indication of the UE AI / ML model representation or quantization preferences is received prior to a training of the AI / ML model.
[0152] In an eleventh aspect, alone or in combination with one or more of the first through tenth aspects, process 1100 includes transmitting, prior to a training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model, or receiving, prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model.
[0153] In a twelfth aspect, alone or in combination with one or more of the first through eleventh aspects, one or more of a data collection, an AI / ML model training, and an AI / ML model quantization is based at least in part on the UE AI / ML model representation or quantization preferences.
[0154] In a thirteenth aspect, alone or in combination with one or more of the first through twelfth aspects, process 1100 includes transmitting an indication in system information regarding AI / ML model representation or quantization techniques supported by a network node, wherein the indication is applicable to a cell or a cell group associated with the system information, or the indication is applicable to a whole network, or the indication is separate for different types of AI / ML models.
[0155] In a fourteenth aspect, alone or in combination with one or more of the first through thirteenth aspects, process 1100 includes transmitting a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, wherein the indication is received based at least in part on the request.
[0156] Although Fig. 11 shows example blocks of process 1100, in some aspects, process 1100 may include additional blocks, fewer blocks, different blocks, or differently arranged blocks than those depicted in Fig. 11. Additionally, or alternatively, two or more of the blocks of process 1100 may be performed in parallel.
[0157] Fig. 12 is a diagram of an example apparatus 1200 for wireless communication, in accordance with the present disclosure. The apparatus 1200 may be a UE, or a UE may include the apparatus 1200. In some aspects, the apparatus 1200 includes a reception component 1202, a transmission component 1204, and / or a communication manager 1206, which may be in communication with one another (for example, via one or more buses and / or one or more other components) . In some aspects, the communication manager 1206 is the communication manager 140 described in connection with Fig. 1. As shown, the apparatus 1200 may communicate with another apparatus 1208, such as a UE or a network node (such as a CU, a DU, an RU, or a base station) , using the reception component 1202 and the transmission component 1204.
[0158] In some aspects, the apparatus 1200 may be configured to perform one or more operations described herein in connection with Figs. 6-9. Additionally, or alternatively, the apparatus 1200 may be configured to perform one or more processes described herein, such as process 1000 of Fig. 10. In some aspects, the apparatus 1200 and / or one or more components shown in Fig. 12 may include one or more components of the UE described in connection with Fig. 2. Additionally, or alternatively, one or more components shown in Fig. 12 may be implemented within one or more components described in connection with Fig. 2. Additionally, or alternatively, one or more components of the set of components may be implemented at least in part as software stored in one or more memories. For example, a component (or a portion of a component) may be implemented as instructions or code stored in a non-transitory computer-readable medium and executable by one or more controllers or one or more processors to perform the functions or operations of the component.
[0159] The reception component 1202 may receive communications, such as reference signals, control information, data communications, or a combination thereof, from the apparatus 1208. The reception component 1202 may provide received communications to one or more other components of the apparatus 1200. In some aspects, the reception component 1202 may perform signal processing on the received communications (such as filtering, amplification, demodulation, analog-to-digital conversion, demultiplexing, deinterleaving, de-mapping, equalization, interference cancellation, or decoding, among other examples) , and may provide the processed signals to the one or more other components of the apparatus 1200. In some aspects, the reception component 1202 may include one or more antennas, one or more modems, one or more demodulators, one or more MIMO detectors, one or more receive processors, one or more controllers / processors, one or more memories, or a combination thereof, of the UE described in connection with Fig. 2.
[0160] The transmission component 1204 may transmit communications, such as reference signals, control information, data communications, or a combination thereof, to the apparatus 1208. In some aspects, one or more other components of the apparatus 1200 may generate communications and may provide the generated communications to the transmission component 1204 for transmission to the apparatus 1208. In some aspects, the transmission component 1204 may perform signal processing on the generated communications (such as filtering, amplification, modulation, digital-to-analog conversion, multiplexing, interleaving, mapping, or encoding, among other examples) , and may transmit the processed signals to the apparatus 1208. In some aspects, the transmission component 1204 may include one or more antennas, one or more modems, one or more modulators, one or more transmit MIMO processors, one or more transmit processors, one or more controllers / processors, one or more memories, or a combination thereof, of the UE described in connection with Fig. 2. In some aspects, the transmission component 1204 may be co-located with the reception component 1202 in one or more transceivers.
