Training dataset updates for training datasets partitioned into multiple dataset groups

Abstract

Aspects of the present disclosure relate to training dataset updates. The training dataset is partitioned into a plurality of dataset groups, and each dataset group contains one or more training data points. Each data set group is associated with a first tag and a second tag. The first tag corresponds to a time or time domain related parameter, such as a timestamp or duration. The second tag is at least one of a weight or a value associated with a characteristic of the data set. The training data set is updated based on at least one of the first tag or the second tag, such as by updating a subset of values of the second tag, removing a set of data sets, or adding a new set of data sets to the training data set. Updated information corresponding to the updated training data set may then be sent from one device to another device.

Description

[0001] Related Applications

[0002] This application claims priority to U.S. Patent Application No. 63 / 526,531, filed July 13, 2023, entitled “TRAINING DATASET UPDATES FOR A TRAINING DATASET PARTITIONED INTO MULTIPLE DATASET GROUPS,” the entire disclosure of which is incorporated herein by reference. This application also claims priority to U.S. Patent Application No. 18 / 769,061, filed July 10, 2024, entitled “TRAINING DATASET UPDATES FOR A TRAINING DATASET PARTITIONED INTO MULTIPLE DATASET GROUPS,” the entire disclosure of which is incorporated herein by reference. TECHNICAL FIELD

[0003] The present disclosure relates to wireless communications, and more specifically to partitioning a training dataset into multiple dataset groups. BACKGROUND

[0004] A wireless communication system can include one or more network communication devices, such as base stations, which can support wireless communication for one or more user communication devices, which can be otherwise referred to as user equipment (UE) or other suitable terminology. A wireless communication system can support wireless communication with one or more user communication devices by utilizing resources of the wireless communication system, such as time resources (e.g., symbols, time slots, subframes, frames, etc.) or frequency resources (e.g., subcarriers, carriers, etc.). Additionally, a wireless communication system can support wireless communication across various radio access technologies, including third generation (3G) radio access technologies, fourth generation (4G) radio access technologies, fifth generation (5G) radio access technologies, and other suitable radio access technologies beyond 5G (e.g., sixth generation (6G)).

[0005] Artificial intelligence or machine learning (AI / ML) techniques can be used to implement various aspects of a wireless communication system. An AI / ML system or algorithm can be initially trained to generate certain information. Over time, additional data can be collected and the AI / ML system or algorithm can be retrained based on the additional data. SUMMARY

[0006] The article “a” preceding an element does not, unless otherwise specified, preclude the presence of additional identical elements and the article “an” preceding an element does not, unless otherwise specified, preclude the presence of additional identical elements or of at least one additional different element. The terms “comprise,” “comprises,” “comprising,” “include,” “includes,” and “including” when used in this document are used in their open, non- limiting sense and can be used interchangeably with the term “comprising.” The term “or” as used in this document is used in the inclusive sense, i.e., the term “or” means any one of the items, or any combination of the items, or both. As used herein, the term “if’ can be construed to mean “when” or “upon” or “in response to the occurrence of” in addition to “in the event of” or “upon the condition of.” As used herein, the term “plurality” can include one or more elements.

[0007] Some implementations of the methods and apparatuses described herein can further include transmitting, to network equipment over a physical channel, first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; updating the second label after the transmission of the first signaling; updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset; and transmitting, to the network equipment over the physical channel, second signaling indicating a second training dataset report including updated information corresponding to the updated training dataset.

[0008] Some implementations of the methods and apparatuses described herein can further include receiving, from network equipment over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset, and receiving, from the network equipment over the physical channel, second signaling indicative of a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset.

[0009] Some implementations of the methods and apparatuses described herein can further include transmitting, to a user equipment (UE) over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset, updating the second label after transmission of the first signaling, updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset, and transmitting, to the UE over the physical channel, second signaling indicative of a second training dataset report including updated information corresponding to the updated training dataset.

[0010] Some implementations of the methods and apparatuses described herein can further include receiving, from a user equipment (UE) on a physical channel, first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; and receiving, from the UE on the physical channel, second signaling indicating a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset.

[0011] In some implementations of the method and apparatus described herein, the physical channel is an uplink channel. Additionally or alternatively, the limit on the number of the plurality of dataset groups is a maximum value for the number of dataset groups. Additionally or alternatively, the time or time-domain related parameter is at least one of a timestamp or a duration. Additionally or alternatively, the first label is a duration that includes a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a timestamp corresponding to one of a transmission time of the data points of the dataset group, or a collection time of the data points of the dataset group, or a combination thereof. Additionally or alternatively, the weight of the dataset group is selected from a codebook of values associated with the weight. Additionally or alternatively, the weight of the dataset group is updated based on an event, and the event is one of periodic or semi-persistent based on a configuration for updating the dataset, triggered by at least one of a network configuration signal, a downlink control information, or a medium access control control element (MAC-CE) signal, or a combination thereof. Additionally or alternatively, the training dataset is updated by replacing dataset groups associated with smaller values of the weight with dataset groups associated with larger values of the weight. Additionally or alternatively, dataset groups associated with timestamps corresponding to more recent values are replaced with dataset groups associated with timestamps corresponding to earlier values. Additionally or alternatively, the data set points are associated with dataset groups of the plurality of dataset groups based on one or more characteristics of the data set points. Additionally or alternatively, the data set points are associated with the dataset groups of the plurality of dataset groups based on an artificial intelligence (AI) based model maintenance process. Additionally or alternatively, a characteristic of the one or more characteristics of the data set points is an observable characteristic derived via at least one of a deterministic formula for a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point relative to one or more values of other data set points. Additionally or alternatively, the characteristic of the one or more characteristics of the data set points is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set.Additionally or alternatively, the training dataset corresponds to at least one of channel state information (CSI), precoding information, or beam-based information, and wherein a first dataset group of the plurality of dataset groups is associated with one or more of a timestamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first dataset group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger relevance of a dataset point of the first dataset group to the CSI and a smaller value corresponds to a weaker relevance of a dataset point of the first dataset group to the CSI; or is classified based on an observable property of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a precoding matrix singular value to a minimum value of the singular value, a power delay profile associated with the CSI, or an unobservable property of the CSI based on the observable property. Additionally or alternatively, the observable property is one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag corresponding to whether a channel associated with the CSI is a line-of-sight (LoS) or a non-line-of-sight (NLoS) channel, the dominant basis indices corresponding to indices with a minimum power threshold. Additionally or alternatively, the training dataset corresponds to positioning information, and wherein a first data group of a plurality of data groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first dataset group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger relevance of a dataset point of the first dataset group to an actual position and a smaller value corresponds to a weaker relevance of a dataset point of the first dataset group to the actual position; or is classified based on an observable property of the position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or an unobservable property of the actual position based on the observable property. Additionally or alternatively, the observable property is one or more of a value based on the angle of arrival, the angle of departure, the round trip time, or the time difference of arrival, a flag corresponding to whether a channel associated with the positioning information is a indoor or an outdoor UE.Additionally or alternatively, the training data set corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval in which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a greater value corresponds to a stronger correlation of data set points of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of data set points of the first data set group to the heuristic cell association or selection; or is classified based on observable characteristics of the UE mobility including one or more of reference signal received power (RSRP), signal to interference and noise ratio (SINR), beam-based information, channel state information (CSI), or unobservable characteristics of the mobility information based on the observable characteristics. Additionally or alternatively, the observable characteristics are flags as to whether the UE is associated with a best cell based on one or more of a value of the RSRP, a value of the SINR, beam-based information, or CSI.

[0012] In some implementations of the methods and apparatuses described herein, the physical channel is a downlink channel. Additionally or alternatively, the event is triggered by at least one of a UE-based signal on uplink control information or multiplexed with a CSI report. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1 An example of a wireless communication system in accordance with aspects of the present disclosure is illustrated.

[0014] Figure 2 An example of training data set updates in accordance with aspects of the present disclosure is illustrated.

[0015] Figures 3 to 6 An example of a matrix related to channel state information reporting using mixed reference signal types is illustrated.

[0016] Figure 7 An example of an aperiodic trigger state defining a list of CSI report settings is illustrated.

[0017] Figure 8 An example of an aperiodic trigger state is illustrated.

[0018] Figure 9 An example of RRC configuration of non-zero power (NZP) CSI reference signal (RS) resources is illustrated.

[0019] Figure 10 An example of RRC configuration of CSI interference management (IM) resources is illustrated.

[0020] Figure 11 An example of a CSI report is described.

[0021] Figure 12 An example of a UE according to aspects of the disclosure is described.

[0022] Figure 13 An example of a processor according to aspects of the disclosure is described.

[0023] Figure 14 An example of a network equipment (NE) according to aspects of the disclosure is described.

[0024] Figure 15 A flowchart of a method performed by a UE according to aspects of the disclosure is described.

[0025] Figure 16 A flowchart of a method performed by a UE according to aspects of the disclosure is described.

[0026] Figure 17 A flowchart of a method performed by a NE according to aspects of the disclosure is described.

[0027] Figure 18 A flowchart of a method performed by a NE according to aspects of the disclosure is described. DETAILED DESCRIPTION

[0028] In Third Generation Partnership Project (3GPP) New Radio (NR) networks, AI / ML techniques are considered a strong candidate for wireless networks. One challenge in supporting AI / ML techniques is that ubiquitous training data for the AI / ML-enabled schemes helps maintain the robustness of the AI / ML techniques to variations in the environment that would cause shifts in real-world measurements compared to measurements within the training dataset. Accordingly, the techniques discussed herein provide a training dataset approach that enables partial updates of the training dataset to enable capturing the latest channel variations without naively ignoring older measurements of dataset points.

[0029] More specifically, the dataset is partitioned into multiple dataset groups. Each dataset group is associated with a different timestamp corresponding to an approximate collection time of the data points of the dataset group, and dataset points of a dataset group associated with a more recent time replace dataset points of a second dataset group associated with an earlier time. The dataset points of the dataset groups and their corresponding labels can be signaled from one communication node to another communication node.

[0030] Each data set group is also associated with different weights corresponding to relevance to actual or real-time data, and the weight values can be signaled from one communication node to another. Based on these weights, one of a plurality of operations is performed by the communication node that possesses the ground truth, e.g., the communication node that collected the data points in the data set group. These operations can include one or more of, for example, updating the value of a tag for a data set partition that has already been shared, adding a new data set partition with a new set of data points, or omitting information corresponding to a data set partition with a minimal weight (e.g., due to memory limitations).

[0031] Based on the auxiliary information, e.g., based on information obtained during the AI / ML model maintenance phase, data points of the data set can be further broken down into multiple data set groups based on characteristics of the data set points, e.g., based on classifier functionality. The grouping of data points in the data set is based on characteristics, which can be observable (e.g., can be computed directly from the values of the data points) or unobservable (e.g., inferred from analyzing characteristics, e.g., statistical characteristics, of a large set of data points).

[0032] The techniques discussed herein provide a training data set approach that enables partial updates of the training data set to enable capturing the latest channel variations without naively ignoring older measurements of data set points. After a configured or fixed period of time, or based on an event, the training data set is not replaced entirely by a new data set. Thus, the techniques discussed herein avoid a situation where some data points in a conventional data set that can have helped improve model training, and can still be relevant to real-time measurements currently obtained, are replaced by a new data set.

[0033] Further, the training data set is not updated by appending new data points to the same data set at every configured or fixed period of time or based on an event. Thus, the techniques discussed herein avoid a situation where the size of the data set grows without limit after each iteration of adding new data points. Further, the techniques discussed herein avoid a situation where the data set is contaminated if the newly added data points are associated with the same weight as the stale data points previously collected.

[0034] Aspects of the disclosure are described in the context of a wireless communication system.

[0035] Figure 1An example of a wireless communication system 100 in accordance with aspects of the present disclosure is illustrated. The wireless communication system 100 can include one or more NEs 102, one or more UEs 104, and a core network (CN) 106. The wireless communication system 100 can support various radio access technologies. In some implementations, the wireless communication system 100 can be a 4G network, such as an LTE network or an LTE-Advanced (LTE-A) network. In some other implementations, the wireless communication system 100 can be an NR network, such as a 5G network, a 5G-Advanced (5G-A) network, or a 5G-Ultra Wideband (5G-UWB) network. In other implementations, the wireless communication system 100 can be a combination of 4G networks and 5G networks, or other suitable radio access technologies, including Institute of Electrical and Electronics Engineers (IEEE) 802.11 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802.20. The wireless communication system 100 can support radio access technologies other than 5G, such as, for example, 6G. Moreover, the wireless communication system 100 can support technologies such as time division multiple access (TDMA), frequency division multiple access (FDMA), or code division multiple access (CDMA).

