WO2024211555A1 - Methods for artificial intelligence (ai) / machine learning (ml) model switching - Google Patents

Abstract

A method performed by a wireless transmit / receive unit (WTRU) may comprise: transmitting, to a base station, information indicating an artificial intelligence / machine learning (AI/ML) capability of the WTRU; receiving, from the base station, configuration information including AI/ML model switching restrictions, wherein the model switching restrictions include a first prohibit time period; switching to a new AI/ML model based on the performance of the current AI/ML model and on a condition that a first prohibit timer associated with the current AI/ML model is not running. The method may further comprise starting a second prohibit time period associated with the new AI/ML model, wherein the second prohibit time period is included in the model switching restrictions.

Description

METHODS FOR ARTIFICIAL INTELLIGENCE (Al) / MACHINE LEARNING (ML) MODEL SWITCHING

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 63/494,029, filed April 4, 2023, the contents of which are incorporated herein by reference.

BACKGROUND

[0002] Artificial intelligence may be broadly defined as the behavior exhibited by machines that mimic cognitive functions to sense, reason, adapt, and act.

[0003] Machine learning may refer to type of algorithms that solve a problem based on learning through experience (‘data’), without explicitly being programmed (‘configuring set of rules’). Machine learning may be considered as a subset of Al. Different machine learning paradigms may be envisioned based in the nature of data or feedback available to the learning algorithm. For example, a supervised learning approach may involve learning a function that maps input to an output based on labeled training example, wherein each training example may be a pair consisting of input and the corresponding output. For example, an unsupervised learning approach may involve detecting patterns in the data with no preexisting labels. For example, reinforcement learning approach may involve performing sequence of actions in an environment in order to maximize the cumulative reward. In some solutions, it is possible to apply machine learning algorithms using a combination or interpolation of the above-mentioned approaches. For example, semi-supervised learning approach may use a combination of a small amount of labeled data with a large amount of unlabeled data during training. In this regard semi-supervised learning falls between unsupervised learning (with no labeled training data) and supervised learning (with only labeled training data).

[0004] Deep learning refers to a class of machine learning algorithms that employ artificial neural networks (specifically DNNs) which were loosely inspired from biological systems. The Deep Neural Networks (DNNs) are a special class of machine learning models inspired by the human brain wherein the input is linearly transformed and pass-through non-linear activation function multiple times. DNNs typically consists of multiple layers where each layer consists of linear transformation and a given nonlinear activation functions. The DNNs may be trained using the training data via back-propagation algorithm. Recently, DNNs have shown state-of-the-art performance in a variety of domains, e.g . , speech, vision, natural language etc. and for various machine learning settings supervised, un-supervised, and semi-supervised. [0005] Auto-encoders are specific class of DNNs that arise in context of un-supervised machine learning setting wherein the high-dimensional data is non-linearly transformed to a lower dimensional latent vector using the DNN based encoder and the lower dimensional latent vector is then used to reproduce the high-dimensional data using a non-linear decoder. The encoder is represented as E(x;W_e) where x is the high-dimensional data and W_e represents the parameters of the encoder. The decoder is represented as D(z;W_d) where z is the low-dimensional latent representation and W_d represents the parameters of the encoder. Further, using training data { x_1 ,•• -,x_N} the auto-encoder may be trained by solving the following optimization equation: t^r, ^tr} = arg Eq. 1

[0006] The above equation may be approximately solved using a backpropagation algorithm. The trained encoder E(x;W_e^Atr) may be used to compress the high-dimensional data and trained decoder D(z;W_d^Atr) may be used to decompress the latent representation.

SUMMARY

[0007] A method performed by a wireless transmit / receive unit (WTRU) may comprise: transmitting, to a base station, information indicating an artificial intelligence / machine learning (AI/ML) capability of the WTRU; receiving, from the base station, configuration information including AI/ML model switching restrictions, wherein the model switching restrictions include a first prohibit time period; switching to a new AI/ML model based on the performance of the current AI/ML model and on a condition that a first prohibit timer associated with the current AI/ML model is not running. The method may further comprise starting a second prohibit time period associated with the new AI/ML model, wherein the second prohibit time period is included in the model switching restrictions.

[0008] The AI/ML capability may include enhancements for one or more of a beam management, a channel state information (CSI), a positioning, a mobility management, a CSI compression, a CSI prediction, or a beam prediction. The configuration information may include configuration information associated with the first prohibit time period. The performance of the current AI/ML model may be determined based on a comparison of radio quality measurements to a threshold. The performance of the current AI/ML model is determined based on an indication from a target cell.

BRIEF DESCRIPTION OF THE DRAWINGS

[0009] A more detailed understanding may be had from the following description, given by way of example in conjunction with the accompanying drawings, wherein like reference numerals in the figures indicate like elements, and wherein: [0010] FIG. 1A is a system diagram illustrating an example communications system in which one or more disclosed embodiments may be implemented;

[0011] FIG. 1 B is a system diagram illustrating an example wireless transmit/receive unit (WTRU) that may be used within the communications system illustrated in FIG. 1 A according to an embodiment;

[0012] FIG. 1 C is a system diagram illustrating an example radio access network (RAN) and an example core network (CN) that may be used within the communications system illustrated in FIG. 1A according to an embodiment;

[0013] FIG. 1 D is a system diagram illustrating a further example RAN and a further example CN that may be used within the communications system illustrated in FIG. 1 A according to an embodiment; and [0014] FIG. 2 is a flow chart illustrating an example procedure performed by a WTRU.

DETAILED DESCRIPTION

[0015] FIG. 1A is a diagram illustrating an example communications system 100 in which one or more disclosed embodiments may be implemented. The communications system 100 may be a multiple access system that provides content, such as voice, data, video, messaging, broadcast, etc., to multiple wireless users. The communications system 100 may enable multiple wireless users to access such content through the sharing of system resources, including wireless bandwidth. For example, the communications systems 100 may employ one or more channel access methods, such as code division multiple access (CDMA), time division multiple access (TDMA), frequency division multiple access (FDMA), orthogonal FDMA (OFDMA), single-carrier FDMA (SC-FDMA), zero-tail unique-word discrete Fourier transform Spread OFDM (ZT-UW-DFT-S-OFDM), unique word OFDM (UW-OFDM), resource block-filtered OFDM, filter bank multicarrier (FBMC), and the like.

[0016] As shown in FIG. 1A, the communications system 100 may include wireless transmit/receive units (WTRUs) 102a, 102b, 102c, 102d, a radio access network (RAN) 104, a core network (CN) 106, a public switched telephone network (PSTN) 108, the Internet 110, and other networks 1 12, though it will be appreciated that the disclosed embodiments contemplate any number of WTRUs, base stations, networks, and/or network elements. Each of the WTRUs 102a, 102b, 102c, 102d may be any type of device configured to operate and/or communicate in a wireless environment. By way of example, the WTRUs 102a, 102b, 102c, 102d, any of which may be referred to as a station (STA), may be configured to transmit and/or receive wireless signals and may include a user equipment (UE), a mobile station, a fixed or mobile subscriber unit, a subscription-based unit, a pager, a cellular telephone, a personal digital assistant (PDA), a smartphone, a laptop, a netbook, a personal computer, a wireless sensor, a hotspot or Mi-Fi device, an Internet of Things (loT) device, a watch or other wearable, a head-mounted display (HMD), a vehicle, a drone, a medical device and applications (e.g., remote surgery), an industrial device and applications (e.g., a robot and/or other wireless devices operating in an industrial and/or an automated processing chain contexts), a consumer electronics device, a device operating on commercial and/or industrial wireless networks, and the like. Any of the WTRUs 102a, 102b, 102c and 102d may be interchangeably referred to as a UE.

[0017] The communications systems 100 may also include a base station 114a and/or a base station 114b. Each of the base stations 1 14a, 114b may be any type of device configured to wirelessly interface with at least one of the WTRUs 102a, 102b, 102c, 102d to facilitate access to one or more communication networks, such as the CN 106, the Internet 110, and/or the other networks 112. By way of example, the base stations 114a, 114b may be a base transceiver station (BTS), a NodeB, an eNode B (eNB), a Home Node B, a Home eNode B, a next generation NodeB, such as a gNode B (gNB), a new radio (NR) NodeB, a site controller, an access point (AP), a wireless router, and the like. While the base stations 114a, 114b are each depicted as a single element, it will be appreciated that the base stations 114a, 114b may include any number of interconnected base stations and/or network elements.

[0018] The base station 114a may be part of the RAN 104, which may also include other base stations and/or network elements (not shown), such as a base station controller (BSC), a radio network controller (RNC), relay nodes, and the like. The base station 114a and/or the base station 1 14b may be configured to transmit and/or receive wireless signals on one or more carrier frequencies, which may be referred to as a cell (not shown). These frequencies may be in licensed spectrum, unlicensed spectrum, or a combination of licensed and unlicensed spectrum. A cell may provide coverage for a wireless service to a specific geographical area that may be relatively fixed or that may change over time. The cell may further be divided into cell sectors. For example, the cell associated with the base station 114a may be divided into three sectors. Thus, in one embodiment, the base station 114a may include three transceivers, i.e . , one for each sector of the cell. In an embodiment, the base station 114a may employ multiple-input multiple output (MIMO) technology and may utilize multiple transceivers for each sector of the cell. For example, beamforming may be used to transmit and/or receive signals in desired spatial directions.

[0019] The base stations 114a, 114b may communicate with one or more of the WTRUs 102a, 102b, 102c, 102d over an air interface 116, which may be any suitable wireless communication link (e.g., radio frequency (RF), microwave, centimeter wave, micrometer wave, infrared (IR), ultraviolet (UV), visible light, etc.). The air interface 1 16 may be established using any suitable radio access technology (RAT). [0020] More specifically, as noted above, the communications system 100 may be a multiple access system and may employ one or more channel access schemes, such as CDMA, TDMA, FDMA, OFDMA, SC-FDMA, and the like. For example, the base station 1 14a in the RAN 104 and the WTRUs 102a, 102b, 102c may implement a radio technology such as Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access (UTRA), which may establish the air interface 116 using wideband CDMA (WCDMA). WCDMA may include communication protocols such as High-Speed Packet Access (HSPA) and/or Evolved HSPA (HSPA+). HSPA may include High-Speed Downlink (DL) Packet Access (HSDPA) and/or High-Speed Uplink (UL) Packet Access (HSUPA).