[0161] The communication manager 1206 may support operations of the reception component 1202 and / or the transmission component 1204. For example, the communication manager 1206 may receive information associated with configuring reception of communications by the reception component 1202 and / or transmission of communications by the transmission component 1204. Additionally, or alternatively, the communication manager 1206 may generate and / or provide control information to the reception component 1202 and / or the transmission component 1204 to control reception and / or transmission of communications.
[0162] The transmission component 1204 may transmit an indication ofUE AI / ML model representation or quantization preferences. The reception component 1202 may receive an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0163] The transmission component 1204 may transmit a request for labeled training data. The reception component 1202 may receive labeled training data based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences. The transmission component 1204 may transmit a request for labeled training data. The reception component 1202 may receive a reject message based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences and an exclusion of the labeled training data.
[0164] The reception component 1202 may receive an indication of a restriction associated with a UE side AI / ML model parameter representation or quantization. The transmission component 1204 may transmit an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, wherein the UE AI / ML model representation or quantization scheme is used for AI / ML functionality or model life cycle management. The reception component 1202 may receive, prior to a training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model. The transmission component 1204 may transmit, prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model.
[0165] The reception component 1202 may receive an indication in system information regarding AI / ML model representation or quantization techniques supported by a network node, wherein the indication is applicable to a cell or a cell group associated with the system information, or the indication is applicable to a whole network, or the indication is separate for different types of AI / ML models. The reception component 1202 may receive a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, wherein the indication is transmitted based at least in part on the request.
[0166] The number and arrangement of components shown in Fig. 12 are provided as an example. In practice, there may be additional components, fewer components, different components, or differently arranged components than those shown in Fig. 12. Furthermore, two or more components shown in Fig. 12 may be implemented within a single component, or a single component shown in Fig. 12 may be implemented as multiple, distributed components. Additionally, or alternatively, a set of (one or more) components shown in Fig. 12 may perform one or more functions described as being performed by another set of components shown in Fig. 12.
[0167] Fig. 13 is a diagram of an example apparatus 1300 for wireless communication, in accordance with the present disclosure. The apparatus 1300 may be a network node, or a network node may include the apparatus 1300. In some aspects, the apparatus 1300 includes a reception component 1302, a transmission component 1304, and / or a communication manager 1306, which may be in communication with one another (for example, via one or more buses and / or one or more other components) . In some aspects, the communication manager 1306 is the communication manager 150 described in connection with Fig. 1. As shown, the apparatus 1300 may communicate with another apparatus 1308, such as a UE or a network node (such as a CU, a DU, an RU, or a base station) , using the reception component 1302 and the transmission component 1304.
[0168] In some aspects, the apparatus 1300 may be configured to perform one or more operations described herein in connection with Figs. 6-9. Additionally, or alternatively, the apparatus 1300 may be configured to perform one or more processes described herein, such as process 1100 of Fig. 11. In some aspects, the apparatus 1300 and / or one or more components shown in Fig. 13 may include one or more components of the network node described in connection with Fig. 2. Additionally, or alternatively, one or more components shown in Fig. 13 may be implemented within one or more components described in connection with Fig. 2. Additionally, or alternatively, one or more components of the set of components may be implemented at least in part as software stored in one or more memories. For example, a component (or a portion of a component) may be implemented as instructions or code stored in a non-transitory computer-readable medium and executable by one or more controllers or one or more processors to perform the functions or operations of the component.