[0036] The one or more NEs 102 can be dispersed throughout a geographic region to form the wireless communication system 100. One or more of the NEs 102 described herein can be or include or can be referred to as a network node, base station, network element, network function, network entity, radio access network (RAN), NodeB, eNodeB (eNB), next generation NodeB (gNB), or other suitable terminology. The NEs 102 and the UEs 104 can communicate via communication links, which can be wireless or wired connections. For example, the NEs 102 and the UEs 104 can perform wireless communication (e.g., receive signaling, transmit signaling) over a Uu interface.

[0037] The NEs 102 can provide geographic coverage to support service for one or more UEs 104 within a geographic coverage area. For example, the NEs 102 and the UEs 104 can support wireless communication of signals related to service (e.g., voice, video, packet data, messaging, broadcast, etc.) in accordance with one or more radio access technologies. In some implementations, the NEs 102 can be mobile, such as satellites associated with non-terrestrial networks (NTNs). In some implementations, different geographic coverage areas associated with the same or different radio access technologies can overlap, although different geographic coverage areas can be associated with different NEs 102.

[0038] One or more UEs 104 can be dispersed throughout the wireless communication system 100. A UE 104 can include or can be referred to as a remote unit, a mobile device, a wireless device, a remote device, a subscriber device, a transmitter device, a receiver device, or some other suitable terminology. In some implementations, a UE 104 can be referred to as a unit, a station, a terminal, or a client, among other examples. Also or alternatively, a UE 104 can be referred to as an Internet of Things (IoT) device, an Internet of Everything (IoE) device, or a machine type communications (MTC) device, among other examples.

[0039] A UE 104 can be capable of supporting wireless communications directly with other UEs 104 over a communication link. For example, a UE 104 can support wireless communications directly with another UE 104 over a device-to-device (D2D) communication link. In some implementations, the communication link can be referred to as a sidelink, such as in vehicle-to-vehicle (V2V) deployments, vehicle-to-everything (V2X) deployments, or cellular-V2X deployments. For example, a UE 104 can support wireless communications directly with another UE 104 over a PC5 interface.

[0040] The NEs 102 can support communication with the CN 106 or with another NE 102 or both. For example, NEs 102 can interface with other NEs 102 or the CN 106 through one or more backhaul links, such as an SI, N2, N6, or other network interface. In some implementations, the NEs 102 can communicate directly with one another. In some other implementations, the NEs 102 can communicate indirectly with one another (e.g., via the CN 106). In some implementations, one or more NEs 102 can include subcomponents, such as an access network entity, which can be an example of an access node controller (ANC). An ANC can communicate with one or more UEs 104 through one or more other access network transmission entities, which can be referred to as a radio head, a smart radio head, or a transmission-reception point (TRP).

[0041] The CN 106 can support user authentication, access authorization, tracking, connectivity, and other access, routing, or mobility functions. The CN 106 can be an evolved packet core (EPC) or 5G core (5GC), which can include control plane entities (e.g., a mobility management entity (MME), an access and mobility management function (AMF)) that manage access and mobility, and user plane entities (e.g., a serving gateway (S-GW), a packet data network (PDN) gateway (P-GW), or a user plane function (UPF)) that route packets or interconnect to external networks. In some embodiments, the control plane entities can manage non-access stratum (NAS) functions such as mobility, authentication, and bearer management (e.g., for signaling bearers, data bearers, etc.) for one or more UEs 104 served by one or more NEs 102 associated with the CN 106.

[0042] CN 106 can communicate with a packet data network through one or more backhaul links (e.g., via an SI, N2, N6, or other network interface). The packet data network can include an application server. In some embodiments, one or more UEs 104 can communicate with the application server. A UE 104 can establish a session (e.g., a protocol data unit (PDU) session, etc.) with the CN 106 via the NE 102. The CN 106 can route traffic (e.g., control information, data, etc.) between the UE 104 and the application server using the established session (e.g., the established PDU session). The PDU session can be an instance of a logical connection between the UE 104 and the CN 106 (e.g., one or more network functions of the CN 106).

[0043] In the wireless communication system 100, the NEs 102 and the UEs 104 can use resources (e.g., time resources (e.g., symbols, time slots, subframes, frames, etc.) or frequency resources (e.g., subcarriers, carriers)) of the wireless communication system 100 to perform various operations (e.g., wireless communications). In some embodiments, the NEs 102 and the UEs 104 can support different resource structures. For example, the NEs 102 and the UEs 104 can support different frame structures. In some embodiments, such as in 4G, the NEs 102 and the UEs 104 can support a single frame structure. In some other embodiments, such as in 5G and among other suitable radio access technologies, the NEs 102 and the UEs 104 can support various frame structures (i.e., multiple frame structures). The NEs 102 and the UEs 104 can support the various frame structures based on one or more numerologies.

[0044] In the wireless communication system 100, one or more numerologies can be supported, and a numerology can include a subcarrier spacing and a cyclic prefix. A first numerology (e.g., m=0) can be associated with a first subcarrier spacing (e.g., 15 kHz) and a normal cyclic prefix. In some embodiments, the first numerology (e.g., m=0) associated with the first subcarrier spacing (e.g., 15 kHz) can utilize one slot per subframe. A second numerology (e.g., m=l) can be associated with a second subcarrier spacing (e.g., 30 kHz) and a normal cyclic prefix. A third numerology (e.g., m=2) can be associated with a third subcarrier spacing (e.g., 60 kHz) and a normal cyclic prefix or an extended cyclic prefix. A fourth numerology (e.g., m=3) can be associated with a fourth subcarrier spacing (e.g., 120 kHz) and a normal cyclic prefix. A fifth numerology (e.g., m=4) can be associated with a fifth subcarrier spacing (e.g., 240 kHz) and a normal cyclic prefix.

[0045] Intervals of time for resources (e.g., communication resources) can be organized according to frames (also referred to as radio frames). Each frame can have a duration of, for example, 10 milliseconds (ms). In some implementations, each frame can include multiple subframes. For example, each frame can include 10 subframes, and each subframe can have a duration of, for example, 1 ms. In some implementations, each frame can have the same duration. In some implementations, each subframe of a frame can have the same duration.

[0046] Additionally or alternatively, intervals of time for resources (e.g., communication resources) can be organized according to slots. For example, a subframe can include a number (e.g., quantity) of slots. The number of slots in each subframe can also depend on one or more numerologies supported in the wireless communications system 100. For example, first, second, third, fourth, and fifth numerologies (i.e., m = 0, m = 1, m = 2, m = 3, m = 4) associated with respective subcarrier spacings of 15 kHz, 30 kHz, 60 kHz, 120 kHz, and 240 kHz can utilize a single slot per subframe, two slots per subframe, four slots per subframe, eight slots per subframe, and 16 slots per subframe, respectively. Each slot can include a number (e.g., quantity) of symbols (e.g., OFDM symbols). In some implementations, the number (e.g., quantity) of slots of a subframe can depend on the numerology. For a normal cyclic prefix, a slot can include 14 symbols. For an extended cyclic prefix (e.g., applicable for 60 kHz subcarrier spacing), a slot can include 12 symbols. The relationship between the number of symbols per slot, the number of slots per subframe, and the number of slots per frame for normal and extended cyclic prefixes can depend on the numerology. It should be understood that reference to a first numerology (e.g., m = 0) associated with a first subcarrier spacing (e.g., 15 kHz) can be used interchangeably between subframes and slots.

[0047] In the wireless communication system 100, the electromagnetic (EM) spectrum can be partitioned into various classes, bands, channels, and so forth based on frequency or wavelength. By way of example, the wireless communication system 100 can support one or more operating bands, such as frequency range designations FR1 (410-7.125 GHz), FR2 (24.25-52.6 GHz), FR3 (7.125-24.25 GHz), FR4 (52.6-114.25 GHz), FR4a or FR4-1 (52.6-71 GHz), and FR5 (114.25-300 GHz). In some implementations, the NEs 102 and the UEs 104 can perform wireless communication on one or more of the operating bands. In some implementations, FR1 can be used by the NEs 102 and the UEs 104, and other equipment or apparatus, for cellular communications traffic (e.g., control information, data), and FR2 can be used by the NEs 102 and the UEs 104, and other equipment or apparatus, for short range, high data rate capabilities.

[0048] FR1 can be associated with one or more numerologies (e.g., at least three numerologies). For example, FR1 can be associated with a first numerology including 15 kHz subcarrier spacing (e.g., m=0), a second numerology including 30 kHz subcarrier spacing (e.g., m=l), and a third numerology including 60 kHz subcarrier spacing (e.g., m=2). FR2 can be associated with one or more numerologies (e.g., at least 2 numerologies). For example, FR2 can be associated with the third numerology including 60 kHz subcarrier spacing (e.g., m=2), and a fourth numerology including 120 kHz subcarrier spacing (e.g., m=3).

[0049] The NEs 102 and the UEs 104 can include one or more AI / ML systems (which can also be referred to as AI / ML algorithms or AI / ML models). These AI / ML systems are initially trained on a training data set that includes a plurality of training data points. Additional training data points are collected, and the AI / ML systems are retrained based at least in part on the additional data points.

[0050] The training data set is partitioned into a plurality of data set groups, and each data set group includes one or more training data points. Each data set group is associated with a first label and a second label. The first label corresponds to a time or time domain related parameter, such as a timestamp or a duration. The second label is at least one of a weight or a value associated with a characteristic of the data set. The training data set is updated based on at least one of the first label or the second label, such as by updating a subset of the values of the second label, removing a data set group, or adding a new data set group to the training data set. Updated information corresponding to the updated training data set can then be sent from one device to another device (e.g., from the NE 102 to the UE 104, or from the UE 104 to the NE 102).

[0051] Figure 2 An example 200 of training data set updating according to aspects of the disclosure is illustrated. The example 200 illustrates a communication node 202 and a communication node 204. In one or more implementations, the communication node 202 is the NE 102 and the communication node 204 is the UE 104. In one or more implementations, the communication node 202 is the UE 104 and the communication node 204 is the NE 102. In one or more implementations, both the communication node 202 and the communication node 204 are the UE 104. In one or more implementations, both the communication node 202 and the communication node 204 are the NE 102.

[0052] The communication node 202 includes an AL / ML system 206 and the communication node 204 includes an AL / ML system 208. The communication node 202 includes a training data point collection process 210 that collects data points and a training data set update process 212 that updates a training data set based on the collected data points. The training data set is partitioned into a plurality of data set groups, and each data set group includes one or more training data points, as discussed in more detail below. The communication node 202 transmits a training data set report 214 to the communication node 204 that includes the updated data set. The communication node 204 also includes a training data set update process 216. Thus, the training data set at the communication node 204 can be updated in the same manner as the training data set at the communication node 202, allowing the AL / ML system 206 and the AL / ML system 208 to retrain using the same updated training data set.

[0053] The techniques discussed herein support various NR codebook types, such as the NR codebook types discussed in 3GPP Technical Specification (TS) 38.214, “Physical layer procedures for data,” March 2020. An overview of these techniques is as follows.

[0054] One codebook type is the NR Rel-15 Type-II codebook. Assume that the NE 102 is equipped with a two-dimensional (2D) antenna array with N1, N2antenna ports per horizontal and vertical placement polarization and communication occurs over N3precoder matrix indicator (PMI) subbands. A PMI subband consists of a set of resource blocks, each of which consists of a set of subcarriers. In this case, 2N1N2CSI-RS ports are utilized to achieve high-resolution downlink (DL) channel estimation with the NR Rel-15 Type-II codebook. To reduce the uplink (UL) feedback overhead, a discrete Fourier transform (DFT)-based CSI compression in the spatial domain is applied to L dimensions per polarization, where L < N1N2. The indices of 2Ldimensions are later referred to as the spatial domain (SD) basis indices. The magnitude and phase values of the linear combination coefficients for each subband are fed back to the NE 102 as part of the CSI report. The 2N1N2x N3codebook for each layer l takes the form

[0055] ,

[0056] where is a 2N1N2x 2L block-diagonal matrix with two identical diagonal blocks (L < N1N2), i.e.,

[0057] ,

[0058] and is an N1N2x L matrix with columns drawn from a 2D oversampled DFT matrix, as follows.

[0059] ,

[0060] ,

[0061] ,

[0062] ,

[0063] ,

[0064] where the superscript denotes the matrix transpose operation. Note that for the 2D DFT matrix from which the matrix is drawn, an oversampling factor of is assumed. Note that is common across all layers. is a 2L x N3matrix, where the i-th column corresponds to the linear combination coefficients of the 2L beams in the i-th subband. Only the indices of the L selected columns of are reported, along with the oversampling indices that take on values. Note that is independent for different layers.