[0021] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a radio technology such as Evolved UMTS Terrestrial Radio Access (E-UTRA), which may establish the air interface 116 using Long Term Evolution (LTE) and/or LTE-Advanced (LTE-A) and/or LTE-Advanced Pro (LTE-A Pro).

[0022] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a radio technology such as NR Radio Access , which may establish the air interface 116 using NR.

[0023] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement multiple radio access technologies. For example, the base station 1 14a and the WTRUs 102a, 102b, 102c may implement LTE radio access and NR radio access together, for instance using dual connectivity (DC) principles. Thus, the air interface utilized by WTRUs 102a, 102b, 102c may be characterized by multiple types of radio access technologies and/or transmissions sent to/from multiple types of base stations (e.g., an eNB and a gNB).

[0024] In otherembodiments, the base station 114a and the WTRUs 102a, 102b, 102c may implement radio technologies such as IEEE 802.11 (i.e., Wireless Fidelity (WiFi), IEEE 802.16 (i.e., Worldwide Interoperability for Microwave Access (WiMAX)), CDMA2000, CDMA2000 1X, CDMA2000 EV-DO, Interim Standard 2000 (IS-2000), Interim Standard 95 (IS-95), Interim Standard 856 (IS-856), Global System for Mobile communications (GSM), Enhanced Data rates for GSM Evolution (EDGE), GSM EDGE (GERAN), and the like.

[0025] The base station 114b in FIG. 1 A may be a wireless router, Home Node B, Home eNode B, or access point, for example, and may utilize any suitable RAT for facilitating wireless connectivity in a localized area, such as a place of business, a home, a vehicle, a campus, an industrial facility, an air corridor (e.g., for use by drones), a roadway, and the like. In one embodiment, the base station 114b and the WTRUs 102c, 102d may implement a radio technology such as IEEE 802.11 to establish a wireless local area network (WLAN). In an embodiment, the base station 114b and the WTRUs 102c, 102d may implement a radio technology such as IEEE 802.15 to establish a wireless personal area network (WPAN). In yet another embodiment, the base station 1 14b and the WTRUs 102c, 102d may utilize a cellular-based RAT (e.g., WCDMA, CDMA2000, GSM, LTE, LTE-A, LTE-A Pro, NR etc.) to establish a picocell or femtocell. As shown in FIG. 1 A, the base station 114b may have a direct connection to the Internet 110. Thus, the base station 114b may not be required to access the Internet 110 via the CN 106. [0026] The RAN 104 may be in communication with the CN 106, which may be any type of network configured to provide voice, data, applications, and/or voice over internet protocol (VoIP) services to one or more of the WTRUs 102a, 102b, 102c, 102d. The data may have varying quality of service (QoS) requirements, such as differing throughput requirements, latency requirements, error tolerance requirements, reliability requirements, data throughput requirements, mobility requirements, and the like. The CN 106 may provide call control, billing services, mobile location-based services, pre-paid calling, Internet connectivity, video distribution, etc., and/or perform high-level security functions, such as user authentication. Although not shown in FIG. 1A, it will be appreciated that the RAN 104 and/or the CN 106 may be in direct or indirect communication with other RANs that employ the same RAT as the RAN 104 or a different RAT. For example, in addition to being connected to the RAN 104, which may be utilizing a NR radio technology, the CN 106 may also be in communication with another RAN (not shown) employing a GSM, UMTS, CDMA 2000, WiMAX, E-UTRA, or WiFi radio technology.

[0027] The CN 106 may also serve as a gateway for the WTRUs 102a, 102b, 102c, 102d to access the PSTN 108, the Internet 110, and/or the other networks 112. The PSTN 108 may include circuit- switched telephone networks that provide plain old telephone service (POTS). The Internet 110 may include a global system of interconnected computer networks and devices that use common communication protocols, such as the transmission control protocol (TCP), user datagram protocol (UDP) and/or the internet protocol (IP) in the TCP/IP internet protocol suite. The networks 112 may include wired and/or wireless communications networks owned and/or operated by other service providers. For example, the networks 112 may include another CN connected to one or more RANs, which may employ the same RAT as the RAN 104 or a different RAT.

[0028] Some or all of the WTRUs 102a, 102b, 102c, 102d in the communications system 100 may include multi-mode capabilities (e.g., the WTRUs 102a, 102b, 102c, 102d may include multiple transceivers for communicating with different wireless networks over different wireless links). For example, the WTRU 102c shown in FIG. 1A may be configured to communicate with the base station 114a, which may employ a cellular-based radio technology, and with the base station 114b, which may employ an IEEE 802 radio technology.

[0029] FIG. 1 B is a system diagram illustrating an example WTRU 102. As shown in FIG. 1 B, the WTRU 102 may include a processor 118, a transceiver 120, a transmit/receive element 122, a speaker/microphone 124, a keypad 126, a display/touchpad 128, non-removable memory 130, removable memory 132, a power source 134, a global positioning system (GPS) chipset 136, and/or other peripherals 138, among others. It will be appreciated that the WTRU 102 may include any subcombination of the foregoing elements while remaining consistent with an embodiment.

[0030] The processor 118 may be a general purpose processor, a special purpose processor, a conventional processor, a digital signal processor (DSP), a plurality of microprocessors, one or more microprocessors in association with a DSP core, a controller, a microcontroller, Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs), any other type of integrated circuit (IC), a state machine, and the like. The processor 118 may perform signal coding, data processing, power control, input/output processing, and/or any other functionality that enables the WTRU 102 to operate in a wireless environment. The processor 118 may be coupled to the transceiver 120, which may be coupled to the transmit/receive element 122. While FIG. 1 B depicts the processor 118 and the transceiver 120 as separate components, it will be appreciated that the processor 118 and the transceiver 120 may be integrated together in an electronic package or chip.

[0031] The transmit/receive element 122 may be configured to transmit signals to, or receive signals from, a base station (e.g., the base station 114a) over the air interface 116. For example, in one embodiment, the transmit/receive element 122 may be an antenna configured to transmit and/or receive RF signals. In an embodiment, the transmit/receive element 122 may be an emitter/detector configured to transmit and/or receive IR, UV, or visible light signals, for example. In yet another embodiment, the transmit/receive element 122 may be configured to transmit and/or receive both RF and light signals. It will be appreciated that the transmit/receive element 122 may be configured to transmit and/or receive any combination of wireless signals.

[0032] Although the transmit/receive element 122 is depicted in FIG. 1 B as a single element, the WTRU 102 may include any number of transmit/receive elements 122. More specifically, the WTRU 102 may employ MIMO technology. Thus, in one embodiment, the WTRU 102 may include two or more transmit/receive elements 122 (e.g., multiple antennas) for transmitting and receiving wireless signals over the air interface 116.

[0033] The transceiver 120 may be configured to modulate the signals that are to be transmitted by the transmit/receive element 122 and to demodulate the signals that are received by the transmit/receive element 122. As noted above, the WTRU 102 may have multi-mode capabilities. Thus, the transceiver 120 may include multiple transceivers for enabling the WTRU 102 to communicate via multiple RATs, such as NR and IEEE 802.11 , for example.

[0034] The processor 1 18 of the WTRU 102 may be coupled to, and may receive user input data from, the speaker/microphone 124, the keypad 126, and/or the display/touchpad 128 (e.g., a liquid crystal display (LCD) display unit or organic light-emitting diode (OLED) display unit). The processor 118 may also output user data to the speaker/microphone 124, the keypad 126, and/or the display/touchpad 128. In addition, the processor 118 may access information from, and store data in, any type of suitable memory, such as the non-removable memory 130 and/or the removable memory 132. The nonremovable memory 130 may include random-access memory (RAM), read-only memory (ROM), a hard disk, or any other type of memory storage device. The removable memory 132 may include a subscriber identity module (SIM) card, a memory stick, a secure digital (SD) memory card, and the like. In other embodiments, the processor 118 may access information from, and store data in, memory that is not physically located on the WTRU 102, such as on a server or a home computer (not shown).

[0035] The processor 118 may receive power from the power source 134, and may be configured to distribute and/or control the power to the other components in the WTRU 102. The power source 134 may be any suitable device for powering the WTRU 102. For example, the power source 134 may include one or more dry cell batteries (e.g., nickel-cadmium (NiCd), nickel-zinc (NiZn), nickel metal hydride (NiMH), lithium-ion (Li-ion), etc.), solar cells, fuel cells, and the like.

[0036] The processor 118 may also be coupled to the GPS chipset 136, which may be configured to provide location information (e.g., longitude and latitude) regarding the current location of the WTRU 102. In addition to, or in lieu of, the information from the GPS chipset 136, the WTRU 102 may receive location information over the air interface 116 from a base station (e.g., base stations 114a, 114b) and/or determine its location based on the timing of the signals being received from two or more nearby base stations. It will be appreciated that the WTRU 102 may acquire location information by way of any suitable location-determination method while remaining consistent with an embodiment.

[0037] The processor 118 may further be coupled to other peripherals 138, which may include one or more software and/or hardware modules that provide additional features, functionality and/or wired or wireless connectivity. For example, the peripherals 138 may include an accelerometer, an e-compass, a satellite transceiver, a digital camera (for photographs and/or video), a universal serial bus (USB) port, a vibration device, a television transceiver, a hands free headset, a Bluetooth® module, a frequency modulated (FM) radio unit, a digital music player, a media player, a video game player module, an Internet browser, a Virtual Reality and/or Augmented Reality (VR/AR) device, an activity tracker, and the like. The peripherals 138 may include one or more sensors. The sensors may be one or more of a gyroscope, an accelerometer, a hall effect sensor, a magnetometer, an orientation sensor, a proximity sensor, a temperature sensor, a time sensor; a geolocation sensor, an altimeter, a light sensor, a touch sensor, a magnetometer, a barometer, a gesture sensor, a biometric sensor, a humidity sensor and the like.