[0169] The reception component 1302 may receive communications, such as reference signals, control information, data communications, or a combination thereof, from the apparatus 1308. The reception component 1302 may provide received communications to one or more other components of the apparatus 1300. In some aspects, the reception component 1302 may perform signal processing on the received communications (such as filtering, amplification, demodulation, analog-to-digital conversion, demultiplexing, deinterleaving, de-mapping, equalization, interference cancellation, or decoding, among other examples) , and may provide the processed signals to the one or more other components of the apparatus 1300. In some aspects, the reception component 1302 may include one or more antennas, one or more modems, one or more demodulators, one or more MIMO detectors, one or more receive processors, one or more controllers / processors, one or more memories, or a combination thereof, of the network node described in connection with Fig. 2. In some aspects, the reception component 1302 and / or the transmission component 1304 may include or may be included in a network interface. The network interface may be configured to obtain and / or output signals for the apparatus 1300 via one or more communications links, such as a backhaul link, a midhaul link, and / or a fronthaul link.
[0170] The transmission component 1304 may transmit communications, such as reference signals, control information, data communications, or a combination thereof, to the apparatus 1308. In some aspects, one or more other components of the apparatus 1300 may generate communications and may provide the generated communications to the transmission component 1304 for transmission to the apparatus 1308. In some aspects, the transmission component 1304 may perform signal processing on the generated communications (such as filtering, amplification, modulation, digital-to-analog conversion, multiplexing, interleaving, mapping, or encoding, among other examples) , and may transmit the processed signals to the apparatus 1308. In some aspects, the transmission component 1304 may include one or more antennas, one or more modems, one or more modulators, one or more transmit MIMO processors, one or more transmit processors, one or more controllers / processors, one or more memories, or a combination thereof, of the network node described in connection with Fig. 2. In some aspects, the transmission component 1304 may be co-located with the reception component 1302 in one or more transceivers.
[0171] The communication manager 1306 may support operations of the reception component 1302 and / or the transmission component 1304. For example, the communication manager 1306 may receive information associated with configuring reception of communications by the reception component 1302 and / or transmission of communications by the transmission component 1304. Additionally, or alternatively, the communication manager 1306 may generate and / or provide control information to the reception component 1302 and / or the transmission component 1304 to control reception and / or transmission of communications.
[0172] The reception component 1302 may receive an indication of UE AI / ML model representation or quantization preferences. The transmission component 1304 may transmit an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0173] The reception component 1302 may receive a request for labeled training data. The transmission component 1304 may transmit labeled training data based at least in part on the request wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences. The reception component 1302 may receive a request for labeled training data. The transmission component 1304 may transmit a reject message based at least in part on the request wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences and an exclusion of the labeled training data.
[0174] The transmission component 1304 may transmit an indication of a restriction associated with a UE side AI / ML model parameter representation or quantization. The reception component 1302 may receive an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, wherein the UE AI / ML model representation or quantization scheme is used for AI / ML functionality or model life cycle management. The transmission component 1304 may transmit, prior to a training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model. The reception component 1302 may receive, prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model.
[0175] The transmission component 1304 may transmit an indication in system information regarding AI / ML model representation or quantization techniques supported by a network node, wherein the indication is applicable to a cell or a cell group associated with the system information, or the indication is applicable to a whole network, or the indication is separate for different types of AI / ML models. The transmission component 1304 may transmit a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, wherein the indication is received based at least in part on the request.
[0176] The number and arrangement of components shown in Fig. 13 are provided as an example. In practice, there may be additional components, fewer components, different components, or differently arranged components than those shown in Fig. 13. Furthermore, two or more components shown in Fig. 13 may be implemented within a single component, or a single component shown in Fig. 13 may be implemented as multiple, distributed components. Additionally, or alternatively, a set of (one or more) components shown in Fig. 13 may perform one or more functions described as being performed by another set of components shown in Fig. 13.
[0177] The following provides an overview of some Aspects of the present disclosure:
[0178] Aspect 1: A method of wireless communication performed by a user equipment (UE) , comprising: transmitting an indication of UE artificial learning or machine learning (AI / ML) model representation or quantization preferences; and receiving an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0179] Aspect 2: The method of Aspect 1, wherein a network node side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on labeled training data and the UE AI / ML model representation or quantization preferences.
[0180] Aspect 3: The method of any of Aspects 1-2, wherein the AI / ML model is a quantized AI / ML model and is based at least in part on the UE AI / ML model representation or quantization preferences.