[0065] One codebook type is the NR Rel-15 Type-II port selection codebook. For the Type-II port selection codebook, only K (where K < 2N1N2) beamformed CSI-RS ports are utilized in the DL transmission in order to reduce complexity. The K x N3 codebook matrix for each layer takes the following form

[0066] .

[0067] Here, follows the same structure as the regular NR Rel-15 Type-II codebook and is layer-specific. is a K x 2L block-diagonal matrix with two identical diagonal blocks, i.e.,

[0068] ,

[0069] and is a K x 2L matrix whose columns are standard unit vectors, i.e., matrix as follows.

[0070] ,

[0071] where is a standard unit vector with 1 at the i-th position. Here, is an RRC parameter taking values {1, 2, 3, 4} under the condition takes values and is reported as part of the UL CSI feedback overhead. is common across all layers.

[0072] Figure 3 An example 300 of matrices related to channel state information reporting with mixed reference signal types is illustrated. For K = 16, L = 4, and = 1, 8 possible implementations of are illustrated in example 300 corresponding to

[0073] Figure 4 An example 400 of matrices related to channel state information reporting with mixed reference signal types is illustrated. When = 2, 4 possible implementations of are illustrated in example 400 corresponding to

[0074] Figure 5 An example 500 of matrices related to channel state information reporting with mixed reference signal types is illustrated. When ​​​= 3, two possible implementations of

[0075] Figure 6 An example 600 illustrating matrices related to channel state information reporting with mixed reference signal types is illustrated in example 600 when = 4, two possible implementations of

[0076] One codebook type is the NR Rel-15 Type-I codebook. The NR Rel-15 Type-I codebook is the baseline codebook for NR with various configurations. The most common utility of the Rel-15 Type-I codebook is a special case of the NR Rel-15 Type-II codebook where for rank indicator (RI) = 1, 2, L = 1, where a phase coupling value is reported for each subband, i.e., is 2 x N3, where the first row is equal to and the second row is equal to In a particular configuration, i.e., wideband reporting. For RI > 2, different beams are used for each pair of layers. The NR Rel-15 Type-I codebook can be depicted as a low-resolution version of the NR Rel-15 Type-II codebook with only spatial beam selection and phase combination per layer pair.

[0077] One codebook type is the NR Rel-16 Type-II codebook. Assume that the NE 102 is equipped with a two-dimensional (2D) antenna array with N1, N2antenna ports per horizontal and vertical placement polarization and communication occurs over N3PMI subbands. A PMI subband consists of a set of resource blocks, each of which consists of a set of subcarriers. In this case, a high-resolution DL channel estimation with the NR Rel-16 Type-II codebook is achieved with 2N1N2N3CSI-RS ports. To reduce the UL feedback overhead, a discrete Fourier transform (DFT)-based CSI compression in the spatial domain is applied to L dimensions per polarization, where L < N1N2. Similarly, an additional compression in the frequency domain is applied, where each beam of the frequency domain precoding vector is transformed to the delay domain using an inverse DFT matrix, and a subset of the delay domain coefficients’ magnitude and phase values are selected and fed back to the NE 102 as part of the CSI report. The 2N1N2x N3codebook per layer takes the form

[0078] ,

[0079] where is a 2N1N2x 2L block-diagonal matrix with two identical diagonal blocks (L < N1N2), i.e.,

[0080] ​​​​ ,

[0081] and is an N1N2xL matrix with columns drawn from a 2D oversampled DFT matrix as follows.

[0082] ,

[0083] ,

[0084] ,

[0085] ,

[0086] ,

[0087] where the superscript denotes the matrix transpose operation. Note that for the 2D DFT matrix from which the matrix is drawn, it is assumed an oversampling factor. Note that is common across all layers. is an N3xM matrix (M

[0088] ,

[0089] .

[0090] Only the indices of the L selected columns of are reported, as well as the oversampling indices taking values . Similarly, for , only the indices of the M selected columns among the pre-defined size-N3DFT matrix are reported. Later, the indices of M dimensions are referred to as the selected frequency domain (FD) basis indices. Thus, L, M represent the equivalent spatial and frequency dimensions after compression, respectively. Finally, the 2LxM matrix represents the linear combination coefficients (LCC) of the spatial and frequency DFT basis vectors. and are selected independently for different layers. As part of the CSI report, the NE 102 is reported the amplitudes and phase values of the approximate fraction of the 2LM available coefficients. Coefficients with zero amplitude values are indicated via a layer-specific bitmap matrix of size 2LxM, where the bitmap matrix whether each of the indicated coefficients has a zero amplitude value, for which the quantized amplitude and phase values do not need to be reported. Since all non-zero coefficients reported within a layer are normalized with respect to the coefficient with the largest amplitude value (the strongest coefficient), for which the amplitude and phase values corresponding to the strongest coefficient are set to 1 and 0, respectively, and thus no other amplitude and phase information of this coefficient is explicitly reported and only the indication of the index of the strongest coefficient per layer is reported. Thus, for single-layer transmission, each layer reports the amplitude and phase values of the strongest coefficient (as well as the index of the selected L, M DFT vector), which results in a significant reduction of the CSI report size compared to reporting the information of 2N1N2xN3-1 coefficients. The amplitude and phase values of the strongest coefficient (as well as the index of the selected L, M DFT vector) per layer, results in a significant reduction of the CSI report size compared to reporting the information of 2N1N2xN3-1 coefficients.

[0091] One codebook type is the NR Rel. 16 Type-II port selection codebook. For the Type-II port selection codebook, only K (with K < 2N1N2) beamformed CSI-RS ports are utilized in the DL transmission in order to reduce complexity. The KxN3 codebook matrix per layer takes the form [1].

[0092] .

[0093] Here, and follow the same structure as the regular NR Rel. 16 Type-II codebook, which is layer-specific. The matrix is a Kx2L block-diagonal matrix with the same structure as in the NR Rel. 15 Type-II port selection codebook.

[0094] One codebook type is the NR Rel. 17 Type-II port selection codebook. The Rel. 17 Type-II port selection codebook follows a similar structure as the Rel. 15 and Rel. 16 port selection codebooks, as follows

[0095] .

[0096] However, unlike the Rel. 15 and Rel. 16 Type-II port selection codebooks, the port selection matrix supports a free selection of K ports, or more precisely, K / 2 ports per polarization out of the N1N2 CSI-RS ports per polarization, i.e., bits are used to identify the K / 2 selected ports per polarization, where this selection is common across all layers. Here, and follow the same structure as the regular NR Rel. 16 Type-II codebook, however, M is limited to 1, 2, where the network configures windows of size N = {2, 4} for M = 2. Furthermore, unless = 1, otherwise report bitmap and UE reports all coefficients up to rank of value 2.

[0097] One codebook type is the NR Rel-18 Type-II codebook. For the Rel-18 potential Type-II codebook, the time domain corresponding to a slot is further compressed via a DFT-based transform, where the codebook takes the form

[0098]

[0099] where follows the same structure as the Rel-16 Type-II codebook, is an N4xQ matrix (Q < N4) with columns selected from a critically sampled size N4DFT matrix, as follows

[0100]

[0101]

[0102] only the indices of the Q selected columns of are reported. Note that may be layer-specific, e.g., , or layer-common, i.e., where RI corresponds to the total number of layers, and the operator corresponds to the Kronecker matrix product. Here, is a 2LxMQ size matrix with layer-specific entries representing the LCCs corresponding to the spatial, frequency, and time domain DFT basis vectors. Thus, a size 2LxMQ bitmap associated with the Rel-18 Type-II codebook can need to be reported.

[0103] For codebook reporting, the codebook reporting is partitioned into two parts based on the priority of the reported information. Each part is encoded separately (part 1 has potentially higher code rate). Only the parameters for the NR Rel-16 Type-II codebook are listed below.

[0104] The content of the CSI report contains part 1 and part 2. Part 1 contains the RI plus the channel quality indicator (CQI) plus the total number of coefficients. Part 2 contains the SD basis indicator plus the FD basis indicator / layer plus the bitmap / layer plus the coefficient amplitude information / layer plus the coefficient phase information / layer plus the strongest coefficient indicator / layer. Furthermore, part 2 CSI can be decomposed into subparts each with different priority (higher priority information is listed first). This partitioning is used to allow dynamic reporting size of the codebook based on the available resources in the uplink phase.

[0105] ​​​Furthermore, Type-II codebook is based on aperiodic CSI reporting and is reported in the physical uplink shared channel (PUSCH) triggered only via downlink control information (DCI) (with one exception). Type-I codebook can be based on periodic CSI reporting physical uplink control channel (PUCCH) or semi-persistent CSI reporting (PUSCH or PUCCH) or aperiodic reporting (PUSCH).

[0106] For priority reporting of part 2 CSI, it should be noted that multiple CSI reports can be transmitted with different priorities as shown in Table 1. Table 1 shows the reporting hierarchy of part 2 CSI.

[0107] Table 1

[0108]

[0109] Furthermore, The priority of a CSI report is based on the following: 1) one CSI report corresponding to one CSI report configuration of one cell can have higher priority than another CSI report corresponding to one other CSI report configuration of the same cell, 2) a CSI report intended to one cell can have higher priority than other CSI report intended to another cell, 3) a CSI report can have higher priority based on the CSI report content, for example, a CSI report carrying L1-RSRP information has higher priority, 4) a CSI report can have higher priority based on its type, for example, whether the CSI report is aperiodic, semi-persistent or periodic, and whether the report is sent via PUSCH or PUCCH can affect the priority of the CSI report.

[0110] In view of this, CSI reports can be prioritized as follows, where a CSI report with lower identifier (ID) has higher priority

[0111]

[0112] where s refers to the CSI report configuration index, M s refers to the maximum number of CSI report configurations, c refers to the cell index, N cells refers to the number of serving cells, k is 0 for CSI report carrying L1-RSRP or L1-SINR, otherwise 1, y is 0 for aperiodic reporting, 1 for semi-persistent reporting on PUSCH, 2 for semi-persistent reporting on PUCCH, 3 for periodic reporting.

[0113] Regarding triggering of aperiodic CSI reporting on PUSCH, the UE reports the required CSI information of the network using the CSI framework in NR Rel-15. The triggering mechanism between the reporting setting and the resource setting can be summarized in Table 2 below, which refers to the Medium Access Control Element (MAC CE), Semi-Persistent (SP), and Aperiodic (AP).

[0114] Table 2

[0115]

[0116] Furthermore, all associated resource settings of a CSI reporting setting need to have the same time-domain behavior. Once configured by RRC, periodic CSI-RS / IM resources and CSI reporting are always assumed to be present and active. Aperiodic and semi-persistent CSI-RS / IM resources and CSI reporting need to be explicitly triggered or activated. Aperiodic CSI-RS / IM resources and aperiodic CSI reporting are jointly triggered by transmitting DCI format 0-1. Semi-persistent CSI-RS / IM resources and semi-persistent CSI reporting are activated independently.

[0117] For aperiodic CSI-RS / IM resources and aperiodic CSI reporting, the triggering is jointly done by transmitting DCI format 0-1. DCI format 0_1 contains a CSI request field (0 to 6 bits). A non-zero request field points to a so-called aperiodic trigger state (see Figure 7 ) configured by RRC. The aperiodic trigger state is in turn defined as a list of up to 16 aperiodic CSI reporting settings for which the UE computes the CSI simultaneously and transmits its CSI reporting setting ID identification on the scheduled PUSCH transmission.

[0118] Figure 7 An example 700 of an aperiodic trigger state defining a list of CSI reporting settings is illustrated.

[0119] When a CSI reporting setting is linked with an aperiodic resource setting (which can include multiple resource sets), the aperiodic NZP CSI-RS resource set for channel measurement, the aperiodic CSI-IM resource set (if used), and the aperiodic NZP CSI-RS resource set for IM (if used) to be used for a given CSI reporting setting are also included in the aperiodic trigger state definition.

[0120] Figure 8 An example 800 of an aperiodic trigger state is illustrated. Example 800 illustrates that the aperiodic trigger state indicates resource sets and quasi co-location (QCL) information.

[0121] For aperiodic NZP CSI-RS, the QCL source to be used is also configured as aperiodic trigger state. The UE assumes that the resources used for channel and interference computation are processed with the same spatial filter, i.e. on “QCL-TypeD” quasi co-location

[0122] Figure 9 An example 900 of RRC configuration of NZP CSI-RS resources is illustrated.