[0038] The WTRU 102 may include a full duplex radio for which transmission and reception of some or all of the signals (e.g . , associated with particular subframes for both the UL (e.g . , for transmission) and DL (e.g., for reception) may be concurrent and/or simultaneous. The full duplex radio may include an interference management unit to reduce and or substantially eliminate self-interference via either hardware (e.g., a choke) or signal processing via a processor (e.g., a separate processor (not shown) or via processor 1 18). In an embodiment, the WTRU 102 may include a half-duplex radio for which transmission and reception of some or all of the signals (e.g., associated with particular subframes for either the UL (e.g., for transmission) or the DL (e.g., for reception)).

[0039] FIG. 1 C is a system diagram illustrating the RAN 104 and the CN 106 according to an embodiment. As noted above, the RAN 104 may employ an E-UTRA radio technology to communicate with the WTRUs 102a, 102b, 102c over the air interface 116. The RAN 104 may also be in communication with the CN 106.

[0040] The RAN 104 may include eNode-Bs 160a, 160b, 160c, though it will be appreciated that the RAN 104 may include any number of eNode-Bs while remaining consistent with an embodiment. The eNode-Bs 160a, 160b, 160c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, 102c over the air interface 116. In one embodiment, the eNode-Bs 160a, 160b, 160c may implement MIMO technology. Thus, the eNode-B 160a, for example, may use multiple antennas to transmit wireless signals to, and/or receive wireless signals from, the WTRU 102a.

[0041] Each of the eNode-Bs 160a, 160b, 160c may be associated with a particular cell (not shown) and may be configured to handle radio resource management decisions, handover decisions, scheduling of users in the UL and/or DL, and the like. As shown in FIG. 1 C, the eNode-Bs 160a, 160b, 160c may communicate with one another over an X2 interface.

[0042] The CN 106 shown in FIG. 1 C may include a mobility management entity (MME) 162, a serving gateway (SGW) 164, and a packet data network (PDN) gateway (PGW) 166. While the foregoing elements are depicted as part of the CN 106, it will be appreciated that any of these elements may be owned and/or operated by an entity other than the CN operator.

[0043] The MME 162 may be connected to each of the eNode-Bs 162a, 162b, 162c in the RAN 104 via an S1 interface and may serve as a control node. For example, the MME 162 may be responsible for authenticating users of the WTRUs 102a, 102b, 102c, bearer activation/deactivation, selecting a particular serving gateway during an initial attach of the WTRUs 102a, 102b, 102c, and the like. The MME 162 may provide a control plane function for switching between the RAN 104 and other RANs (not shown) that employ other radio technologies, such as GSM and/or WCDMA.

[0044] The SGW 164 may be connected to each of the eNode Bs 160a, 160b, 160c in the RAN 104 via the S1 interface. The SGW 164 may generally route and forward user data packets to/from the WTRUs 102a, 102b, 102c. The SGW 164 may perform other functions, such as anchoring user planes during inter-eNode B handovers, triggering paging when DL data is available for the WTRUs 102a, 102b, 102c, managing and storing contexts of the WTRUs 102a, 102b, 102c, and the like.

[0045] The SGW 164 may be connected to the PGW 166, which may provide the WTRUs 102a, 102b, 102c with access to packet-switched networks, such as the Internet 110, to facilitate communications between the WTRUs 102a, 102b, 102c and IP-enabled devices.

[0046] The CN 106 may facilitate communications with other networks. For example, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to circuit-switched networks, such as the PSTN 108, to facilitate communications between the WTRUs 102a, 102b, 102c and traditional land-line communications devices. For example, the CN 106 may include, or may communicate with, an IP gateway (e.g., an IP multimedia subsystem (IMS) server) that serves as an interface between the CN 106 and the PSTN 108. In addition, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to the other networks 112, which may include other wired and/or wireless networks that are owned and/or operated by other service providers.

[0047] Although the WTRU is described in FIGS. 1 A-1 D as a wireless terminal, it is contemplated that in certain representative embodiments that such a terminal may use (e.g., temporarily or permanently) wired communication interfaces with the communication network.

[0048] In representative embodiments, the other network 112 may be a WLAN.

[0049] A WLAN in Infrastructure Basic Service Set (BSS) mode may have an Access Point (AP) for the BSS and one or more stations (STAs) associated with the AP. The AP may have access or an interface to a Distribution System (DS) or another type of wired/wireless network that carries traffic in to and/or out of the BSS. Traffic to STAs that originates from outside the BSS may arrive through the AP and may be delivered to the STAs. Traffic originating from STAs to destinations outside the BSS may be sent to the AP to be delivered to respective destinations. Traffic between STAs within the BSS may be sent through the AP, for example, where the source STA may send traffic to the AP and the AP may deliver the traffic to the destination STA. The traffic between STAs within a BSS may be considered and/or referred to as peer-to-peer traffic. The peer-to-peer traffic may be sent between (e.g., directly between) the source and destination STAs with a direct link setup (DLS). In certain representative embodiments, the DLS may use an 802.11 e DLS or an 802.11z tunneled DLS (TDLS). A WLAN using an Independent BSS (IBSS) mode may not have an AP, and the STAs (e.g., all of the STAs) within or using the IBSS may communicate directly with each other. The IBSS mode of communication may sometimes be referred to herein as an “ad-hoc” mode of communication. [0050] When using the 802.11ac infrastructure mode of operation or a similar mode of operations, the AP may transmit a beacon on a fixed channel, such as a primary channel. The primary channel may be a fixed width (e.g . , 20 MHz wide bandwidth) or a dynamically set width. The primary channel may be the operating channel of the BSS and may be used by the STAs to establish a connection with the AP. In certain representative embodiments, Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) may be implemented, for example in 802.11 systems. For CSMA/CA, the STAs (e.g., every STA), including the AP, may sense the primary channel. If the primary channel is sensed/detected and/or determined to be busy by a particular STA, the particular STA may back off. One STA (e.g., only one station) may transmit at any given time in a given BSS.

[0051] High Throughput (HT) STAs may use a 40 MHz wide channel for communication, for example, via a combination of the primary 20 MHz channel with an adjacent or nonadjacent 20 MHz channel to form a 40 MHz wide channel.

[0052] Very High Throughput (VHT) STAs may support 20MHz, 40 MHz, 80 MHz, and/or 160 MHz wide channels. The 40 MHz, and/or 80 MHz, channels may be formed by combining contiguous 20 MHz channels. A 160 MHz channel may be formed by combining 8 contiguous 20 MHz channels, or by combining two non-contiguous 80 MHz channels, which may be referred to as an 80+80 configuration. For the 80+80 configuration, the data, after channel encoding, may be passed through a segment parser that may divide the data into two streams. Inverse Fast Fourier Transform (IFFT) processing, and time domain processing, may be done on each stream separately. The streams may be mapped on to the two 80 MHz channels, and the data may be transmitted by a transmitting STA. At the receiver of the receiving STA, the above described operation for the 80+80 configuration may be reversed, and the combined data may be sent to the Medium Access Control (MAC).

[0053] Sub 1 GHz modes of operation are supported by 802.11af and 802.11ah. The channel operating bandwidths, and carriers, are reduced in 802.11af and 802.11ah relative to those used in 802.11n, and 802.11ac. 802.1 1af supports 5 MHz, 10 MHz, and 20 MHz bandwidths in the TV White Space (TVWS) spectrum, and 802.11 ah supports 1 MHz, 2 MHz, 4 MHz, 8 MHz, and 16 MHz bandwidths using non-TVWS spectrum. According to a representative embodiment, 802.11ah may support Meter Type Control/Machine-Type Communications (MTC), such as MTC devices in a macro coverage area. MTC devices may have certain capabilities, for example, limited capabilities including support for (e.g., only support for) certain and/or limited bandwidths. The MTC devices may include a battery with a battery life above a threshold (e.g., to maintain a very long battery life).

[0054] WLAN systems, which may support multiple channels, and channel bandwidths, such as 802.11 n , 802.11 ac, 802.11 af, and 802.11 ah, include a channel which may be designated as the primary channel. The primary channel may have a bandwidth equal to the largest common operating bandwidth supported by all STAs in the BSS. The bandwidth of the primary channel may be set and/or limited by a STA, from among all STAs in operating in a BSS, which supports the smallest bandwidth operating mode. In the example of 802.11 ah , the primary channel may be 1 MHz wide for STAs (e.g., MTC type devices) that support (e.g., only support) a 1 MHz mode, even if the AP, and other STAs in the BSS support 2 MHz, 4 MHz, 8 MHz, 16 MHz, and/or other channel bandwidth operating modes. Carrier sensing and/or Network Allocation Vector (NAV) settings may depend on the status of the primary channel. If the primary channel is busy, for example, due to a STA (which supports only a 1 MHz operating mode) transmitting to the AP, all available frequency bands may be considered busy even though a majority of the available frequency bands remains idle.

[0055] In the United States, the available frequency bands, which may be used by 802.11 ah, are from 902 MHz to 928 MHz. In Korea, the available frequency bands are from 917.5 MHz to 923.5 MHz. In Japan, the available frequency bands are from 916.5 MHz to 927.5 MHz. The total bandwidth available for 802.11 ah is 6 MHz to 26 MHz depending on the country code.

[0056] FIG. 1 D is a system diagram illustrating the RAN 104 and the CN 106 according to an embodiment. As noted above, the RAN 104 may employ an NR radio technology to communicate with the WTRUs 102a, 102b, 102c over the air interface 116. The RAN 104 may also be in communication with the CN 106.

[0057] The RAN 104 may include gNBs 180a, 180b, 180c, though it will be appreciated that the RAN 104 may include any number of gNBs while remaining consistent with an embodiment. The gNBs 180a, 180b, 180c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, 102c over the air interface 1 16. In one embodiment, the gNBs 180a, 180b, 180c may implement MIMO technology. For example, gNBs 180a, 108b may utilize beamforming to transmit signals to and/or receive signals from the gNBs 180a, 180b, 180c. Thus, the gNB 180a, for example, may use multiple antennas to transmit wireless signals to, and/or receive wireless signals from, the WTRU 102a. In an embodiment, the gNBs 180a, 180b, 180c may implement carrier aggregation technology. For example, the gNB 180a may transmit multiple component carriers to the WTRU 102a (not shown). A subset of these component carriers may be on unlicensed spectrum while the remaining component carriers may be on licensed spectrum. In an embodiment, the gNBs 180a, 180b, 180c may implement Coordinated Multi-Point (CoMP) technology. For example, WTRU 102a may receive coordinated transmissions from gNB 180a and gNB 180b (and/or gNB 180c).