[0181] Aspect 4: The method of any of Aspects 1-3, wherein the UE AI / ML model representation or quantization preferences are based at least in part on a standard definition or a network node configuration.
[0182] Aspect 5: The method of any of Aspects 1-4, wherein the UE AI / ML model representation or quantization preferences indicate one or more of: a number of bits to represent or quantize a single parameter, metrics associated with determining an AI / ML model parameter representation or quantization, an indication of whether the UE prefers a network node to represent or quantize the AI / ML model or whether the UE prefers to represent or quantize the AI / ML model, an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, or components or layer specific AI / ML model parameter representation or quantization preferences.
[0183] Aspect 6: The method of any of Aspects 1-5, further comprising: transmitting a request for labeled training data; and receiving labeled training data based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences.
[0184] Aspect 7: The method of any of Aspects 1-6, further comprising: transmitting a request for labeled training data; and receiving a reject message based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences and an exclusion of the labeled training data.
[0185] Aspect 8: The method of any of Aspects 1-7, further comprising: receiving an indication of a restriction associated with a UE side AI / ML model parameter representation or quantization.
[0186] Aspect 9: The method of any of Aspects 1-8, further comprising: transmitting an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, wherein the UE AI / ML model representation or quantization scheme is used for AI / ML functionality or model life cycle management.
[0187] Aspect 10: The method of any of Aspects 1-9, wherein transmitting the indication of the UE AI / ML model representation or quantization preferences is based at least in part on a user plane protocol, radio resource control (RRC) signaling, a medium access control control element (MAC-CE) , uplink control information (UCI) , or an application layer protocol.
[0188] Aspect 11: The method of any of Aspects 1-10, wherein the indication of the UE AI / ML model representation or quantization preferences is transmitted prior to a training of the AI / ML model.
[0189] Aspect 12: The method of any of Aspects 1-11, further comprising: receiving, prior to a training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model; or transmitting, prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model.
[0190] Aspect 13: The method of any of Aspects 1-12, wherein one or more of a data collection, an AI / ML model training, and an AI / ML model quantization is based at least in part on the UE AI / ML model representation or quantization preferences.
[0191] Aspect 14: The method of any of Aspects 1-13, further comprising: receiving an indication in system information regarding AI / ML model representation or quantization techniques supported by a network node, wherein: the indication is applicable to a cell or a cell group associated with the system information, or the indication is applicable to a whole network, or the indication is separate for different types of AI / ML models.
[0192] Aspect 15: The method of any of Aspects 1-14, further comprising: receiving a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, wherein the indication is transmitted based at least in part on the request.
[0193] Aspect 16: A method of wireless communication performed by a network node, comprising: receiving an indication of user equipment (UE) artificial learning or machine learning (AI / ML) model representation or quantization preferences; and transmitting an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
[0194] Aspect 17: The method of Aspect 16, wherein a network node side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on labeled training data and the UE AI / ML model representation or quantization preferences.
[0195] Aspect 18: The method of any of Aspects 16-17, wherein the AI / ML model is a quantized AI / ML model and is based at least in part on the UE AI / ML model representation or quantization preferences.
[0196] Aspect 19: The method of any of Aspects 16-18, wherein the UE AI / ML model representation or quantization preferences are based at least in part on a standard definition or a network node configuration.
[0197] Aspect 20: The method of any of Aspects 16-19, wherein the UE AI / ML model representation or quantization preferences indicate one or more of: a number of bits to represent or quantize a single parameter, metrics associated with determining a AI / ML model parameter representation or quantization, an indication of whether the UE prefers a network node to represent or quantize the AI / ML model or whether the UE prefers to represent or quantize the AI / ML model, an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, or components or layer specific AI / ML model parameter representation or quantization preferences.
[0198] Aspect 21: The method of any of Aspects 16-20, further comprising: receiving a request for labeled training data; and transmitting labeled training data based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences.
[0199] Aspect 22: The method of any of Aspects 16-21, further comprising: receiving a request for labeled training data; and transmitting a reject message based at least in part on the request, wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences and an exclusion of the labeled training data.