[0123] Figure 10 An example 1000 of RRC configuration of CSI-IM resources is illustrated.

[0124] Table 3 summarizes the type of uplink channel used for CSI reporting depending on the CSI codebook type, and refers to subband (SB) and wideband (WB).

[0125] Table 3

[0126]

[0127] For aperiodic CSI reporting, PUSCH-based reporting is split into two CSI parts: CSI part 1 and CSI part 2. The reason is that the size of the CSI payload varies significantly, and therefore worst-case uplink control information (UCI) payload size design would result in large overhead.

[0128] CSI part 1 has a fixed payload size (and can be decoded by the NE 102 without a priori information), and contains the following: RI (if reported) of the first codeword, CSI-RS resource index (CRI) (if reported), and CQI; number of non-zero wideband amplitude coefficients per layer of Type-II CSI feedback on PUSCH. CSI part 2 has a variable payload size that can be derived from the CSI parameters in CSI part 1, and contains the PMI and CQI of the second codeword when RI > 4.

[0129] Figure 11 An example 1100 of CSI reporting is illustrated. The example 1100 illustrates the ordering of aperiodic CSI reporting of CSI part 2 in case of 3 reporting settings x, y, and z defined by aperiodic trigger state indicated by DCI format 0_1. The example 1100 is based on partial CSI omission of Rel. 15 PUSCH-based CSI.

[0130] As previously mentioned, CSI reporting is prioritized according to the following: 1) time domain behavior and physical channel, where reports with higher degree of dynamics are given priority over reports with lower degree of dynamics, and PUSCH has priority over PUCCH; 2) CSI content, where beam reports (i.e., L1-RSRP reports) have priority over regular CSI reports; 3) serving cell the CSI corresponds to (in case of carrier aggregation (CA) operation), CSI corresponding to PCell has priority over CSI corresponding to Scell; 4) reportConfigID.

[0131] With respect to CQI reporting, a CSI report can contain a number of CQI reports corresponding to channel quality assuming a maximum target block error rate, which indicates a modulation order, a code rate, and a corresponding spectral efficiency associated with the modulation order and code rate pair. Examples of maximum block error rates are 0.1 and 0.00001. Modulation orders can vary from QPSK up to 1024 QAM, while code rates can vary from 30 / 1024 up to 948 / 1024. Table 4 describes one example of a CQI table for identifying a 4-bit CQI indicator with possible CQI values corresponding to a modulation order, a code rate, and an efficiency.

[0132] Table 4

[0133]

[0134] CQI values can be reported in two formats: a wideband format, where one CQI value corresponds to each physical downlink shared channel (PDSCH) transport block report; and a subband format, where one wideband CQI value is reported for the entire transport block in addition to a set of subband CQI values corresponding to the CQI subbands on which the transport block is transmitted. The CQI subband size is configurable and depends on the number of PRBs in the bandwidth part. Table 5 shows an example of configurable subband sizes for a given bandwidth part (BWP) size.

[0135] Table 5

[0136]

[0137] If the higher layer parameter cqi-BitsPerSubband in CSI reporting settings CSI-ReportConfig is configured, subband CQI values are reported in full form, i.e., 4 bits are used for each subband CQI based on the CQI table (e.g., Table 4). If the higher layer parameter cqi-BitsPerSubband in CSI-ReportConfig is not configured, a 2-bit subband differential CQI value is reported for each subband s, which is defined as:

[0138] Subband offset level (s) = Subband CQI index (s) - Wideband CQI index.

[0139] Table 6 shows the mapping from 2-bit subband differential CQI values to offset levels.

[0140] Table 6

[0141]

[0142] Regarding the AI / ML model for CSI measurement and reporting, for an AI / ML based CSI framework, there are multiple alternatives for the overview of AI / ML algorithm functionality as follows.

[0143] In one or more embodiments, the AI / ML model is trained at the UE. This alternative can seem reasonable as the UE is the node that can seamlessly collect training data for CSI acquisition using DL pilot signals (e.g., CSI-RS for channel measurement), however, the AI / ML model is typically retrained every time the environment changes (e.g., change in UE location or orientation), and each training example involves significant memory and computational complexity requirements.

[0144] In one or more embodiments, the AI / ML model is trained at the network node (e.g., NE 102). One advantage of this approach is that the network has significantly more power and computational capability compared to the UE, and thus can manage training moderately complex AI / ML models, as well as store large amounts of training data. Furthermore, since the network nodes are mostly assumed to be fixed, their coverage area is expected to be the same, and thus a single AI / ML model can be applicable to UEs within a specific area of a cell for a reasonable period of time. One challenge of this approach is related to obtaining training data at the network node, especially for frequency division duplex (FDD) systems where UL / DL channel reciprocity can not hold. It should be noted that the overhead corresponding to the feedback of training data from the UE to the network should be considered as one of the metrics when evaluating the efficiency of the AI / ML algorithm.

[0145] Later, due to the advantages in terms of memory, computation, and cell center characteristics corresponding to network based AI / ML model computation, we assume that the AI / ML model is trained at the network. The challenges corresponding to obtaining training data for the DL channel at the network side are discussed below.

[0146] Regarding obtaining training data, assuming that the AI / ML model is trained at the network, several aspects of DL training data acquisition at the network side to enable efficient AI / ML modeling are as follows.

[0147] To maintain the robustness of the AI / ML model with respect to channel variations, DL training data is usually continuously fed back to the network to keep up with changes in the environment (e.g., traffic, weather, and moving scatterers). It should be noted that this can not necessarily correspond to online learning; even for offline learning algorithms, the framework for obtaining new training data corresponding to channel variations is usually characterized.

[0148] Based on the current codebook-based DL CSI feedback scheme in NR, the CSI is compressed in at least one of the spatial or frequency domain. One intuitive approach would be to use codebook-based CSI feedback, e.g., Type-I and / or Type-II codebooks for obtaining the training data. One drawback of this approach is that the training data would include CSI feedback that has already been compressed via conventional methods, which would have an adverse effect on the AI / ML model inference accuracy. For example, if the AI / ML model compares the output of the AI / ML model with the channel corresponding to the CSI feedback to assess its own inference accuracy, this assessment would not be accurate because it is based on i.e., an estimate of the channel based on the pre-defined compression, rather than i.e., a digitally quantized channel without further compression in the spatial or frequency domain. On the other hand, if the UE feeds back the training data corresponding to the DL CSI feedback without compression in the spatial and / or frequency dimensions, the feedback overhead of the training data would be large, which would defeat the purpose of using an AI / ML model, which is primarily to reduce the overall CSI feedback overhead. Numerically, AI / ML-based CSI feedback aims to minimize the following metric

[0149]

[0150] where denotes the digital domain representation of the channel matrix. On the other hand, the compressed channel after the codebook-based transformation represents the recovered channel

[0151]

[0152] Since , the outputs of the two optimizations would result in different channel estimates.

[0153] Regarding the output of the AI / ML model, for DL CSI acquisition in NR, whether the network is operating in FDD mode or in time division duplex (TDD) mode, AI / ML is unlikely to completely replace the RS-based CSI feedback for high-resolution precoding design, as some channel parameters can vary from one time instance to another without strong correlation across the two time instances, e.g., the initial random phase of the channel. In view of this, an AI / ML-based CSI framework can be envisioned as a means to further reduce the CSI feedback overhead compared to the conventional approach, e.g., reducing the number of dominant spatial domain basis indices, frequency / delay domain basis indices, and time / Doppler domain basis indices after spatial domain, frequency domain, and time domain transformations, respectively. While the current CSI feedback framework already provides CSI feedback overhead reduction via utilizing such transformations, if a wider range of transformation techniques are preconfigured, the CSI dimensionality can be further reduced, where different transformations can be selected for a given UE based on the variation of the channel.

[0154] The following contains additional information regarding antenna panels / ports, quasi co-location, transmission configuration indication (TCI) states, and spatial relations.

[0155] In one or more implementations, the terms antenna, panel, and antenna panel are used interchangeably. An antenna panel can be hardware used to transmit and / or receive radio signals at frequencies below 6 GHz, e.g., frequency range 1 (FR1), or above 6 GHz, e.g., frequency range 2 (FR2), or millimeter wave (mmWave). In some implementations, an antenna panel can include an array of antenna elements, where each antenna element is connected to hardware, e.g., a phase shifter, that allows a control module to apply spatial parameters to the transmission and / or reception of a signal. The resulting radiation pattern can be referred to as a beam, which can or can not be unimodal, and can allow the device to amplify a signal transmitted or received from a spatial direction.

[0156] Additionally or alternatively, antenna panels can or can not be virtualized into antenna ports in the specification. An antenna panel can be connected to a baseband processing module through a radio frequency (RF) chain for each of the transmit (outbound) and receive (inbound) directions. The capabilities of the devices in terms of the number of antenna panels, their duplexing capabilities, their beamforming capabilities, etc. can or can not be transparent to other devices. In some implementations, the capability information can be conveyed via signaling, or in some implementations, the capability information can be provided to the devices without the need for signaling. In cases where such information can be used by other devices, it can be used for signaling or local decision making.

[0157] Additionally or alternatively, a device (e.g., UE, node) antenna panel can be a physical or logical antenna array that includes a set of antenna elements or antenna ports that share a common or important portion of an RF chain (e.g., in-phase / quadrature (I / Q) modulator, analog / digital (A / D) converter, local oscillator, phase shift network). A device antenna panel or “device panel” can be a logical entity with physical device antennas mapped to the logical entity. The mapping of physical device antennas to logical entities can depend on the device implementation. Communicating (receiving or transmitting) on at least a subset of the antenna elements or antenna ports of an antenna panel that are active for radiating energy (also referred to herein as active elements) requires biasing or energizing of the RF chain, which results in current consumption or power consumption in the device associated with the antenna panel (including power amplifier / low noise amplifier (LNA) power consumption associated with the antenna elements or antenna ports). The phrase “active for radiating energy” as used herein does not mean limited to a transmit function and also encompasses a receive function. Thus, an antenna element active for radiating energy can be coupled to a transmitter to transmit radio frequency energy or coupled to a receiver to receive radio frequency energy simultaneously or sequentially, or generally can be coupled to a transceiver to perform its intended functionality. Communicating on the active elements of an antenna panel enables generation of a radiation pattern or beam.

[0158] Additionally or alternatively, depending on the device’s own implementation, a “device panel” can have at least one of the following functionalities as the following operational roles: a unit to independently control its set of Tx beams, a unit to independently control its set of transmit power, a unit to independently control its set of transmit timing. The “device panel” can be transparent to the NE 102. For certain conditions, the NE 102 or network can assume that the mapping between the device’s physical antennas to logical entities “device panels” can be invariant. The conditions can include, for example, until the next update or report from the device, or including a duration for which the NE 102 assumes the mapping does not change. The device can report its capabilities regarding “device panels” to the NE 102 or network. The device capabilities can include at least the number of “device panels.” In one implementation, the device can support UL transmission from one beam within a panel; for multiple panels, more than one beam (one per panel) can be used for UL transmission. In another implementation, more than one beam per panel can be supported / used for UL transmission.

[0159] Additionally or alternatively, an antenna port is defined such that a channel on which a symbol related to the antenna port is transmitted can be inferred from a channel on which another symbol related to the same antenna port is transmitted.

[0160] Two antenna ports are said to be quasi co-located (QCL) if the large-scale properties of the channel over which a symbol related to one antenna port is transmitted can be inferred from the channel over which a symbol related to another antenna port is transmitted. The large-scale properties include one or more of delay spread, Doppler spread, Doppler shift, average gain, average delay, or spatial Rx parameters. Two antenna ports can be quasi-located with respect to a subset of the large-scale properties, and different subsets of the large-scale properties can be indicated by a QCL type. A QCL type can indicate which channel properties are the same between two reference signals (e.g., on two antenna ports). Thus, reference signals can be linked to each other with respect to what assumptions a UE can make about their channel statistics or QCL properties. For example, a qcl-Type can take one of the following values:

[0161] - 'QCL-TypeA': {Doppler shift, Doppler spread, average delay, delay spread}

[0162] - 'QCL-TypeB': {Doppler shift, Doppler spread}

[0163] - 'QCL-TypeC': {Doppler shift, average delay}

[0164] - 'QCL-TypeD': {spatial Rx parameters}.

[0165] The spatial Rx parameters can include one or more of angle of arrival (AoA), dominant AoA, average AoA, angular spread, power angular spectrum (PAS) of AoA, average AoD (angle of departure), PAS of AoD, transmit / receive channel correlation, transmit / receive beamforming, spatial channel correlation, etc.