[0058] The WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c using transmissions associated with a scalable numerology. For example, the OFDM symbol spacing and/or OFDM subcarrier spacing may vary for different transmissions, different cells, and/or different portions of the wireless transmission spectrum. The WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c using subframe or transmission time intervals (TTIs) of various or scalable lengths (e.g., containing a varying number of OFDM symbols and/or lasting varying lengths of absolute time).

[0059] The gNBs 180a, 180b, 180c may be configured to communicate with the WTRUs 102a, 102b, 102c in a standalone configuration and/or a non-standalone configuration. In the standalone configuration, WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c without also accessing other RANs (e.g., such as eNode-Bs 160a, 160b, 160c). In the standalone configuration, WTRUs 102a, 102b, 102c may utilize one or more of gNBs 180a, 180b, 180c as a mobility anchor point. In the standalone configuration, WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c using signals in an unlicensed band. In a non-standalone configuration WTRUs 102a, 102b, 102c may communicate with/connect to gNBs 180a, 180b, 180c while also communicating with/connecting to another RAN such as eNode-Bs 160a, 160b, 160c. For example, WTRUs 102a, 102b, 102c may implement DC principles to communicate with one or more gNBs 180a, 180b, 180c and one or more eNode-Bs 160a, 160b, 160c substantially simultaneously. In the non-standalone configuration, eNode- Bs 160a, 160b, 160c may serve as a mobility anchor for WTRUs 102a, 102b, 102c and gNBs 180a, 180b, 180c may provide additional coverage and/or throughput for servicing WTRUs 102a, 102b, 102c.

[0060] Each of the gNBs 180a, 180b, 180c may be associated with a particular cell (not shown) and may be configured to handle radio resource management decisions, handover decisions, scheduling of users in the UL and/or DL, support of network slicing, DC, interworking between NR and E-UTRA, routing of user plane data towards User Plane Function (UPF) 184a, 184b, routing of control plane information towards Access and Mobility Management Function (AMF) 182a, 182b and the like. As shown in FIG. 1 D, the gNBs 180a, 180b, 180c may communicate with one another over an Xn interface.

[0061] The CN 106 shown in FIG. 1 D may include at least one AMF 182a, 182b, at least one UPF 184a, 184b, at least one Session Management Function (SMF) 183a, 183b, and possibly a Data Network (DN) 185a, 185b. While the foregoing elements are depicted as part of the CN 106, it will be appreciated that any of these elements may be owned and/or operated by an entity other than the CN operator.

[0062] The AMF 182a, 182b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 104 via an N2 interface and may serve as a control node. For example, the AMF 182a, 182b may be responsible for authenticating users of the WTRUs 102a, 102b, 102c, support for network slicing (e.g., handling of different protocol data unit (PDU) sessions with different requirements), selecting a particular SMF 183a, 183b, management of the registration area, termination of non-access stratum (NAS) signaling, mobility management, and the like. Network slicing may be used by the AMF 182a, 182b in order to customize CN support for WTRUs 102a, 102b, 102c based on the types of services being utilized WTRUs 102a, 102b, 102c. For example, different network slices may be established for different use cases such as services relying on ultra-reliable low latency (URLLC) access, services relying on enhanced massive mobile broadband (eMBB) access, services for MTC access, and the like. The AMF 182a, 182b may provide a control plane function for switching between the RAN 104 and other RANs (not shown) that employ other radio technologies, such as LTE, LTE-A, LTE-A Pro, and/or non-3GPP access technologies such as WiFi.

[0063] The SMF 183a, 183b may be connected to an AMF 182a, 182b in the CN 106 via an N1 1 interface. The SMF 183a, 183b may also be connected to a UPF 184a, 184b in the CN 106 via an N4 interface. The SMF 183a, 183b may select and control the UPF 184a, 184b and configure the routing of traffic through the UPF 184a, 184b. The SMF 183a, 183b may perform other functions, such as managing and allocating UE IP address, managing PDU sessions, controlling policy enforcement and QoS, providing DL data notifications, and the like. A PDU session type may be IP-based, non-IP based, Ethernet-based, and the like.

[0064] The UPF 184a, 184b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 104 via an N3 interface, which may provide the WTRUs 102a, 102b, 102c with access to packet- switched networks, such as the Internet 110, to facilitate communications between the WTRUs 102a, 102b, 102c and IP-enabled devices. The UPF 184, 184b may perform other functions, such as routing and forwarding packets, enforcing user plane policies, supporting multi-homed PDU sessions, handling user plane QoS, buffering DL packets, providing mobility anchoring, and the like.

[0065] The CN 106 may facilitate communications with other networks. For example, the CN 106 may include, or may communicate with, an IP gateway (e.g., an IP multimedia subsystem (IMS) server) that serves as an interface between the CN 106 and the PSTN 108. In addition, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to the other networks 1 12, which may include other wired and/or wireless networks that are owned and/or operated by other service providers. In one embodiment, the WTRUs 102a, 102b, 102c may be connected to a local DN 185a, 185b through the UPF 184a, 184b via the N3 interface to the UPF 184a, 184b and an N6 interface between the UPF 184a, 184b and the DN 185a, 185b.

[0066] In view of FIGs. 1A-1 D, and the corresponding description of FIGs. 1A-1 D, one or more, or all, of the functions described herein with regard to one or more of: WTRU 102a-d, Base Station 114a-b, eNode-B 160a-c, MME 162, SGW 164, PGW 166, gNB 180a-c, AMF 182a-b, UPF 184a-b, SMF 183a-b, DN 185a-b, and/or any other device(s) described herein, may be performed by one or more emulation devices (not shown). The emulation devices may be one or more devices configured to emulate one or more, or all, of the functions described herein. For example, the emulation devices may be used to test other devices and/or to simulate network and/or WTRU functions.

[0067] The emulation devices may be designed to implement one or more tests of other devices in a lab environment and/or in an operator network environment. For example, the one or more emulation devices may perform the one or more, or all, functions while being fully or partially implemented and/or deployed as part of a wired and/or wireless communication network in order to test other devices within the communication network. The one or more emulation devices may perform the one or more, or all, functions while being temporarily implemented/deployed as part of a wired and/or wireless communication network. The emulation device may be directly coupled to another device for purposes of testing and/or performing testing using over-the-air wireless communications.

[0068] The one or more emulation devices may perform the one or more, including all, functions while not being implemented/deployed as part of a wired and/or wireless communication network. For example, the emulation devices may be utilized in a testing scenario in a testing laboratory and/or a nondeployed (e.g., testing) wired and/or wireless communication network in order to implement testing of one or more components. The one or more emulation devices may be test equipment. Direct RF coupling and/or wireless communications via RF circuitry (e.g., which may include one or more antennas) may be used by the emulation devices to transmit and/or receive data.

[0069] The following abbreviations and acronyms may be referred to:

Af Sub-carrier Spacing gNB NR NodeB

ACK Acknowledgement

Al Artificial Intelligence

BLER Block Error Rate

BRS Beam Reference Signal

CA Carrier Aggregation

CB Contention-Based (e.g., access, channel, resource)

CG Cell Group

CoMP Coordinated Multi-Point transmission/reception

CP Cyclic Prefix

CP-OFDM Conventional OFDM (relying on cyclic prefix)

CQI Channel Quality Indicator

CN Core Network (e.g., LTE packet core or NR core)

CRC Cyclic Redundancy Check

CSI Channel State Information

CU Central Unit

D2D Device to Device transmissions (e.g., LTE Sidelink)

DC Dual Connectivity

DCI Downlink Control Information

DL Downlink

DM-RS Demodulation Reference Signal DNN Deep Neural Network

DRB Data Radio Bearer

DU Distributed Unit

EN-DC E-UTRA - NR Dual Connectivity

EPC Evolved Packet Core

FDD Frequency Division Duplexing

FDM Frequency Division Multiplexing

IP Internet Protocol

LCH Logical Channel

LCP Logical Channel Prioritization

LTE Long Term Evolution

MAC Medium Access Control

NACK Negative ACK

MBS Multicast Broadcast System

MBMS Multimedia Broadcast Multicast System

MCG Master Cell Group

MCS Modulation and Coding Scheme

MIMO Multiple Input Multiple Output

ML Machine Learning

MTC Machine-Type Communications

MR-DC Multi-RAT Dual Connectivity

NAS Non-Access Stratum

NE-DC NR-RAN - E-UTRA Dual Connectivity

NR New Radio

NR-DC Dual Connectivity with

NW Network

OFDM Orthogonal Frequency-Division Multiplexing

OOB Out-Of-Band (emissions)

Pcmax Total available WTRU power in a given transmission interval

Pcell Primary cell of Master Cell Group

PCG Primary Cell Group

PDU Protocol Data Unit

PER Packet Error Rate

PHY Physical Layer

PLMN Public Land Mobile Network

PLR Packet Loss Rate

PRACH Physical Random-Access Channel

PRB Physical Resource Block

PScell Primary cell of a Secondary cell group

PSS Primary Synchronization Signal

QoS Quality of Service (from the physical layer perspective)

RAB Radio Access Bearer

RAN PA Radio Access Network Paging Area

RACH Random Access Channel (or procedure)

RAR Random Access Response

RAT Radio Access Technology

RCU Radio access network Central Unit

RF Radio Front end

RLF Radio Link Failure

RLM Radio Link Monitoring RNN Recurrent Neural Network

RNTI Radio Network Identifier

RRC Radio Resource Control

RRM Radio Resource Management

RS Reference Signal

RTT Round-Trip Time

SCG Secondary Cell Group

SCS Sub Carrier Spacing

SCMA Single Carrier Multiple Access

SDU Service Data Unit

SOM Spectrum Operation Mode

SpCell Primary cell of a master or secondary cell group.