[0200] Aspect 23: The method of any of Aspects 16-22, further comprising: transmitting an indication of a restriction associated with a UE side AI / ML model parameter representation or quantization.
[0201] Aspect 24: The method of any of Aspects 16-23, further comprising: receiving an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, wherein the UE AI / ML model representation or quantization scheme is used for AI / ML functionality or model life cycle management.
[0202] Aspect 25: The method of any of Aspects 16-24, wherein receiving the indication of the UE AI / ML model representation or quantization preferences is based at least in part on a user plane protocol, radio resource control (RRC) signaling, a medium access control control element (MAC-CE) , uplink control information (UCI) , or an application layer protocol.
[0203] Aspect 26: The method of any of Aspects 16-25, wherein the indication of the UE AI / ML model representation or quantization preferences is received prior to a training of the AI / ML model.
[0204] Aspect 27: The method of any of Aspects 16-26, further comprising: transmitting, prior to a training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model; or receiving, prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model.
[0205] Aspect 28: The method of any of Aspects 16-27, wherein one or more of a data collection, an AI / ML model training, and an AI / ML model quantization is based at least in part on the UE AI / ML model representation or quantization preferences.
[0206] Aspect 29: The method o f any of Aspects 16-28, further comprising: transmitting an indication in system information regarding AI / ML model representation or quantization techniques supported by a network node, wherein: the indication is applicable to a cell or a cell group associated with the system information, or the indication is applicable to a whole network, or the indication is separate for different types of AI / ML models.
[0207] Aspect 30: The method of any of Aspects 16-29, further comprising: transmitting a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, wherein the indication is received based at least in part on the request.
[0208] Aspect 31: An apparatus for wireless communication at a device, the apparatus comprising one or more processors; one or more memories coupled with the one or more processors; and instructions stored in the one or more memories and executable by the one or more processors to cause the apparatus to perform the method of one or more of Aspects 1-15.
[0209] Aspect 32: An apparatus for wireless communication at a device, the apparatus comprising one or more memories and one or more processors coupled to the one or more memories, the one or more processors configured to cause the device to perform the method of one or more of Aspects 1-15.
[0210] Aspect 33: An apparatus for wireless communication, the apparatus comprising at least one means for performing the method of one or more of Aspects 1-15.
[0211] Aspect 34: A non-transitory computer-readable medium storing code for wireless communication, the code comprising instructions executable by one or more processors to perform the method of one or more of Aspects 1-15.
[0212] Aspect 35: A non-transitory computer-readable medium storing a set of instructions for wireless communication, the set of instructions comprising one or more instructions that, when executed by one or more processors of a device, cause the device to perform the method of one or more of Aspects 1-15.
[0213] Aspect 36: A device for wireless communication, the device comprising a processing system that includes one or more processors and one or more memories coupled with the one or more processors, the processing system configured to cause the device to perform the method of one or more of Aspects 1-15.
[0214] Aspect 37: An apparatus for wireless communication at a device, the apparatus comprising one or more memories and one or more processors coupled to the one or more memories, the one or more processors individually or collectively configured to cause the device to perform the method of one or more of Aspects 1-15.
[0215] Aspect 38: An apparatus for wireless communication at a device, the apparatus comprising one or more processors; one or more memories coupled with the one or more processors; and instructions stored in the one or more memories and executable by the one or more processors to cause the apparatus to perform the method of one or more of Aspects 16-30.
[0216] Aspect 39: An apparatus for wireless communication at a device, the apparatus comprising one or more memories and one or more processors coupled to the one or more memories, the one or more processors configured to cause the device to perform the method of one or more of Aspects 16-30.
[0217] Aspect 40: An apparatus for wireless communication, the apparatus comprising at least one means for performing the method of one or more of Aspects 16-30.
[0218] Aspect 41: A non-transitory computer-readable medium storing code for wireless communication, the code comprising instructions executable by one or more processors to perform the method of one or more of Aspects 16-30.
[0219] Aspect 42: A non-transitory computer-readable medium storing a set of instructions for wireless communication, the set of instructions comprising one or more instructions that, when executed by one or more processors of a device, cause the device to perform the method of one or more of Aspects 16-30.