[0166] QCL-TypeA, QCL-TypeB, and QCL-TypeC can be applicable for all carrier frequencies, but QCL-TypeD can be applicable only for higher carrier frequencies (e.g., mmWave, FR2 and above), where essentially a UE can not be able to perform omni-directional transmission, i.e., the UE will need to form a beam for directional transmission. For QCL-TypeD between two reference signals A and B, reference signal A is considered to be spatially co-located with reference signal B, and the UE can assume that reference signals A and B can be received with the same spatial filter (e.g., with the same receiver (RX) beamforming weights).

[0167] An“antenna port” according to embodiments can be a logical port that can correspond to a beam (resulting from beamforming), or can correspond to a physical antenna on the device. In some embodiments, a physical antenna can be mapped directly to a single antenna port, where the antenna port corresponds to the actual physical antenna. Alternatively, a set or subset of physical antennas, or an antenna set or antenna array or antenna subarray, can be mapped to one or more antenna ports after applying complex weights, cyclic delays, or both to the signal on each physical antenna. The set of physical antennas can have antennas from a single module or panel or from multiple modules or panels. The weights can be fixed, as in an antenna virtualization scheme, e.g., cyclic delay diversity (CDD). The procedure for deriving an antenna port from a physical antenna can be specific to the device implementation and transparent to other devices.

[0168] In one or more embodiments, a TCI state (transmission configuration indication) associated with a target transmission can indicate parameters for configuring a quasi co- location relationship between the target transmission (e.g., target RS for DM-RS ports of the target transmission during a transmission occasion) and one or more source reference signals (e.g., synchronization signal block (SSB) / CSI-RS / sounding reference signal (SRS)) with respect to one or more quasi co-location type parameters indicated in the corresponding TCI state. The TCI describes which reference signals are used as QCL sources and what QCL properties can be derived from each reference signal. A device can receive a configuration of multiple transmission configuration indicator states of a serving cell for transmissions on the serving cell. In some described embodiments, a TCI state includes at least one source RS to provide a reference (UE assumption) for determining QCL and / or spatial filters.

[0169] Additionally or alternatively, spatial relation information associated with a target transmission can indicate parameters for configuring a spatial setting between the target transmission and a reference RS (e.g., SSB / CSI-RS / SRS). For example, a device can transmit a target transmission with the same spatial domain filter used to receive a reference RS (e.g., a DL RS such as a SSB / CSI-RS). In another example, a device can transmit a target transmission with the same spatial domain transmit filter used to transmit a reference RS (e.g., a UL RS such as a SRS). A device can receive a configuration of multiple spatial relation information configurations of a serving cell for transmissions on the serving cell.

[0170] Additionally or alternatively, if a device is configured with separate DL / UL TCI by RRC signaling, an UL TCI state is provided. An UL TCI state can include a source reference signal that provides a reference for determining an UL spatial domain transmit filter for UL transmissions (e.g., dynamic grant / configured grant based PUSCH, dedicated PUCCH resources) in a component carrier (CC) or across a configured set of CCs / BWPs.

[0171] Additionally or alternatively, if the apparatus is configured with joint DL / UL TCI by RRC signaling (e.g., configuration of joint TCI or separate DL / UL TCI is based on RRC signaling), a joint DL / UL TCI state is provided. The joint DL / UL TCI state at least indicates a common source reference RS used for determining both DL QCL information and UL spatial transmit filter. The source RS determined from the indicated joint (or common) TCI state provides QCL Type-D indication (e.g., for apparatus-specific physical downlink control channel (PDCCH) / PDSCH) and is used for determining UL spatial transmit filter (e.g., for UE-specific PUSCH / PUCCH) for a CC or across a configured set of CCs / BWPs. In one example, the UL spatial transmit filter is derived from the RS of DL QCL Type-D in the joint TCI state. The spatial setting of UL transmission can be according to the spatial relation on the source RS in the joint TCI state configured with qcl-Type set to “typeD”.

[0172] Assume that the UE 104 has a channel with P channel paths (index ) between the UE 104 and the NE 102 (e.g., gNB), which occupies number of frequency bands (index ), where the gNB is equipped with K number of antennas (index ). Then, the channel at time index can be represented as

[0173]

[0174] where is the complex gain of path p at antenna k, is the PMI subband spacing, is the delay of path p, is the carrier frequency, c is the speed of light, d is the antenna spacing at the NE 102 (e.g., gNB), is the angular spatial displacement at the NE 102 (e.g., gNB) antenna array corresponding to path p, is the time index, v is the relative speed between the NE 102 (e.g., gNB) and the UE, and is the angle between the moving direction of path p and the signal incident direction.

[0175] The above channel is parameterized by three dimensions: spatial, frequency, and time dimensions. To construct a precoder codebook with reasonable CSI feedback overhead, the CSI corresponding to the three dimensions is compressed. In Rel-16 e-Type-II codebook, both spatial and frequency domains are compressed via DFT transformation of the columns of two-dimensional and one-dimensional DFT matrices, respectively, while in a potential Rel-18 e-Type-II codebook for high speed, the time domain is further compressed via DFT transformation in the form of a column of one-dimensional DFT matrix. Additionally or alternatively, the CSI feedback can be transmitted in an explicit format, e.g., in terms of explicit channel coefficients, to enhance the CSI feedback resolution. However, the CSI feedback overhead will be significantly increased, especially for the scenario of training dataset transmission, where the CSI feedback includes a large number of training dataset points corresponding to different realizations of the CSI. In the present disclosure, we propose an AI-based CSI framework, where a statistical CSI training data is reported via aggregating similar training dataset points corresponding to the CSI, where the corresponding weight or occurrence rate of this dataset point is fed back as part of the CSI feedback corresponding to the training data. Additionally, a likelihood ratio corresponding to whether a given coefficient of the channel or precoding matrix is associated with a non-zero amplitude value is reported, where the likelihood ratio is based on the weight or occurrence rate of the given CSI data point as part of the training dataset. Furthermore, the aforementioned AI-based CSI framework facilitates inferring the characteristics of the channel distribution based on the training dataset, such that the CSI feedback can exploit a distribution-aware data compression scheme, e.g., Huffman coding, where the CSI parameters are encoded such that the values with higher occurrence probability are mapped to shorter bit sequences, while the CSI parameters are encoded such that the values with lower occurrence probability are mapped to longer bit sequences.

[0176] In the present disclosure, the “training dataset” can also be simply referred to as “dataset”.

[0177] Regarding the indication of the CSI training dataset transmission, in one or more embodiments, the training dataset is transmitted from the network node to the UE. In one example, the training dataset is transmitted on PDSCH. In another example, the training dataset is transmitted on PDCCH. In another example, the training dataset is transmitted via higher layer signaling, e.g., as part of RRC configuration.

[0178] In one or more embodiments, the training dataset is transmitted from the UE to the network node (e.g., NE 102). In one example, the training dataset is transmitted on PUSCH. In another example, the training dataset is transmitted on PUCCH. In another example, the training dataset is further divided into two parts, a first part of the two parts of the training dataset is transmitted on PUCCH, and a second part of the two parts of the training data is transmitted on PUSCH.

[0179] In one or more embodiments, the training data set corresponds to a reporting type configured via a reporting setting. In one example, the reporting setting includes a higher layer configuration parameter, where the higher layer configuration parameter is set to true if the data set report corresponds to the training data set report. For example, the higher layer configuration parameter is part of an RRC configuration.

[0180] In one or more embodiments, the training data set is configured via a dedicated higher layer configuration, e.g., a training data setting or an AI / ML setting.

[0181] With respect to grouping of data set points, in one or more embodiments, a data set containing data set points is partitioned into one or more data set groups. In one example, all data set points of a data set are mapped to the same group, i.e., a data set group is equivalent to a data set. In another example, each data set point is mapped to a different group, i.e., each data set point is uniquely mapped to a data set group. In another example, data set points are partitioned into two or more data set groups, where each data set group includes multiple data set points.

[0182] In one or more embodiments, a limit of a number of data set groups in a data set is a maximum number of data set groups. In one example, a data set containing a number of data set groups equal to the maximum number of data set groups cannot contain additional data set groups without removing a data set group from the data set groups in the data set. In another example, each data set group is associated with a set of labels, features, identification (ID) values, or a combination thereof.

[0183] In one or more embodiments, a limit of a number of data set points in a data set group is a maximum number of data set points. In one example, all data set groups include the same number of data set points. In another example, different data set groups of a data set are configured with different values of the maximum number of data set points. In another example, each data set point is mapped to a different group, i.e., each data set point is uniquely mapped to a data set group.

[0184] In one or more embodiments, an indication corresponding to a total number of bits of a size of a training data set is reported in a first portion of a training data set report, where the training data set report includes multiple portions, segments, partitions, or a combination thereof.

[0185] In one or more embodiments, an identifier or indicator of the data set group is signaled from the first communication node to the second communication node. In one example, the first communication node is the UE 104 and the second communication node is the network node (e.g., NE 102). In another example, the first communication node is the network node (e.g., NE 102) and the second communication node is the UE 104. In another example, the identifier or indicator corresponds to a data set group from a set of configured data set groups, where each data set group in the set of configured data set groups is associated with an ID.

[0186] With respect to the time marking of the data set points, in one or more embodiments, each data set group is marked with a timestamp. In one example, the timestamp identifies the time at which the data set points of the data set group are collected. In another example, the timestamp identifies the time at which the data set points of the data set group are added to the data set. In another example, the timestamp identifies the time at which the data set points of the data set group are set to be omitted from the data set.

[0187] In one or more embodiments, each data set group is marked with at least one of a start time, a time interval, and a time periodicity. In one example, the start time corresponds to the time at which the data set group is activated. In another example, the time interval corresponds to the interval of time for which the data set group is active from the start time. In another example, the time periodicity corresponds to the periodic interval at which the start time and the time interval repeat. In another example, for a start time of 4 seconds, a time interval of 2 seconds, and a time periodicity of 10 seconds, the data set group is activated in the following time periods [4, 6], [14, 16], [24, 26],...

[0188] In one or more embodiments, each data set group is marked with a parameter corresponding to at least one of a time or a delay. In one example, the marking of the data set is based on the output of an AI / ML algorithm. In another example, the marking of the data set is based on the output of a classifier function.

[0189] In one or more embodiments, a value corresponding to a label of the data set group is signaled from the first communication node to the second communication node. In one example, the first communication node is the UE 104 and the second communication node is the network node (e.g., NE 102). In another example, the first communication node is the network node (e.g., NE 102) and the second communication node is the UE 104. In another example, the value of the label is extracted from a codebook of values known to both communication nodes.

[0190] With respect to the weight association of the set of data sets, in one or more embodiments, each data set group is associated with at least one of a weight, a probability of occurrence, a scaling factor, or a priority index. In one example, a data set point of a first data set group associated with at least one of a larger value of a weight, a probability of occurrence, a scaling factor, or a priority index corresponds to a greater significance of the data set point of the first data set group than a data set point of a second data set group associated with at least one of a smaller value of a weight, a probability of occurrence, a scaling factor, or a priority index. In another example, a sum of at least one of a weight, a probability of occurrence, a scaling factor, or a priority index across all data set groups equals a unit value. In another example, at least one of a weight, a scaling factor, or a priority index is normalized by a size of a data set. In another example, at least one of a weight, a scaling factor, or a priority index is normalized by a value of a maximum weight, e.g., at least one weight is set to 1. In another example, a value corresponding to at least one of a weight, a probability of occurrence, a scaling factor, or a priority index is selected from a predefined or preconfigured codebook of values.

[0191] In one or more embodiments, at least one of a weight, a probability of occurrence, a scaling factor, a priority index, or a combination thereof associated with each of the set of data sets is updated based on an event. In one example, the event is one of a periodic or semi-persistent event based on a configuration for updating the data set, where the data set is partially updated. In another example, the event is triggered by at least one of a network configuration signal, a downlink control information, or a MAC-CE signal. In another example, the event is triggered by at least one of a UE-based signal on an uplink control information or multiplexed with a CSI report.

[0192] In one or more embodiments, a set of data set groups constituting a data set is updated after each event, where a data set group associated with at least one of a smaller value of a weight, a probability of occurrence, a scaling factor, or a priority index is replaced with a data set group associated with at least one of a larger value of a weight, a probability of occurrence, a scaling factor, or a priority index.

[0193] In one or more embodiments, data set points corresponding to different data set groups are ordered based on a characteristic of the corresponding data set group. In one example, the characteristic is based on at least one of a weight, a probability of occurrence, a scaling factor, or a priority index of the data set group. In another example, the characteristic is based on a time label or a timestamp associated with the data set group. In another example, the ordering of the data set groups is based on an ascending order of values associated with the characteristic. In another example, the ordering of the data set groups is based on a descending order of values associated with the characteristic.