SS Synchronization Signal

SSS Secondary Synchronization Signal

SRB Signaling Radio Bearer

SWG Switching Gap (in a self-contained subframe)

TB Transport Block

TBS Transport Block Size

TDD Time-Division Duplexing

TDM Time-Division Multiplexing

Tl Time Interval (in integer multiple of one or more symbols)

TTI T ransmission Time Interval (in integer multiple of one or more symbols)

TRP Transmission / Reception Point

TRPG Transmission / Reception Point Group

TRx Transceiver

UL Uplink

URC Ultra-Reliable Communications

URLLC Ultra-Reliable and Low Latency Communications

V2X Vehicular communications

WLAN Wireless Local Area Networks and related technologies (IEEE 8O2.xx domain)

WTRU Wireless T ransmit / Receive Unit

[0070] The following descriptions are for exemplary purposes and are not intended to limit in any way the applicability of the methods described herein to any wireless technology and/or to other technology, when applicable. The term “network” in this disclosure may refer to one or more gNBs which in turn may be associated with one or more Transmission/Reception Points (TRPs), or to any other node in the radio access network.

[0071] The terms Artificial Intelligence (Al), Machine Learning (ML), Deep Learning (DL), DNNs may be used interchangeably. Methods described herein are exemplified based on learning in wireless communication systems. The methods are not limited to such scenarios, systems and services and may be applicable to any type of transmissions and/or services etc.

[0072] Hereinafter, the terms “AIML model”, “AI/ML model”, “ML model”, “Al model”, “model” may be used interchangeably to refer to an Artificial Intelligence/Machine Learning model that emulates logical decision-making based on available/collected/requested data. [0073] Model switching is one of the LCM processes being considered by RAN2. There may be scenarios where overly frequent model switching prevents the network from monitoring/predicting WTRU behavior. Frequent model switching may increase the signaling overhead and/or latency, for example, due to a WTRU’s indication of model switching to the network. Overly frequent model switching may also impact WTRU performance. As such, the network may want to control the number of times model switching is allowed at the WTRU.

[0074] A WTRU may send a capability indication to the network. In one example, the capability indication may be a simple (e.g., one bit) indication to convey AI/ML capability. The AI/ML capability indication may be on a higher granularity. In one example the, the AI/ML capability indication may be feature/functionality level, which, for example, may include AI/ML capability for enhancements for beam management (e.g., higher accuracy and lower overhead compare); CSI; positioning; and/or mobility management.

[0075] In another example, the AI/ML capability indication may be sub-feature/sub-functionality level, which, for example, may include: CSI compression; CSI prediction; temporal beam prediction; and/or spatial beam prediction.

[0076] In another example, the AI/ML capability indication may be sub-sub-feature/functionality level, which, for example, may include: temporal beam prediction for up to a certain time in the future; spatial beam prediction for certain frequency ranges; and/or CSI compression for up to a certain compression ratio.

[0077] In one example, the granularity of AI/ML capability reporting, whether on a (sub)- functionality/feature level and definition of the functionality/feature, may be determined by the network and sent to the WTRU as part of configuration information (e.g., in terms of a bitmap/mapping table during RRC). In one example, the granularity of AI/ML capability reporting may be the same across all gNBs/operators and the corresponding configuration may be sent to the WTRU by the serving gNB or hard-coded in the WTRU. In one example, the granularity of AI/ML capability reporting may be cellspecific, such that the WTRU may receive the configuration information following every change of cell. The WTRU may also receive configuration specific to a group of cells, a ‘designated’ area, etc., in which case the WTRU may receive information pertaining to that group of cells (e.g., in terms of cell IDs, or area ID). In the case of ‘designated area’ the WTRU may be preconfigured with the location/perimeter/demarcation lines of the area.

[0078] For example, when in cell A, the WTRU may be configured to perform AI/ML capability reporting for CSI enhancements and while in cell B, the WTRU may be configured to perform AI/ML capability reporting specific to CSI compression enhancements or CSI prediction enhancements. [0079] For example, when in designated area C, the WTRU may be configured to perform AI/ML capability reporting for beam management enhancements and while in cell C, the WTRU may be configured to perform AI/ML capability reporting specific to the scenarios (e.g., spatial domain or temporal domain beam prediction).

[0080] A WTRU may receive, from the network, configuration information for model switching. In some instances, a WTRU may not be able to perform any autonomous model switching. Model switching or switching back to legacy systems may only be allowed on indication from the network.

[0081] In one example, the indication from the network may be explicit. For example, the WTRU may receive a switch indication from the network to change AI/ML model. For example, the WTRU may receive an indication from the network to switch back to legacy systems.

[0082] In another example, the indication from the network may be implicit. For example, the WTRU may receive configuration information to switch to another AI/ML model if performance of the current AI/ML model falls below X, where X may be any standardized or non-standardized KPI used to assess AI/ML model performance (e.g., throughput, NMSE, cosine similarity, beam accuracy (e.g., RSRP, RSRP difference, etc.). In a case where the network is performing performance monitoring, reception of a performance report from the network indicating a performance below X may trigger the WTRU to switch to another AI/ML model. In one example, the switching indication from the network may be accompanied with information on which model to switch to, for example, accompanied with a model ID (local or global model ID). In one example, the switching indication from the network may be accompanied with metadata information on the model to switch to. For example, accompanied with an indication to switch to a different type of model, for example, from DNN to RNN model, in which case, the WTRU may switch to an RNN model performing the same function as the DNN function was performing. For example, the WTRU may receive configuration information to switch to legacy systems if performance of the current AI/ML model falls below Y, where Y may be any standardized or non-standardized KPI used to assess AI/ML model performance (e.g., throughput, NMSE, cosine similarity, beam accuracy (e.g., RSRP, RSRP difference, etc.). In the case where the network is doing performance monitoring, reception of a performance report from the network indicating a performance below Y may trigger the WTRU to switch to legacy systems.

[0083] In some examples, a WTRU may be able to perform autonomous model switching, in which case, it may receive, from the network, the corresponding configuration information for autonomous model switching.

[0084] The WTRU may be able to make any switching decision. The WTRU may perform blind switching, which is not based on performance monitoring. For example, the WTRU may be configured and/or receive more than one AI/ML model and the WTRU may periodically switch between the models. In one example, the WTRU may switch between two (or more) models based on the latency of the model to produce an output. For example, if Model A takes a shorter time to produce an output during inference compared to Model B, the WTRU may switch to Model A. If one AI/ML model is not good enough, the WTRU may switch to another model.

[0085] The WTRU may perform performance monitoring of its AI/ML models and perform model switching based on the monitoring. In one example, the criteria for performance monitoring may be hard- coded in the WTRU or received by the WTRU from the WTRU vendor for model monitoring purposes for example. In one example, if the performance of the WTRU is lower than a threshold configured by the WTRU vendor, the WTRU may deactivate the model. Performance metrics configured by the WTRU vendor may be KPIs such as model inference latency, model training latency, etc. For example, if it takes longer than a certain amount of time to train/retrai n/fine-tune the model, the WTRU may either deactivate or permanently retire that model and switch to another model. In one example, the conditions related to performance monitoring may be configured by the network, as will be described in the next section.

[0086] The WTRU may perform autonomous model switching to any AI/ML model based on an indication from the network. For example, the WTRU may receive a simple indication from the network that the throughput of the WTRU is low. Based on this indication, the WTRU may autonomously select another AI/ML model and switch to the other model.

[0087] The WTRU may be able to perform conditional model switching decisions, based on one or more of features/functionalities and/or sub features/functionalities. For example, the WTRU may be able to perform autonomous model switching only for CSI prediction, not for CSI compression. For example, the WTRU may be able to perform autonomous model switching for beam management (e.g ., best beam prediction), not for CSI enhancements, etc.

[0088] For example, a WTRU may make autonomous switching decisions only within different versions of the same model. Different versions of the same model may be the result of training/retraining/fine- tuning of the model. Different versions of the same model may be the result of training/retraining/fine- tuning on a different data set. For example, in one example, the model may be trained with a fixed set of measurement beams/beam pairs for beam management (set B) resulting in one version. The same model may be trained with a random/mixed set of measurement beams/beam pairs (different set B) resulting in another version of the model.

[0089] A WTRU may make autonomous switching decisions only within a designated group of AI/ML models. The group of AI/ML models may be indicated by the network to the WTRU via model ID. For example, the WTRU receives an indication that it may autonomously switch between models with model IDs X, Y, Z. The group of AI/ML models may be indicated by the network to the WTRU via a part of model ID. For example, the WTRU may receive an indication that it can autonomously switch between models with model IDs X-1 , X-2, and X-3. X may be a prefix corresponding to a reference ID within the model ID or it may correspond to some model metadata. The group of AI/ML models may correspond to any AI/ML received/downloaded within a certain time window.

[0090] A WTRU may make autonomous switching decisions based on miscellaneous conditions related to, for example, mobility state, speed at which WTRU is moving, radio conditions of current or neighbor cells, battery level, location of WTRU within the cell, etc. The WTRU may be allowed to switch AI/ML model only within certain conditions (e.g., only if battery level of WTRU > X%). The WTRU may not be allowed to switch AI/ML models if certain conditions are met (e.g., WTRU cannot switch is radio conditions of current cell is below a quality threshold, e.g., below a RSRP). The WTRU may perform autonomous model switching when in some locations within the cell. However, at cell edge, the WTRU has to validate any model switching decision with the network, or the WTRU has to fall back to legacy. For example, the WTRU may perform autonomous model switching if the WTRU is stationary. If the WTRU becomes mobile or if the WTRU mobility is changing at a rate faster than a threshold, the WTRU is no longer allowed to perform autonomous model switching.

[0091] A WTRU may make autonomous switching decisions only following network confirmation/validation. For example, any model switching decision made by the WTRU has to be validated by the network first before the WTRU may actually perform the model switching. For example, if the WTRU determines to perform a model switch based on performance monitoring of the model, the network may have to confirm/validate that the performance is indeed poor enough to warrant a model switch. For example, the WTRU may make switching decisions for different models, but only the network may indicate which version of the model to use.