[0220] Aspect 43: A device for wireless communication, the device comprising a processing system that includes one or more processors and one or more memories coupled with the one or more processors, the processing system configured to cause the device to perform the method of one or more of Aspects 16-30.
[0221] Aspect 44: An apparatus for wireless communication at a device, the apparatus comprising one or more memories and one or more processors coupled to the one or more memories, the one or more processors individually or collectively configured to cause the device to perform the method of one or more of Aspects 16-30.
[0222] The foregoing disclosure provides illustration and description but is not intended to be exhaustive or to limit the aspects to the precise forms disclosed. Modifications and variations may be made in light of the above disclosure or may be acquired from practice of the aspects.
[0223] As used herein, the term “component” is intended to be broadly construed as hardware and / or a combination of hardware and software. “S oftware” shall be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures, and / or functions, among other examples, whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise. As used herein, a “processor” is implemented in hardware and / or a combination of hardware and software. It will be apparent that systems and / or methods described herein may be implemented in different forms of hardware 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 are described herein without reference to specific software code, since those skilled in the art will understand that software and hardware can be designed to implement the systems and / or methods based, at least in part, on the description herein.
[0224] The hardware and data processing apparatus used to implement the various illustrative logics, logical blocks, modules and circuits described in connection with the aspects disclosed herein may be implemented or performed with a general purpose single-or multi-chip processor, a digital signal processor (DSP) , an application specific integrated circuit (ASIC) , a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, or any conventional processor, controller, microcontroller, or state machine. A processor also may be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. In some aspects, particular processes and methods may be performed by circuitry that is specific to a given function.
[0225] As used herein, “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, or the like.
[0226] 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. Many of these features may be combined in ways not specifically recited in the claims and / or disclosed in the specification. The disclosure of various aspects includes each dependent claim in combination with every other claim in the claim set. As used herein, a phrase referring to “at least one of” a list of items refers to any combination of those items, including single members. 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) .
[0227] No element, act, or instruction used herein should be construed as critical or essential unless explicitly described as such. Also, as used herein, the articles “a” and “an” are intended to include one or more items and may be used interchangeably with “one or more. ” Further, as used herein, the article “the” is intended to include one or more items referenced in connection with the article “the” and may be used interchangeably with “the one or more. ” Furthermore, as used herein, the terms “set” and “group” are intended to include one or more items 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 herein, the terms “has, ” “have, ” “having, ” or the like are intended to be open-ended terms that do not limit an element that they modify (e.g., an element “having” A may also have B) . Further, the phrase “based on” is intended to mean “based, at least in part, on” unless explicitly stated otherwise. Also, as used herein, the term “or” is intended to be inclusive when used in a series and may be used interchangeably with “and / or, ” unless explicitly stated otherwise (e.g., if used in combination with “either” or “only one of” ) .