[0194] In one or more embodiments, processing of the set of data sets is based on one of a set of decisions including updating a weight or value of a label associated with the set of data sets, not updating a weight or value of a label associated with the set of data sets, omitting data set points associated with the set of data sets associated with small weight values or low probability of occurrence, or adding a new set of data sets associated with a plurality of data set points, where the set of data sets is associated with a positive value of weight or probability of occurrence.

[0195] With respect to classification of the set of data sets, in one or more embodiments, a data set point is associated with one of a plurality of sets of data sets, where grouping is based on one or more characteristics of the data set point. In one example, a data set point is classified into no more than one set of data sets. In another example, classification of the data set point is based on maintenance, closed loop operation, e.g., AI / ML based model maintenance. In another example, classification of the data set point is based on ground truth information associated as a label of the data set point.

[0196] In one or more embodiments, data set points of a data set are classified based on observable characteristics, e.g., characteristics that can be directly computed from values of the data points. In one example, observable characteristics are derived from deterministic formulae of values of the data set points. In another example, observable characteristics are transformed variants of values of the data set points based on a transformation operation, e.g., projection to a k-dimensional plane. In another example, observable characteristics are based on normalization of values of the data set points relative to one or more values of other data set points.

[0197] In one or more embodiments, data set points of a data set are classified based on unobservable characteristics, e.g., inferred from analysis of characteristics, e.g., statistical characteristics, of a large set of data points. In one example, unobservable characteristics correspond to whether a data set point is classified as an outlier or an inlier. In another example, unobservable characteristics correspond to a statistical function, e.g., a second order coherence or correlation function. In another example, unobservable characteristics are based on a power delay profile corresponding to an approximate distribution associated with the data set.

[0198] With respect to data set grouping, in one or more implementations, a data set corresponds to at least one of CSI data, or spatial, time, frequency domain precoded data, spatial beam based data. In the following examples, "CSI" is used to refer to precoding and beam processing. In one example, a data set group of data sets is associated with a timestamp corresponding to at least one of a collection time of CSI, a time at which data corresponding to CSI is signaled, or a time interval in which a data set point is valid. In another example, a data set group of data sets is associated with a value corresponding to at least one of a weight or a probability of occurrence, where a larger value corresponds to a stronger correlation of a data set point of the data set group to actual or reported CSI, and a smaller value corresponds to a weaker correlation of a data set point of the data set group to actual or reported CSI. In another example, data set groups are classified based on observable characteristics of CSI, e.g., at least one of a number of dominant basis indices based on transformed frequency domain basis, a ratio of a maximum value of singular values of a channel tap, channel matrix, or precoding matrix, or a power delay profile associated with CSI. In another example, data set groups are classified based on unobservable characteristics of CSI, where unobservable characteristics are based on observable characteristics, e.g., with respect to a number of dominant basis indices based on transformed frequency domain basis, an indicator of whether a channel corresponds to a line of sight (LoS) or non-line of sight (NLoS) channel.

[0199] In one or more implementations, a data set corresponds to positioning (or localization) data. In one example, a data set group of data sets is associated with a timestamp corresponding to at least one of a collection time of positioning information, a time at which data corresponding to a position is signaled, or a time interval in which a data set point is valid. In another example, a data set group of data sets is associated with a value corresponding to at least one of a weight or a probability of occurrence, where a larger value corresponds to a stronger correlation of a data set point of the data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the data set group to an actual position. In another example, data set groups are classified based on observable characteristics of a position, e.g., at least one of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival. In another example, data set groups are classified based on unobservable characteristics of a position, where unobservable characteristics are based on observable characteristics, e.g., with respect to one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, an indicator of whether a channel corresponds to an indoor or outdoor UE.

[0200] In one or more embodiments, the data set corresponds to UE mobility. In one example, a data set group of the data set is associated with a timestamp corresponding to at least one of a collection time of the mobility or cell association information, a time of signaling data corresponding to cell association, or a time interval in which a data set point of the data set group is valid. In another example, a data set group of the data set is associated with at least one of a value corresponding to a weight or a probability of occurrence, where a larger value corresponds to a stronger correlation of a data set point of the data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the data set group to a heuristic cell association or selection. In another example, a data set group is classified based on an observable characteristic of UE mobility, such as at least one of RSRP, SINR, beam-based information, or CSI. In another example, a data set group is classified based on a non-observable characteristic of mobility, where the non-observable characteristic is based on an observable characteristic, such as at least one of a value regarding RSRP, SINR, beam-based information, or CSI, a flag of whether the UE is associated with an optimal cell.

[0201] Figure 12 An example of a UE 1200 in accordance with aspects of the disclosure is described. The UE 1200 can include a processor 1202, a memory 1204, a controller 1206, and a transceiver 1208. The processor 1202, the memory 1204, the controller 1206, or the transceiver 1208, or various combinations thereof or various components thereof, can be examples of means for performing various aspects of the disclosure as described herein. These components can be coupled (e.g., operatively, communicatively, functionally, electronically, electrically) via one or more interfaces.

[0202] The processor 1202, the memory 1204, the controller 1206, or the transceiver 1208, or various combinations thereof or components thereof, can be implemented in hardware (e.g., circuitry). The hardware can be a processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), or other programmable logic device, or any combination thereof, configured as or otherwise supporting means for performing the functions described in the present disclosure.

[0203] The processor 1202 can include an intelligent hardware device, e.g., a general- purpose processor, a DSP, a CPU, an ASIC, an FPGA, or any combination thereof. In some embodiments, the processor 1202 can be configured to operate the memory 1204. In some other embodiments, the memory 1204 can be integrated into the processor 1202. The processor 1202 can be configured to execute computer-readable instructions stored in the memory 1204 to cause the UE 1200 to perform various functions of the present disclosure.

[0204] Memory 1204 can include volatile or nonvolatile memory. Memory 1204 can store computer-readable, computer-executable code including instructions that, when executed by processor 1202, cause UE 1200 to perform various functions described herein. The code can be stored in a non-transitory computer-readable medium such as memory 1204 or another type of memory. Computer-readable media include both volatile and nonvolatile media, removable and nonremovable media, and communication media. Non-transitory storage media can be any available media that can be accessed by a general purpose or special purpose computer.

[0205] In some implementations, processor 1202 and memory 1204 coupled with processor 1202 can be configured to cause UE 1200 to perform one or more of the functions described herein (e.g., by processor 1202 executing instructions stored in memory 1204). For example, processor 1202 can support wireless communication at UE 1200 in accordance with examples disclosed herein. UE 1200 can be configured to support a means for transmitting, to network equipment, first signaling on a physical channel indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; updating the second label after the transmission of the first signaling; updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset; and transmitting, to the network equipment, second signaling on the physical channel indicating a second training dataset report including updated information corresponding to the updated training dataset.

[0206] Additionally or alternatively, the UE 1200 can be configured to support cases where the physical channel is an uplink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where at least one of the first labels is a duration comprising a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event, and the event is one of periodic or semi-persistent based on a configuration for updating the data set, triggered by at least one of a network configuration signal, a downlink control information, or a MAC-CE signal, or a combination thereof; where the training data set is updated by replacing a data set group associated with a smaller value of the weight with a data set group associated with a larger value of the weight; where a data set group associated with a time stamp corresponding to an earlier value is replaced with a data set group associated with a time stamp corresponding to a more recent value; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point with respect to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix to a minimum value of the singular value, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with a channel of CSI as to whether it corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to an actual position; or classifying based on observable characteristics of a position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, the channel associated with the positioning information is one or more of a flag corresponding to whether it corresponds to an indoor or outdoor UE; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0207] The UE 1200 can be configured to support a means for receiving, from a network equipment, a first signaling on a physical channel indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; and receiving, from the network equipment, a second signaling on the physical channel indicating a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second labels of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset.

[0208] Additionally or alternatively, the UE 1200 can be configured to support cases where the physical channel is a downlink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where at least one of the first labels is a duration including a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event; where the event is based on a configuration for updating the data set being one of a periodic or semi-persistent event; where the event is triggered by at least one of on uplink control information or a UE-based signal multiplexed with the CSI report; where the training data set is updated by replacing data set groups associated with smaller values of the weight with data set groups associated with larger values of the weight; where data set groups associated with time stamps corresponding to more recent values are replaced with data set groups associated with time stamps corresponding to earlier values; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point relative to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or an occurrence probability, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classifying the CSI based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix to a minimum value of the singular value, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with the channel of CSI of whether the channel corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to an actual position; or classifying based on observable characteristics of a position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, the channel associated with the positioning information is one or more of a flag corresponding to whether the UE is indoors or outdoors; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag of whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0209] The UE 1200 can be configured to support at least one memory; and at least one processor coupled with the at least one memory and configured to cause the UE to transmit, to a network equipment over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; update the second label after the transmission of the first signaling; update the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset; and transmit, to the network equipment over the physical channel, second signaling indicative of a second training dataset report including updated information corresponding to the updated training dataset.

[0210] Additionally or alternatively, the UE 1200 can be configured to support cases where the physical channel is an uplink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where at least one of the first labels is a duration comprising a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event, and the event is one of periodic or semi-persistent based on a configuration for updating the data set, triggered by at least one of a network configuration signal, a downlink control information, or a MAC-CE signal, or a combination thereof; where the training data set is updated by replacing a data set group associated with a smaller value of the weight with a data set group associated with a larger value of the weight; where a data set group associated with a time stamp corresponding to an earlier value is replaced with a data set group associated with a time stamp corresponding to a more recent value; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point with respect to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix to a minimum value of the singular value, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with a channel of CSI as to whether it corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to an actual position; or classifying based on observable characteristics of a position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, the channel associated with the positioning information is one or more of a flag corresponding to whether it corresponds to an indoor or outdoor UE; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0211] The UE 1200 can be configured to support at least one memory; and at least one processor coupled with the at least one memory and configured to cause the UE to receive, from a network equipment over a physical channel, first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; and receive, from the network equipment over the physical channel, second signaling indicating a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second labels of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset.

[0212] Additionally or alternatively, the UE 1200 can be configured to support cases where the physical channel is a downlink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where at least one of the first labels is a duration including a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event; where the event is based on a configuration for updating the data set being one of a periodic or semi-persistent event; where the event is triggered by at least one of on uplink control information or a UE-based signal multiplexed with the CSI report; where the training data set is updated by replacing data set groups associated with smaller values of the weight with data set groups associated with larger values of the weight; where data set groups associated with time stamps corresponding to more recent values are replaced with data set groups associated with time stamps corresponding to earlier values; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point relative to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or an occurrence probability, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classifying the CSI based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix to a minimum value of the singular value, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with the channel of the CSI as to whether it corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or one or more of values corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of data set points of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of data set points of the first data set group to an actual position; or classifying based on observable characteristics of a location including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, associated with the positioning information, one or more of a flag corresponding to whether the channel is associated with an indoor or outdoor UE; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or one or more of values corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of data set points of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of data set points of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0213] The controller 1206 can manage input and output signals for the UE 1200. The controller 1206 can also manage peripherals not integrated into the UE 1200. In some embodiments, the controller 1206 can utilize an operating system, such as iOS®, ANDROID®, WINDOWS®,

[0214] In some implementations, the UE 1200 can include at least one transceiver 1208. In some other implementations, the UE 1200 can have more than one transceiver 1208. The transceiver 1208 can represent a wireless transceiver. The transceiver 1208 can include one or more receiver chains 1210, one or more transmitter chains 1212, or a combination thereof.

[0215] The receiver chain 1210 can be configured to receive signals (e.g., control information, data, packets) over a wireless medium. For example, the receiver chain 1210 can include one or more antennas to receive signals over the air or wireless medium. The receiver chain 1210 can include at least one amplifier (e.g., a low noise amplifier (LNA)) configured to amplify a received signal. The receiver chain 1210 can include at least one demodulator configured to demodulate a received signal and obtain transmitted data by reversing a modulation technique applied during transmission of the signal. The receiver chain 1210 can include at least one decoder to decode a demodulated signal to receive transmitted data.

[0216] The transmitter chain 1212 can be configured to generate and transmit signals (e.g., control information, data, packets). The transmitter chain 1212 can include at least one modulator to modulate data onto a carrier signal, preparing a signal for transmission over a wireless medium. The at least one modulator can be configured to support one or more techniques, such as amplitude modulation (AM), frequency modulation (FM), or digital modulation schemes like phase-shift keying (PSK) or quadrature amplitude modulation (QAM). The transmitter chain 1212 can also include at least one power amplifier configured to amplify the modulated signal to an appropriate power level for transmission over a wireless medium. The transmitter chain 1212 can also include one or more antennas to transmit the amplified signal into the air or wireless medium.