[0092] A WTRU may make autonomous switching decisions following model performance monitoring. For example, the WTRU may perform model switching if it determines the performance of the current model is poor, or receives an indication from the network that the performance of the current model is poor. The WTRU may assess the performance of the model against a single or multi-KPI metric. For example, for a beam prediction model, if the RSRP of the ‘best’ predicted beam is below a threshold X, the WTRU may switch to another model. For example, for a CSI prediction model, if the CQI is below a threshold Y and the BLER is above a threshold Z, the WTRU may switch to another model. Following reception of an indication of poor performance from the network, the WTRU may perform model switching.

[0093] A WTRU may make autonomous switching decisions based on information on other/alternate candidate AI/ML models. The WTRU may make switching decision based on information on other AI/ML models, either at the WTRU or that the WTRU has the option to download from the network. Information on other candidate models may include: performance and or expected performance of other/alternate AI/ML models; indication that another AI/ML model may be more suitable to current radio conditions; historical information from use of that model at a previous time; training stage/state of other AI/ML models, for example, if another model has been trained for a longer time (and/or for a larger number of iterations), the WTRU may assume higher accuracy during model inference and switch to that model; model latency to produce an output; model overhead; and/or metadata associated with other AI/ML models.

[0094] In addition to model a switching configuration, the WTRU may receive model switching restrictions from the network. The WTRU may receive, from the network, model switching restrictions related to timing, which may or may not be in addition to the conditions described above.

[0095] In one example, the WTRU may receive prohibit timer configurations that prevent frequent model switching. The prohibit timer may be functionality or sub-functionality based or model-specific. For example, a different timer value may be configured for AI/ML models for CSI enhancements versus AI/ML models for beam management. For example, the timer value may depend on how time-critical the (sub)feature/(sub)functionality/(sub)application is. For example, an AI/ML model for beam prediction may be more time-critical compared to an AI/ML model for SCell prediction. For example, a different timer value may be configured for every AI/ML model. For example, the WTRU may receive the timer value as part of the model and/or model metadata when downloading the model (e.g., from the network). For example, the WTRU may receive the timer value associated the AI/ML model from the network at any time (possibly tagged with the model ID).

[0096] Switching restrictions related to timing may be temporary or may be applied at all times. For example, a prohibit timer may be applied for a certain amount of time or at all time or at all time unless indicated otherwise. For example, during certain times, if network loading is high, it may be harder for the network to monitor the WTRU performance if it does frequent model switching versus other times when the NW loading is average.

[0097] Switching restrictions related to timing may be conditional on any one or more of the conditions described above. For example, the condition may be to apply a prohibit timer t to model A only if radio conditions are good (e.g., above a RSRP threshold). If radio conditions are poor, the WTRU may be allowed to override the prohibit timer and switch to another AI/ML model before the prohibit timer expires or switch to legacy mechanisms before the prohibit timer expires. For example, apply a prohibit timer t1 to model B if radio conditions are good (e.g., RSRP > Th 1 ) and apply a prohibit timer t2 to model B if radio conditions are average (e.g., Th 1 < RSRP < Th 2). [0098] In one example, the WTRU may receive configuration information that it has to switch models with a minimum frequency. For example, the WTRU has to switch model every t s/min/hours etc. In some cases, the WTRU may need to switch model every t s/min/hours, unless it does a new performance assessment of the model.

[0099] In one embodiment, a WTRU may receive a prohibit timer indication from the network at different granularities. For example, the WTRU may receive prohibit timer indication via AI/ML model deactivation command MAC CE. For example, the AI/ML model deactivation command may explicitly indicate the value of prohibit timer. For example, if the value of prohibit timer is absent in the MAC CE, the WTRU may apply a RRC configured value for prohibit timer. For example, if the value of prohibit timer is absent in MAC CE and RRC configuration, the WTRU may assume a value of infinity. For example, the AIML model deactivation command many carry an indication of the AI/ML model for which the prohibit timer applies. For example, if the value of the AI/ML model ID is absent in the MAC CE, the WTRU may assume that the deactivation command and prohibit timer applies to all AI/ML models at the WTRU.

[0100] In one embodiment, the AI/ML model deactivation command may carry an indication of the functionality (e.g., CSI enhancements, CSI compression, CSI prediction, beam prediction, etc.) for which the prohibit timer applies. The WTRU may assume that the timer value applies to any AI/ML model linked to that functionality. In one example, the WTRU may apply this timer value to any model linked to this functionality unless the WTRU is otherwise indicated (e.g., via reception of a different timer value linked to a model ID).

[0101] In an embodiment, a WTRU may receive an activation command for AI/ML model activation from the network via MAC CE. For example, the AI/ML model activation command many carry an indication of the AI/ML model for which the activation command applies. For example, the WTRU may receive an indication related to handling of prohibit timer in the activation command. For example, the activation command indicate explicitly that the WTRU may stop any prohibit timer, if running, and activate the AIML model. For example, the activation command may indicate that the WTRU may activate the AIML model after the expiration of the prohibit timer, if running. The activation command may indicate that the WTRU may consider the AIML model for switching and the actual model activation may happen when one or more model switching conditions and/or model validation conditions described herein are satisfied.

[0102] In an embodiment, a WTRU may receive updated indications of the prohibit timer value, for example, in an activation/deactivation message. An indication to activate an AI/ML model may also carry an updated prohibit timer value to apply to that AI/ML model. [0103] In an embodiment, the WTRU may receive configuration information from the network on how to determine the length of prohibit timers applied to model switching.

[0104] The configuration information from the network on how to determine the length of the prohibit timers applied to model switching may be model parameters. In one example, the model parameters may be static model parameters. For example, the static parameters may include the size of AI/ML model, latency to get an output from model, RS overhead associated with model, and/or training parameters (e.g., training time, number of iterations to achieve convergence, training overhead, dataset size, etc.). If the latency to generate an output from the model is larger, the length of the prohibit timer may be larger accordingly.

[0105] In one example, the model parameters may be dynamic model parameters (e.g., model performance). For example, if the performance of the model is below a performance threshold, the WTRU may configure a shorter length of the prohibit timer to allow switching to a better performing model, or switching to legacy.

[0106] The configuration information from the network on how to determine the length of the prohibit timers applied to model switching may be radio conditions/channel measurements (e.g., SINR, SINR range, RSRP, channel coherence time, channel coherence bandwidth, etc.). For example, in some radio conditions, it may be more acceptable to switch between AI/ML models and the length of the prohibit timer may be shorter.

[0107] The configuration information from the network on how to determine the length of the prohibit timers applied to model switching may be based on loading. For example, on detection of congestion (e.g., via a high number of HARQ NACKs, delay in expected traffic arrival times, etc.,) the WTRU may apply a longer prohibit timer.

[0108] In an embodiment, the prohibit timer may only prevent switching between multiple AI/ML models. In another embodiment, the prohibit timer may also prevent back and forth switching between AI/ML based operation and legacy-based operation.

[0109] In example, the WTRU may be configured to stop using AI/ML methods and switch to legacy methods implicitly via configuration of status of prohibit timer or a value thereof. For example, the WTRU may receive an indication to start the prohibit timers for all the AI/ML models. This may implicitly indicate that the WTRU may not use any AI/ML models (e.g., at least until the prohibit timer expires). For example, the value of prohibit timer may be preconfigured. For example, the default value of the prohibit timer may be infinity. [0110] A WTRU may be configured to measure/report one of more AI/ML models being used for inference. In one example, unless otherwise configured by the network, the WTRU may monitor performance of active AI/ML models by default. In another example, the WTRU may be configured to perform performance monitoring on request from the network. In another example, the WTRU may be configured to perform performance monitoring with a minimum frequency (e.g., an assessment every X s/min) etc. In any case where the WTRU is performing performance monitoring, the WTRU may report the result of the performance monitoring to the network (e.g., in a performance report) and/or it may report the resulting action from the performance monitoring exercise (e.g., a switch to another model). In the case the WTRU is performing performance monitoring for multiple models, the WTRU may send one performance report for each model (possible associated with a model ID and/or metadata about the model being assessed) or one performance report aggregating the assessment of multiple models.

[0111] Performance monitoring may also be done at the network (even if the AI/ML model may be at the WTRU). The network may request that the WTRU report metrics to enable the performance monitoring at the network. For example, the WTRU may have to report the output of the CSI parameters from a CSI prediction model as well as the CSI parameters computed using the traditional CSI framework (using CSI-RS from the network). The network may compare the two to evaluate the model. In the case where the network is performing performance monitoring, the WTRU may receive an indication of the result of the performance monitoring, for example, in the form of one or multiple performance monitoring reports. The WTRU may also receive the resulting recommendation from the performance monitoring at the network (e.g., switch to another model).

[0112] If performance monitoring is done simultaneously (both at the WTRU and the network), the performance monitoring at the network and any resulting request/indication from the network may take precedence over the one at the WTRU.

[0113] A WTRU may switch to another cell/g N B/e N B/g roup of cells/tracking area/any other designated area. The switch to another cell may be the result of WTRU measurements or a request from the network. If the switch is a result of a request from the network and the network knows about the model being used at the WTRU, the request to switch cells from the network may also be accompanied with an indication on whether the current active AI/ML model at the WTRU may be valid for the target cell. Based on the indication received, the WTRU may determine to deactivate the AI/ML model.

[0114] In one example, a WTRU may have received from the network a list of cells/group of cells/tracking areas where the AI/ML model is still valid (e.g., via some cell ID or tracking area ID). Following a switch from one cell to another, the WTRU may determine whether the AI/ML model is valid for the target cell based on the preconfigured list. [0115] In one example, the WTRU may be configured to perform validation of the AI/ML model following every cell change. In another example, the WTRU may perform model validation only following some cell changes (e.g . , only for cells outside a target area or within a target area etc.). The KPIs used by the WTRU to perform model validation may be general KPIs (e.g., throughput, RSRP) or more cellspecific, for example, beam quality (RSRP), NMSE, cosine similarity, difference from ground truth labeling for positioning, etc.

[0116] To perform the model validation, the WTRU may request, from the network, additional information/resources. For example, the WTRU may request for CSI-RS from the network to compute the CSI parameters using the traditional CSI framework. For example, the WTRU may request for beam resources to measure additional beams.