Claims
1.An apparatus for wireless communication at a user equipment (UE) , comprising:one or more memories; andone or more processors, coupled to the one or more memories, which, individually or in any combination, are operable to cause the apparatus to:transmit an indication of UE artificial learning or machine learning (AI / ML) model representation or quantization preferences; andreceive an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.2.The apparatus of claim 1, wherein a network node side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on labeled training data and the UE AI / ML model representation or quantization preferences.3.The apparatus of any of claims 1-2, wherein the AI / ML model is a quantized AI / ML model and is based at least in part on the UE AI / ML model representation or quantization preferences.4.The apparatus of any of claims 1-3, wherein the UE AI / ML model representation or quantization preferences are based at least in part on a standard definition or a network node configuration.5.The apparatus of any of claims 1-4, wherein the UE AI / ML model representation or quantization preferences indicate one or more of:a number of bits to represent or quantize a single parameter,metrics associated with determining an AI / ML model parameter representation or quantization,an indication of whether the UE prefers a network node to represent or quantize the AI / ML model or whether the UE prefers to represent or quantize the AI / ML model,an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, orcomponents or layer specific AI / ML model parameter representation or quantization preferences.6.The apparatus of any of claims 1-5, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:transmit a request for labeled training data; andreceive labeled training data based at least in part on the request,wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences.7.The apparatus of any of claims 1-6, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:transmit a request for labeled training data; andreceive a reject message based at least in part on the request,wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences and an exclusion of the labeled training data.8.The apparatus of any of claims 1-7, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:receive an indication of a restriction associated with a UE side AI / ML model parameter representation or quantization.9.The apparatus of any of claims 1-8, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:transmit an indication of a UE AI / ML model representation or quantization scheme used by the UE for the AI / ML model, wherein the UE AI / ML model representation or quantization scheme is used for AI / ML functionality or model life cycle management.10.The apparatus of any of claims 1-9, wherein the one or more processors, individually or in any combination, are operable to cause the apparatus to:transmit the indication of the UE AI / ML model representation or quantization preferences based at least in part on a user plane protocol, radio resource control (RRC) signaling, a medium access control control element (MAC-CE) , uplink control information (UCI) , or an application layer protocol.11.The apparatus of any of claims 1-10, wherein the indication of the UE AI / ML model representation or quantization preferences is transmitted prior to a training of the AI / ML model.12.The apparatus of any of claims 1-11, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:receive, prior to a training of the AI / ML model, an indication of a model identifier associated with the AI / ML model and a model structure associated with the AI / ML model; ortransmit, prior to the training of the AI / ML model, the indication of the model identifier associated with the AI / ML model and the model structure associated with the AI / ML model.13.The apparatus of any of claims 1-12, wherein one or more of a data collection, an AI / ML model training, and an AI / ML model quantization is based at least in part on the UE AI / ML model representation or quantization preferences.14.The apparatus of any of claims 1-13, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:receive an indication in system information regarding AI / ML model representation or quantization techniques supported by a network node, wherein:the indication is applicable to a cell or a cell group associated with the system information, or the indication is applicable to a whole network, orthe indication is separate for different types of AI / ML models.15.The apparatus of any of claims 1-14, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:receive a request for UE feedback regarding the UE AI / ML model representation or quantization preferences, wherein the indication is transmitted based at least in part on the request.16.An apparatus for wireless communication at a network node, comprising:one or more memories; andone or more processors, coupled to the one or more memories, which, individually or in any combination, are operable to cause the apparatus to:receive an indication of user equipment (UE) artificial leaming or machine learning (AI / ML) model representation or quantization preferences; andtransmit an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.17.The apparatus of claim 16, wherein a network node side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on labeled training data and the UE AI / ML model representation or quantization preferences.18.The apparatus of any of claims 16-17, wherein the UE AI / ML model representation or quantization preferences indicate one or more of:a number of bits to represent or quantize a single parameter,metrics associated with determining an AI / ML model parameter representation or quantization,an indication of whether the UE prefers a network node to represent or quantize the AI / ML model or whether the UE prefers to represent or quantize the AI / ML model,an indication of whether a same or different AI / ML model parameter representation or quantization scheme is to be applied to weights, bias, or activation, orcomponents or layer specific AI / ML model parameter representation or quantization preferences.19.The apparatus of any of claims 16-18, wherein the one or more processors, individually or in any combination, are operable to further cause the apparatus to:receive a request for labeled training data; andtransmit labeled training data based at least in part on the request,wherein a UE side AI / ML model parameter representation or quantization of the AI / ML model is based at least in part on the labeled training data and the UE AI / ML model representation or quantization preferences.20.A method of wireless communication performed by a user equipment (UE) , comprising:transmitting an indication of UE artificial learning or machine learning (AI / ML) model representation or quantization preferences; andreceiving an AI / ML model, wherein a representation or quantization of one or more parameters associated with the AI / ML model is based at least in part on the UE AI / ML model representation or quantization preferences.
Citation Information
Patent Citations
Reporting quantized user equipment (UE) orientation for beam selection
US20210336682A1
Techniques for adaptive quantization level selection in federated learning
US20220245527A1
Method and apparatus for transmitting channel information based on machine learning
US20230145844A1
Apparatus and method to enhance efficiency of transferrable machine learning models
WO2023143701A1