[0217] Figure 13 An example of a processor 1300 in accordance with aspects of the present disclosure is illustrated. The processor 1300 can be an example of a processor configured to perform various operations in accordance with examples as described herein. The processor 1300 can include a controller 1302 configured to perform various operations in accordance with examples as described herein. The processor 1300 can optionally include at least one memory 1304, which can be, for example, L1 / L2 / L3 cache. Additionally or alternatively, the processor 1300 can optionally include one or more arithmetic logic units (ALUs) 1306. One or more of these components can be in electronic communication or otherwise coupled (e.g., operatively, communicatively, functionally, electronically, electrically) via one or more interfaces, such as buses.

[0218] The processor 1300 can be a processor chipset and include a protocol stack (e.g., software stack) executed by the processor chipset to perform various operations (e.g., receiving, obtaining, retrieving, transmitting, outputting, forwarding, storing, determining, identifying, accessing, writing, reading) in accordance with examples as described herein. The processor chipset can include one or more cores, one or more caches (e.g., memory local to or included in the processor chipset (e.g., processor 1300) or other memory (e.g., random access memory (RAM), read only memory (ROM), dynamic RAM (DRAM), synchronous dynamic RAM (SDRAM), static RAM (SRAM), ferroelectric RAM (FeRAM), magnetic RAM (MRAM), resistive RAM (RRAM), flash memory, phase change memory (PCM), and others)).

[0219] The controller 1302 can be configured to manage and coordinate the various operations (e.g., signaling, receiving, obtaining, retrieving, transmitting, outputting, forwarding, storing, determining, identifying, accessing, writing, reading) of the processor 1300 to enable the processor 1300 to support various operations in accordance with examples as described herein. For example, the controller 1302 can operate as a control unit of the processor 1300, generating control signals that manage the operation of the various components of the processor 1300. These control signals include enabling or disabling functional units, selecting data paths, initiating memory accesses, and coordinating the timing of operations.

[0220] The controller 1302 can be configured to fetch (e.g., obtain, retrieve, receive) instructions from the memory 1304 and determine subsequent instructions to be executed to enable the processor 1300 to support various operations in accordance with examples as described herein. The controller 1302 can be configured to track memory addresses of instructions associated with the memory 1304. The controller 1302 can be configured to decode instructions to determine operations to be performed and operands involved. For example, the controller 1302 can be configured to interpret instructions and determine control signals to be output to other components of the processor 1300 to enable the processor 1300 to support various operations in accordance with examples as described herein. Additionally or alternatively, the controller 1302 can be configured to manage data flow within the processor 1300. The controller 1302 can be configured to control data transfers between registers, the ALU 1306, and other functional units of the processor 1300.

[0221] Memory 1304 can include one or more caches (e.g., memory local to or included in processor 1300, or other memory such as RAM, ROM, DRAM, SDRAM, SRAM, MRAM, flash memory, etc.). In some implementations, memory 1304 can reside within or on a processor chipset (e.g., local to processor 1300). In some other implementations, memory 1304 can reside outside of a processor chipset (e.g., remote from processor 1300).

[0222] Memory 1304 can store computer-readable, computer-executable code including instructions that, when executed by processor 1300, cause processor 1300 to perform various functions described herein. The code can be stored in a non-transitory computer-readable medium such as system memory or another type of memory. Controller 1302 and / or processor 1300 can be configured to execute the computer-readable instructions stored in memory 1304 to cause processor 1300 to perform various functions. For example, processor 1300 and / or controller 1302 can be coupled with or to memory 1304, which processor 1300 and controller 1302 can be configured to perform the various functions described herein. In some examples, processor 1300 can include multiple processors, and memory 1304 can include multiple memories. One or more of the multiple processors can be coupled with one or more of the multiple memories, which can be individually or collectively configured to perform the various functions herein.

[0223] One or more ALUs 1306 can be configured to support various operations in accordance with examples as described herein. In some implementations, one or more ALUs 1306 can reside within or on a processor chipset (e.g., processor 1300). In some other implementations, one or more ALUs 1306 can reside outside of a processor chipset (e.g., processor 1300). One or more ALUs 1306 can perform one or more calculations such as addition, subtraction, multiplication, and division on data. For example, one or more ALUs 1306 can receive input operands and an opcode that specifies the operation to be performed. One or more ALUs 1306 can be configured with various logic and arithmetic circuits, including adders, subtractors, shifters, and logic gates, to process and manipulate data according to the operation. Additionally or alternatively, one or more ALUs 1306 can support logical operations such as AND, OR, exclusive OR (XOR), NOT OR (NOR), and NOT AND (NAND), enabling one or more ALUs 1306 to handle conditional operations, comparisons, and bitwise operations.

[0224] The processor 1300 can support wireless communication in accordance with examples as disclosed herein. The processor 1300 can include at least one controller coupled with at least one memory and can be configured to or operable to cause the processor to: transmit, to a network equipment over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; update the second label after the transmission of the first signaling; update the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset; and transmit, to the network equipment over the physical channel, second signaling indicative of a second training dataset report including updated information corresponding to the updated training dataset.

[0225] Additionally or alternatively, the processor 1300 can be configured to support a case in which the physical channel is an uplink channel; in which the limit of the number of the plurality of data set groups is a maximum value of the number of the data set groups; in which the time or time domain related parameter is at least one of a time stamp or a duration; in which at least one of the first labels is a duration including a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; in which the weight of the data set group is selected from a codebook of values associated with the weight; in which the weight of the data set group is updated based on an event, and the event is one of periodic or semi-persistent based on a configuration for updating the data set, triggered by at least one of a network configuration signal, a downlink control information or a MAC-CE signal, or a combination thereof; in which the training data set is updated by replacing a data set group associated with a smaller value of the weight with a data set group associated with a larger value of the weight; in which a data set group associated with a time stamp corresponding to an earlier value is replaced with a data set group associated with a time stamp corresponding to a more recent value; in which a data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; in which a data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; in which a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point relative to one or more values of other data set points; in which a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or a normal point, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; in which the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and in which a first data set group of the plurality of data set groups is associated with one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or an occurrence probability, in which a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix to a minimum value of the singular value, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with a channel of CSI as to whether it corresponds to a LoS or NLoS channel, a dominant basis index corresponding to an index with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to an actual position; or classifying based on observable characteristics of a position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, one or more of a flag associated with the positioning information as to whether it corresponds to an indoor or outdoor UE; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether a processor is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0226] The processor 1300 can include at least one controller coupled with at least one memory and can be configured or operable to cause the processor to receive, from network equipment over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; and receive, from the network equipment over the physical channel, second signaling indicative of a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset.

[0227] Additionally or alternatively, the processor 1300 can be configured to support a case in which the physical channel is a downlink channel; in which the limit of the number of the plurality of data set groups is a maximum value of the number of the data set groups; in which the time or time domain related parameter is at least one of a time stamp or a duration; in which at least one of the first labels is a duration including a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; in which the weight of the data set group is selected from a codebook of values associated with the weight; in which the weight of the data set group is updated based on an event; in which the event is based on a configuration for updating the data set being one of a periodic or semi-persistent event; in which the event is triggered by at least one of on uplink control information or a UE-based signal multiplexed with the CSI report; in which the training data set is updated by replacing a data set group associated with a smaller value of the weight with a data set group associated with a larger value of the weight; in which a data set group associated with a time stamp corresponding to an earlier value is replaced with a data set group associated with a time stamp corresponding to a more recent value; in which a data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; in which a data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; in which a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point relative to one or more values of other data set points; in which a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; in which the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and in which a first data set group of the plurality of data set groups is associated with one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or an occurrence probability, in which a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix to a minimum value of the singular value, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with a channel of CSI as to whether the channel corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices having a minimum power threshold; wherein the training data set corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of data set points of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of data set points of the first data set group to an actual position; or classifying based on observable characteristics of a location including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, the channel associated with the positioning information is one or more of a flag as to whether the channel corresponds to an indoor or outdoor UE; wherein the training data set corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to a cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of data set points of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of data set points of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether a processor is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0228] Figure 14 An example of a NE 1400 in accordance with aspects of the disclosure is illustrated. The NE 1400 can include a processor 1402, a memory 1404, a controller 1406, and a transceiver 1408. The processor 1402, the memory 1404, the controller 1406, or the transceiver 1408, or various combinations thereof or various components thereof, can be examples of means for performing various aspects of the disclosure as described herein. These components can be coupled (e.g., operatively, communicatively, functionally, electronically, electrically) via one or more interfaces.

[0229] The processor 1402, the memory 1404, the controller 1406, or the transceiver 1408, or various combinations or configurations thereof, can be implemented in hardware (e.g., circuitry), which can be functional hardware. The hardware can be a processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), or other programmable logic device, or any combination thereof, which is configured as or otherwise supports a means for performing the functions described in the present disclosure.

[0230] The processor 1402 can include an intelligent hardware device, (e.g., a general- purpose processor, a DSP, a CPU, an ASIC, an FPGA, or any combination thereof). In some embodiments, the processor 1402 can be configured to operate the memory 1404. In some other embodiments, the memory 1404 can be integrated into the processor 1402. The processor 1402 can be configured to execute computer-readable instructions stored in the memory 1404 to cause the NE 1400 to perform various functions of the present disclosure.

[0231] The memory 1404 can include volatile or non-volatile memory. The memory 1404 can store computer-readable, computer-executable code including instructions that, when executed by the processor 1402, cause the NE 1400 to perform various functions described herein. The code can be stored in a non-transitory computer-readable medium such as the memory 1404 or another type of memory. Computer-readable media include both volatile and non-volatile media, removable and non-removable media, and communication media. Communication media typically embody computer-readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and include any information delivery media. Non-limiting examples of communication media include a wired or wireless network that enables program code to be downloaded into a computer- readable medium.

[0232] In some embodiments, the processor 1402 and memory 1404 coupled with the processor 1402 can be configured to cause the NE 1400 to perform one or more of the functions described herein (e.g., by the processor 1402 executing instructions stored in the memory 1404). For example, the processor 1402 can support wireless communication at the NE 1400 in accordance with examples disclosed herein. The NE 1400 can be configured to support a means for transmitting, to a UE on a physical channel, first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; updating the second label after the transmission of the first signaling; updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset; and transmitting, to the UE on the physical channel, second signaling indicating a second training dataset report including updated information corresponding to the updated training dataset.

[0233] Additionally or alternatively, the NE 1400 can be configured to support cases where the physical channel is a downlink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where the first tag is a duration that includes a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event; where the event is based on a configuration for updating the data set being one of a periodic or semi-persistent event; where the event is triggered by at least one of on uplink control information or a UE-based signal multiplexed with the CSI report; where the training data set is updated by replacing data set groups associated with smaller values of the weight with data set groups associated with larger values of the weight; where data set groups associated with time stamps corresponding to more recent values are replaced with data set groups associated with time stamps corresponding to earlier values; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point relative to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or an occurrence probability, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with the channel of CSI of whether the channel corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to an actual position; or classifying based on observable characteristics of a position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, the channel associated with the positioning information is one or more of a flag corresponding to whether the UE is indoors or outdoors; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag of whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0234] The NE 1400 can be configured to support a means for receiving, from a UE on a physical channel, a first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; and receiving, from the UE on the physical channel, a second signaling indicating a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second labels of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset.

[0235] Additionally or alternatively, the NE 1400 can be configured to support cases where the physical channel is an uplink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where the first tag is a duration comprising parameters corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event, and the event is one of periodic or semi-persistent based on a configuration for updating the data set, triggered by at least one of a network configuration signal, a downlink control information or a MAC-CE signal, or a combination thereof; where the training data set is updated by replacing data set groups associated with smaller values of the weight with data set groups associated with larger values of the weight; where data set groups associated with time stamps corresponding to more recent values are replaced with data set groups associated with time stamps corresponding to earlier values; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point with respect to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or a normal point, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of the weight or the occurrence probability, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix or a singular value of a precoding matrix, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with a channel of CSI as to whether it corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to an actual position; or classifying based on observable characteristics of a position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, the channel associated with the positioning information is one or more of a flag corresponding to whether it corresponds to an indoor or outdoor UE; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0236] The NE 1400 can be configured to support at least one memory; and at least one processor coupled with the at least one memory and configured to cause the base station to: transmit, to a UE on a physical channel, first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; update the second label after transmission of the first signaling; update the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset; and transmit, to the UE on the physical channel, second signaling indicating a second training dataset report including updated information corresponding to the updated training dataset.