[0117] The AI/ML model at the WTRU may be valid for all (group of) cells/tracking areas or it may be cell-specific/cell group-specific/tracking area-specific. During reception of the model (e.g., from network or external server), the WTRU may receive information on the applicability of the model, for example, the WTRU may receive a list of cells/groups of cells/tracking areas for which the model is valid. In one example, this information or part thereof may be received as metadata associated with the model; in another example, this information or part thereof may be incorporated into a model ID (local or global) received together with the model; in another example, it may be received as separately from the model (e.g., via any L1/L2 signaling). In this case, every time the WTRU is about to switch cell, it may determine whether the AI/ML model would still be valid for the target cell/tracki ng area by checking the preconfigured list of cells/tracking areas. For example, when the WTRU is performing measurements (e.g., RSRP) on the target cell, the WTRU may check whether the model is valid for the target cell (e.g., by checking whether the target cell ID is in the preconfigured list of cells associated to the model). Similarly, the WTRU may check whether the AI/ML model is valid shortly after the cell switch, for example, by checking for the presence of the target cell ID in the preconfigured list.

[0118] The WTRU may validate the AI/ML model using information about the cell. Before transitioning to a new cell or shortly following transition to a new cell, the WTRU may assess the validity of the AI/ML model based on the information for the cell that the WTRU has transitioned to. For example, the AI/ML model may be valid for a type of cell deployment (e.g., small cell, large cell, macro cell, etc.). In another example, the AI/ML model may be valid in certain locations (e.g., urban versus rural deployment).

[0119] Whether or not the cell supports WTRUs using AI/ML. For example, if the target cell does not support WTRUs using AI/ML, the WTRU may switch back to legacy. If the target cell supports WTRUs using AI/ML, the WTRU may assume the AI/ML is valid for the target cell or may check for further validation based on other conditions (e.g., whether AI/ML is supported for specific features/functionalities/scenarios, etc.)

[0120] Whether or not the cell supports WTRUs for certain (sub)functionalities/(sub)features. For example, if the target cell does not support WTRUs using AI/ML for CSI prediction, the WTRU may switch back to legacy CSI framework for CSI estimation. If the target cell supports WTRUs using AI/ML for CSI prediction, the WTRU may assume the AI/ML is valid for the target cell or may check for further validation based on other conditions.

[0121] A WTRU may send a request to the network to validate whether the AI/ML model at the WTRU is valid for the target cell/group of cells/tracking area. The WTRU may send the validation request to the source gNB before the cell switch and/or the target gNB after the cell switch. In one example, the request for model validation sent by the WTRU may be standalone. For example, the WTRU may send a single bit/flag to the target gNB to request for model validation. If the WTRU only has one AI/ML model and/or one AI/ML model per feature/functionality and/or one AI/ML model activated at a time, a simple standalone request for model validation to the network may be sufficient.

[0122] In one example, the request for model validation sent by the WTRU may be accompanied by one or more of the following, which may assist the NW in identifying the model at the WTRU that requires validation: Model ID or part thereof (e.g., global model ID or part thereof, local model ID or part thereof); model metadata; and/or combination of model ID or part thereof and model metadata. The model metadata may include, for example, the (sub)feature/(sub)functionality applicable to the model. For example, the WTRU may send a ‘0’ for CSI compression and a T for CSI prediction in a 1-bit field configured for CSI enhancement models. For example, the WTRU may send a ‘01 ’ for CSI compression with compression ratios > 0.5 and ‘00’ for CSI compression with compression ratios < 0.5, based on configuration at the WTRU.

[0123] To perform the model validation, the network may request from the WTRU for additional information. For example, the network may send CSI-RS to the WTRU and request for the WTRU to compute and transmit the CSI parameters.

[0124] In one example, the network may determine whether or not the AI/ML model at the WTRU is valid for the target cell, unprompted (i.e . , without receiving any request for model validation) by the WTRU. For example, the network may determine that the AI/ML model at the WTRU is a mismatch for the target cell due to a sudden change in some conditions (for example, decline in throughput following the cell switch, drop in RSRP following the cell switch). In one example, the network may determine whether or not the AI/ML model at the WTRU is valid following every cell change. Every time the WTRU changes cell, the network may measure some KPIs (e.g., throughput, RSRP) to determine whether the AI/ML model is valid or not for the new cell. The KPIs used by the network may also be more use case specific, for example, beam quality (RSRP), NMSE, cosine similarity, difference from ground truth labeling for positioning, etc. This may be the case if the network has knowledge about the feature/functionality of the AI/ML model at the WTRU.

[0125] Following the validation step in the case where the WTRU performs the model validation, if the WTRU assesses the model to be suitable for the target cell/group of cells/tracking area, the WTRU may activate/start using the model in the target cell in one example. If there is a prohibit timer associated with the model, the WTRU may also start the associated prohibit timer on activating the model. In another example, the WTRU may need to further validate the assessment of model validity with the target cell before it can activate/start using the model in the target cell. For example, the WTRU may need to send a request with its assessment of model validity to the network and wait for a confirmation/acknowledgement before activating/deploying the model for inference. In another example, the WTRU behavior may be dependent on whether the prohibit timer associated to that model is still running or has expired.

[0126] In one example, if the WTRU determines a model is still valid for a new (target) cell while the prohibit timer associated to that model is still running, the WTRU may keep using the model until expiry of the associated prohibit timer after the switch to the target cell. In one example, if the WTRU determines a model is not valid for a new (target) cell while the prohibit timer associated to that model is still valid, the WTRU may stop using that model and end the associated prohibit timer following transition to the new/target cell. The WTRU may switch to another model which is valid for the new/target cell and/or request to download another model from the network which is valid for the new/target cell and/or switch back to legacy mechanisms.

[0127] In one example, every switch to another (group of) cell/target area may result in the WTRU ending the prohibit timer associated to the AI/ML model that is/was being used with the source cell. The WTRU may validate the applicability of the model against the target cell and if determined to be valid, may reactivate the model (and its associated prohibit timer) following the validation. In one example, the WTRU may only be able to end the prohibit timer associated to the model used in the source cell (and therefore stop using the model) only if some conditions are met. For example, the WTRU may only be able to stop using the model (and end the associated prohibit timer) if a certain amount of time has elapsed in the prohibit timer. For example, at the time of cell switch, if the time elapsed on the prohibit timer is more than halfway through (> 50%), the WTRU may be able to end the prohibit timer and pause/deactivate the use of the model associated with the source cell (for e.g ., until its validation for the target cell is confirmed). If the time elapsed on the prohibit timer is less than halfway through (< 50%), the WTRU may need to keep using the model for longer following switch to the target cell.

[0128] As described in above, a WTRU may send a request for model validation to the network. The WTRU may send the request before and/or after switching to another cell/group of cells NW. In response to the request, the WTRU may receive feedback/information from the NW on applicability of the model. In one example, the WTRU may receive feedback/information from the NW on applicability of the model, unprompted by the WTRU (i.e . , not in response to WTRU request). For example, the NW may determine model validation following any change in cell from the WTRU and/or following a change in KPI (e.g., drop in throughput, change in RSRP, etc.)

[0129] In one example, the WTRU may receive feedback from the NW on the applicability of the model while the prohibit timer associated with the model is still running:

[0130] In one example, the WTRU may receive a positive/ACK response that the model is valid for the target cell as well. In one example, the WTRU may receive a standalone positive response. In another example, the WTRU may receive a positive response which may be accompanied by information on the model (e.g., model ID, model metadata, etc.). In one example, the WTRU may receive a performance report from the NW, which may contain the results of the model validation exercise at the NW, for example, throughput > 90%, NMSE > 95%, etc. The WTRU may be configured to act based on the validation results. If the WTRU receives an indication and/or validation exercise results that the model is valid for the target cell, WTRU actions may include any one or more of the following: the WTRU may keep using the model until the prohibit timer associated with the model expires; the WTRU may keep using the model and restart the prohibit timer associated with the model; the WTRU may keep using the model and confirm with the target gNB whether the same/current prohibit timer value associated with the model is still valid.

[0131] In one example, the WTRU may receive a negative/NACK response that the model is not valid for the target cell. In one example, the WTRU may receive a standalone negative response. In another example, the WTRU may receive a negative response which may be accompanied by information on the model (e.g., model ID, model metadata, etc.). In one example, the WTRU may receive a performance report from the NW, which may contain the results of the model validation exercise at the NW, for example, throughput < 90%, NMSE < 95%, etc. The WTRU may be configured to act based on the validation results. If the WTRU receives an indication and/or validation exercise results that the model is not valid for the target cell, WTRU actions may include any one or more of the following: the WTRU may stop using the model immediately; the WTRU may end the prohibit timer associated to the model; the WTRU may stop using the model only if some condition associated with the model is met. For example, the WTRU stops using the model only if a certain amount of time > a threshold of the prohibit timer associated to the WTRU has elapsed; the WTRU may switch back to legacy mechanism; the WTRU may train/retrain/fine-tune the model and request from the NW to validate the model suitability for the target cell following the training/retraining/fine-tuning; the WTRU may switch to another model; the WTRU may request for another model from the network that is valid for the target gNB; the WTRU may receive another model from the network that is valid for the target gNB (e.g., model T) (the model may be accompanied by the prohibit timer associated to the new model); the WTRU may start using the new model (e.g., model T) and start the prohibit timer associated with model T.

[0132] In another example, the WTRU may receive feedback from the NW on the applicability of the model after the prohibit timer associated to the model has expired.

[0133] In one example, the WTRU may receive a positive/ACK response that the model (e.g., model A) is valid for the target cell as well. In one example, the WTRU may receive a standalone positive response. In another example, the WTRU may receive a positive response which may be accompanied by information on the model (e.g., model ID, model metadata, etc.). In one example, the WTRU may receive a performance report from the NW, which may contain the results of the model validation exercise at the NW, for example, throughput > 90%, NMSE > 95%, etc. The WTRU may be configured to act based on the validation results. If the WTRU receives an indication and/or validation exercise results that the model (e.g., model A) is valid for the target cell, WTRU actions may include any one or more of the following: if the WTRU kept using model A after expiry of the associated prohibit timer, the WTRU may keep on using model A; if the WTRU had switched to another model (e.g., model B) after expiry of the prohibit timer associated to model A or had switched back to legacy systems, the WTRU may start using model A again; if the WTRU had switched to another model (e.g., model B) after expiry of the prohibit timer associated to model A, the WTRU may keep using model B (e.g., until the prohibit timer associated with model B expires); and/or if the WTRU had switched to another model (e.g., model B) after expiry of the prohibit timer associated to model A, the WTRU may decide between model A or model B based on a comparison between the two models (e.g., based on performance history of model A vs model B or based on performance monitoring KPIs).

[0134] In one example, the WTRU may receive a negative/NACK response that the model (e.g., model A) is not valid for the target cell. In one example, the WTRU may receive a standalone negative response. In another example, the WTRU may receive a negative response which may be accompanied by information on the model (e.g., model ID, model metadata, etc.). In one example, the WTRU may receive a performance report from the NW, which may contain the results of the model validation exercise at the NW, for example, throughput < 90%, NMSE < 95%, etc. The WTRU may be configured to act based on the validation results. If the WTRU receives an indication and/or validation exercise results that the model (e.g . , model A) is not valid for the target cell, WTRU actions may include any one or more of the following: the WTRU may temporarily deactivate the model; the WTRU may permanently retire the model; the WTRU may delete the model in its ‘model memory log’; the WTRU may switch to another model; the WTRU may train/retrain/fine-tune the model and request from the NW to validate the model suitability for the target cell following the training/retraining/fine-tu ning.

[0135] Following a model switch, the WTRU may send to the network a simple switch indication (e.g., a one-bit switch indication). In one example, if the WTRU only has two AI/ML models, and the network knows about the two AI/ML models, a simple switch indication may inform the network of the AI/ML model that the WTRU has switched to.

[0136] Following a model switch, the WTRU may send to the network A switch indication per (sub)functionality/(sub)feature. In one example, if the WTRU only has two AI/ML models per (sub)functionality/(sub)feature, and the network knows about the two AI/ML models, a simple switch indication per (sub)fu nctionality/(sub)featu re may inform the network of the AI/ML model that the WTRU has switched to.

[0137] Following a model switch, the WTRU may send to the network a switch indication to another model (e.g., accompanied by metadata and/or model ID or part thereof of the new model).

[0138] Following a model switch, the WTRU may send to the network an acknowledgement/confirmation if the model switch is in response to a request from the network.

[0139] In one embodiment, a WTRU may receive configuration from the network that prevents overly frequent model switching that may impact WTRU performance and the network’s ability to predict WTRU behavior or monitor AI/ML performance.

[0140] The WTRU may send capability indication to a base station (e.g., gNB) including AI/ML capability. The AI/ML capability may be on a higher granularity. For example, The AI/ML capability may be the functionality level or sub-functionality level or any other granularity.

[0141] The WTRU may receive, from the base station, configuration information regarding autonomous model switching including model switching restrictions. For example, the WTRU may receive prohibit timer configurations that prevent frequent switching. The prohibit timer may be functionality or sub-functionality based or model-specific. The prohibit timer value may depend on how time-critical the application is (e.g., beam prediction more time critical than SCell prediction).

[0142] The WTRU may determine performance of the current model to be poor (e.g., below a threshold) or the WTRU may receive an indication from the network that model performance is not good. If the WTRU has a model for which prohibit timer is not running, WTRU may switch to another model and/or starts a prohibit timer for a new model. Otherwise, the WTRU may revert back to legacy behavior. [0143] The WTRU may send an indication to the network on the switch or fallback. The switching may be accompanied with model metadata and/or model ID.

[0144] The WRTU may perform performance monitoring of a new model or receive a performance monitoring report from the network.

[0145] In another embodiment, a WTRU may receive a configuration from the network that prevents overly frequent model switching that may impact WTRU performance and impact the network’s ability to predict WTRU behavior. Following a switch to another cell, the WTRU may validate whether the model is valid to the target cell. Based on the model suitability to the target cell and the status of the prohibit timer associated with the model, the WTRU may determine the next steps.

[0146] The WTRU may send capability indication to the base station (e.g., gNB) including AI/ML capability. Functionality level or sub-functionality level or any other granularity. The WTRU may receive, from the base station, configuration information regarding autonomous model switching including model switching restrictions. For example, the WTRU may receive prohibit timer configurations that prevent frequent switching. The prohibit timer may be functionality or sub-functionality based or model-specific. The prohibit timer value may depend on how time-critical the application is (e.g., beam prediction more time critical than SCell prediction).

[0147] The WTRU may determine performance of the current model to be poor (e.g., below a threshold) or the WTRU may receive an indication from the network that model performance is not good. If WTRU has a model for which prohibit timer is not running, WTRU switches to another model or starts prohibit timer for new model. Otherwise, the WTRU may revert back to legacy.

[0148] While a prohibit timer for the AI/ML model is running, the WTRU may switch to another cell. The WTRU may send a request to a target base station to validate if the AI/ML model is still useful.

[0149] If the receives feedback from network while the prohibit timer is running, the WTRU may receive an ACK (e.g., an indication that model is valid for target cell) or NACK (e.g., an indication that model is not valid for target cell). If the WTRU receives an ACK, the WTRU may use a ML model until the timer expires. If the WTRU receives a NACK, the WTRU may stop using ML model and switches to another ML model or back to legacy.

[0150] If the WTRU receives feedback from the network after the prohibit time has expired, the WTRU may receive an ACK (e.g., an indication that model is valid for target cell) or a NACK (e.g., an indication that model is not valid for target cell). If the WTRU receives an ACK, the WTRU may switch back to the model at a future time. If the WTRU receives an NACK, the WTRU may deactivates the ML model.

[0151] FIG. 2 is a flow chart illustrating an example procedure performed by a WTRU. At 202, the WTRU may transmit, to a base station, information indicating the AI/ML capability of the WTRU. At 204, the WTRU may receive, from the base station, configuration information including AI/ML model switching restrictions. The model switching restrictions may include a first prohibit time period. At 206, the WTRU may switch to a new AI/ML model based on the performance of the current AI/ML model and on a condition that a first prohibit time period associated with the current AI/ML model is not running.

[0152] Although features and elements are described above in particular combinations, one of ordinary skill in the art will appreciate that each feature or element can be used alone or in any combination with the other features and elements. In addition, the methods described herein may be implemented in a computer program, software, or firmware incorporated in a computer-readable medium for execution by a computer or processor. Examples of computer-readable media include electronic signals (transmitted over wired or wireless connections) and computer-readable storage media. Examples of computer-readable storage media include, but are not limited to, a read only memory (ROM), a random access memory (RAM), a register, cache memory, semiconductor memory devices, magnetic media such as internal hard disks and removable disks, magneto-optical media, and optical media such as CD-ROM disks, and digital versatile disks (DVDs). A processor in association with software may be used to implement a radio frequency transceiver for use in a UE, WTRU, terminal, base station, RNC, or any host computer.

Claims

CLAIMS What is Claimed:

1 . A method performed by a wireless transmit / receive unit (WTRU) using a current AI/ML model, the method comprising: transmitting, to a base station, information indicating an artificial intelligence / machine learning (AI/ML) capability of the WTRU; receiving, from the base station, configuration information including AI/ML model switching restrictions, wherein the model switching restrictions include a first prohibit time period; switching to a new AI/ML model based on the performance of the current AI/ML model and on a condition that a first prohibit time period associated with the current AI/ML model is not running.

2. The method of claim 1 , wherein the AI/ML capability includes enhancements for one or more of a beam management, a channel state information (CSI), a positioning, a mobility management, a CSI compression, a CSI prediction, or a beam prediction.

3. The method of claim 1 , wherein the configuration information includes configuration information associated with the first prohibit time period.

4. The method of claim 1 , wherein the performance of the current AI/ML model is determined based on a comparison of radio quality measurements to a threshold.

5. The method of claim 1 , wherein the performance of the current AI/ML model is determined based on an indication from a target cell.

6. The method of claim 1 , further comprising: starting a second prohibit time period associated with the new AI/ML model.

7. The method of claim 6, wherein the model switching restrictions includes the second prohibit time period.

8. A wireless transmit / receive unit (WTRU) using a current AI/ML model, the WTRU comprising: a transceiver; and a processor; wherein the transceiver and processor are configured to: transmit, to a base station, information indicating an artificial intelligence / machine learning (AI/ML) capability of the WTRU; receive, from the base station, configuration information including AI/ML model switching restrictions, wherein the model switching restrictions include a first prohibit time period; switch to a new AI/ML model based on a performance of a current AI/ML model and on a condition that a first prohibit time period associated with the current AI/ML model is not running.

9. The WTRU of claim 8, wherein the AI/ML capability includes enhancements for one or more of a beam management, a channel state information (CSI), a positioning, a mobility management, a CSI compression, a CSI prediction, or a beam prediction.

10. The WTRU of claim 8, wherein the configuration information includes configuration information associated with the first prohibit time period.

11 . The WTRU of claim 8, wherein the performance of the current AI/ML model is determined based on a comparison of radio quality measurements to a threshold.

12. The WTRU of claim 8, wherein the performance of the current AI/ML model is determined based on an indication from a target cell.

13. The WTRU of claim 8, further comprising: starting a second prohibit time period associated with the new AI/ML model.

14. The WTRU of claim 13, wherein the model switching restrictions includes the second prohibit time period.

15. A method performed by a wireless transmit / receive unit (WTRU), the method comprising: transmitting, to a first base station, information indicating an artificial intelligence / machine learning (AI/ML) capability of the WTRU; receiving, from the base station, configuration information including AI/ML model switching restrictions; switching to a second base station based on the performance of a current AI/ML model and on a condition that a first prohibit time period associated with the current AI/ML model is not running; and transmitting, to the second base station, a request to validate the current AI/ML model.

16. The method of claim 15, wherein the AI/ML capability includes enhancements for one or more of a beam management, a channel state information (CSI), a positioning, a mobility management, a CSI compression, a CSI prediction, or a beam prediction.

17. The method of claim 15, wherein the configuration information includes configuration information associated with the first prohibit time period.