[0237] Additionally or alternatively, the NE 1400 can be configured to support cases where the physical channel is a downlink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where the first tag is a duration that includes a parameter corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event; where the event is based on a configuration for updating the data set being one of a periodic or semi-persistent event; where the event is triggered by at least one of on uplink control information or a UE-based signal multiplexed with the CSI report; where the training data set is updated by replacing data set groups associated with smaller values of the weight with data set groups associated with larger values of the weight; where data set groups associated with time stamps corresponding to more recent values are replaced with data set groups associated with time stamps corresponding to earlier values; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point relative to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or an inlier, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or an occurrence probability, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix, or a singular value of a precoding matrix, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with a channel of CSI as to whether it corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to an actual position; or classifying based on observable characteristics of a position including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, the channel associated with the positioning information is one or more of a flag corresponding to whether it corresponds to an indoor or outdoor UE; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or a value corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of a data set point of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of a data set point of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0238] The NE 1400 can be configured to support at least one memory; and at least one processor coupled with the at least one memory and configured to cause the base station to: receive, from the UE over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time-domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset; and receive, from the UE over the physical channel, second signaling indicative of a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second labels of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset.

[0239] Additionally or alternatively, the NE 1400 can be configured to support cases where the physical channel is an uplink channel; where the limit on the number of the plurality of data set groups is a maximum value of the number of data set groups; where the time or time domain related parameter is at least one of a time stamp or a duration; where the first tag is a duration comprising parameters corresponding to at least one of a start time or time interval and a time periodicity, or a time stamp corresponding to one of a transmission time of a data point of the data set group, or a collection time of a data point of the data set group, or a combination thereof; where the weight of the data set group is selected from a codebook of values associated with the weight; where the weight of the data set group is updated based on an event, and the event is one of periodic or semi-persistent based on a configuration for updating the data set, triggered by at least one of a network configuration signal, a downlink control information or a MAC-CE signal, or a combination thereof; where the training data set is updated by replacing data set groups associated with smaller values of the weight with data set groups associated with larger values of the weight; where data set groups associated with time stamps corresponding to more recent values are replaced with data set groups associated with time stamps corresponding to earlier values; where the data set point is associated with a data set group of the plurality of data set groups based on one or more characteristics of the data set point; where the data set point is associated with a data set group of the plurality of data set groups based on an AI-based model maintenance process; where a characteristic of the one or more characteristics of the data set point is an observable characteristic derived via at least one of a deterministic formula of a value of the data set point, a transformed variant of the value of the data set point based on a transformation operation, or a normalization of the value of the data set point with respect to one or more values of other data set points; where a characteristic of the one or more characteristics of the data set point is a non-observable characteristic corresponding to at least one of a parameter identifying whether the data set point is classified as an outlier or a normal point, a statistical dependence parameter corresponding to an approximate distribution associated with the data set, or a parameter corresponding to a power delay profile corresponding to an approximate distribution associated with the data set; where the training data set corresponds to at least one of CSI, precoding information, or beam-based information, and where a first data set group of the plurality of data set groups is associated with one or more of one or more of a time stamp corresponding to a collection time of the CSI, a signaling data time corresponding to the CSI, a time interval for which the first data set group is valid, or a value corresponding to one or more of the weight or the occurrence probability, where a larger value corresponds to a stronger relevance of data set points of the first data set group to the CSI and a smaller value corresponds to a weaker relevance of data set points of the first data set group to the CSI; or classified based on an observable characteristic of the CSI including one or more of a ratio of a maximum value of a channel tap, a channel matrix or a singular value of a precoding matrix, or a power delay profile associated with the CSI, or a non-observable characteristic of the CSI based on the observable characteristic.wherein the observable characteristics are one or more of a number of dominant basis indices based on a transformed frequency domain basis, a flag associated with the channel of the CSI as to whether it corresponds to a LoS or NLoS channel, the dominant basis indices corresponding to indices with a minimum power threshold; wherein the training dataset corresponds to positioning information, and wherein a first data set group of the plurality of data set groups is associated with one or more of a timestamp corresponding to a collection time of the positioning information, a signaling data time corresponding to the positioning information, a time interval for which the first data set group is valid, or one or more of values corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of data set points of the first data set group to an actual position, and a smaller value corresponds to a weaker correlation of data set points of the first data set group to an actual position; or classifying based on observable characteristics of a location including one or more of an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, or unobservable characteristics of an actual position based on the observable characteristics; wherein the observable characteristics are one or more of values based on an angle of arrival, an angle of departure, a round trip time, or a time difference of arrival, associated with the positioning information, one or more of a flag corresponding to whether the channel is associated with an indoor or outdoor UE; wherein the training dataset corresponds to mobility information, and wherein a first data set group of the plurality of data set groups is associated with at least one of a timestamp corresponding to a collection time of the mobility information or cell association information, a signaling data time corresponding to the cell association, a time interval for which the first data set group is valid, or one or more of values corresponding to one or more of a weight or a probability of occurrence, wherein a larger value corresponds to a stronger correlation of data set points of the first data set group to a heuristic cell association or selection, and a smaller value corresponds to a weaker correlation of data set points of the first data set group to a heuristic cell association or selection; or classifying based on observable characteristics of UE mobility including one or more of RSRP, SINR, beam-based information, CSI, or unobservable characteristics of the mobility information based on the observable characteristics; wherein the observable characteristics are a flag as to whether the UE is associated with a best cell based on one or more of a value of RSRP, a value of SINR, beam-based information, or CSI.

[0240] The controller 1406 can manage input and output signals for the NE 1400. The controller 1406 can also manage peripherals not integrated into the NE 1400. In some embodiments, the controller 1406 can employ an operating system, such as iOS®, ANDROID®, WINDOWS®, or other operating system. In some embodiments, the controller 1406 can be implemented as part of a processor 1402.

[0241] In some embodiments, the NE 1400 can include at least one transceiver 1408. In some other embodiments, the NE 1400 can have more than one transceiver 1408. The transceiver 1408 can represent a wireless transceiver. The transceiver 1408 can include one or more receiver chains 1410, one or more transmitter chains 1412, or a combination thereof.

[0242] The receiver chain 1410 can be configured to receive signals (e.g., control information, data, packets) over a wireless medium. For example, the receiver chain 1410 can include one or more antennas to receive signals over the air, or wireless medium. The receiver chain 1410 can include at least one amplifier (e.g., a low noise amplifier (LNA)) configured to amplify a received signal. The receiver chain 1410 can include at least one demodulator configured to demodulate a received signal and obtain transmitted data by reversing the modulation techniques applied during transmission of the signal. The receiver chain 1410 can include at least one decoder to decode the demodulated signal to receive the transmitted data.

[0243] The transmitter chain 1412 can be configured to generate and transmit signals (e.g., control information, data, packets). The transmitter chain 1412 can include at least one modulator to modulate data onto a carrier signal, preparing a signal for transmission over a wireless medium. The at least one modulator can be configured to support one or more techniques, such as amplitude modulation (AM), frequency modulation (FM), or digital modulation schemes, like phase-shift keying (PSK) or quadrature amplitude modulation (QAM). The transmitter chain 1412 can also include at least one power amplifier configured to amplify the modulated signal to an appropriate power level for transmission over a wireless medium. The transmitter chain 1412 can also include one or more antennas to transmit the amplified signal into the air, or wireless medium.

[0244] Figure 15 A flow diagram illustrating a method in accordance with aspects of the disclosure is shown. The operations of the method can be implemented by a UE as described herein. In some embodiments, the UE can execute a set of instructions to control the functional elements of the UE to perform the described functions.

[0245] At 1502, the method can include transmitting, to a network equipment over a physical channel, first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset. The operations of 1502 can be performed according to the examples as described herein. In some embodiments, aspects of the operations of 1502 can be performed by a UE as described with reference to Figure 12 FIG. 9.

[0246] At 1504, the method can include updating the second label after the transmission of the first signaling. The operations of 1504 can be performed according to the examples as described herein. In some embodiments, aspects of the operations of 1504 can be performed by a UE as described with reference to Figure 12 FIG. 9.

[0247] At 1506, the method can include updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset. The operations of 1506 can be performed according to the examples as described herein. In some embodiments, aspects of the operations of 1506 can be performed by a UE as described with reference to Figure 12 FIG. 9.

[0248] At 1508, the method can include transmitting, to a network equipment over a physical channel, second signaling indicating a second training dataset report including updated information corresponding to the updated training dataset. The operations of 1508 can be performed according to the examples as described herein. In some embodiments, aspects of the operations of 1508 can be performed by a UE as described with reference to Figure 12 FIG. 9.

[0249] It should be noted that the methods described herein describe possible implementations, and that the operations and the steps can be rearranged or otherwise modified and that other implementations are possible.

[0250] Figure 16 A flow diagram that describes a method in accordance with an aspect of the present disclosure is illustrated in FIG. 9. The operations of the method can be implemented by a UE as described herein. In some embodiments, a UE can execute a set of instructions to control the functional elements of the UE to perform the described functions.

[0251] At 1602, the method can include receiving, from network equipment over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset. The operations of 1602 can be performed according to examples as described herein. In some embodiments, aspects of the operations of 1602 can be performed by a UE as described with reference to Figure 12 the examples described herein.

[0252] At 1604, the method can include receiving, from network equipment over a physical channel, second signaling indicative of a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset. The operations of 1604 can be performed according to examples as described herein. In some embodiments, aspects of the operations of 1604 can be performed by a UE as described with reference to Figure 12 the examples described herein.

[0253] It should be noted that the methods described herein describe possible implementations, and that the operations and the steps can be rearranged or otherwise modified and that other implementations are possible.

[0254] Figure 17 A flow diagram that illustrates a method in accordance with aspects of the present disclosure is shown. The operations of the method can be implemented by a NE as described herein. In some embodiments, the NE can execute a set of instructions to control the functional elements of the NE to perform the described functions.

[0255] At 1702, the method can include transmitting, to a UE over a physical channel, first signaling indicative of a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset. The operations of 1702 can be performed according to examples as described herein. In some embodiments, aspects of the operations of 1702 can be performed by a NE as described with reference to Figure 14 the examples described herein.

[0256] At 1704, the method can include updating the second label after the transmission of the first signaling. The operations of 1704 can be performed according to examples as described herein. In some embodiments, aspects of the operations of 1704 can be performed by a NE as described with reference to Figure 14 the description.

[0257] At 1706, the method can include updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label for the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset. The operations of 1706 can be performed according to examples as described herein. In some embodiments, aspects of the operations of 1706 can be performed by a NE as described with reference to Figure 14 the description.

[0258] At 1708, the method can include transmitting, to the UE on the physical channel, second signaling indicating a second training dataset report including updated information corresponding to the updated training dataset. The operations of 1708 can be performed according to examples as described herein. In some embodiments, aspects of the operations of 1708 can be performed by a NE as described with reference to Figure 14 the description.

[0259] It should be noted that the methods described herein describe possible implementations, and that the operations and the steps can be rearranged or otherwise modified and that other implementations are possible.

[0260] Figure 18 A flow diagram that illustrates a method in accordance with aspects of the present disclosure is shown. The operations of the method can be implemented by a NE as described herein. In some embodiments, a NE can execute a set of instructions to control the functional elements of the NE to perform the described functions.

[0261] At 1802, the method can include receiving, from a UE on a physical channel, first signaling indicating a first training dataset report identifying a training dataset corresponding to a machine learning or artificial intelligence algorithm, the training dataset including a plurality of data points and partitioned into a plurality of dataset groups each including one or more of the plurality of data points, each of the plurality of dataset groups being associated with a first label and a second label, the first label corresponding to a time or time domain related parameter, the second label being at least one of a weight or a value associated with a characteristic of the dataset. The operations of 1802 can be performed according to examples as described herein. In some embodiments, aspects of the operations of 1802 can be performed by a NE as described with reference to Figure 14 the description.

[0262] At 1804, the method can include receiving, from the UE over the physical channel, a second signaling indicating a second training dataset report including updated information corresponding to an updated training dataset resulting from updating the training dataset based on at least one of the first label or the second label by at least one of updating a subset of values of the second label of the plurality of dataset groups, removing a dataset group of the plurality of dataset groups, or adding a new dataset group to the dataset. The operations of 1804 can be performed according to the examples as described herein. In some embodiments, aspects of the operations of 1804 can be performed by a NE as described with reference to Figure 14 the NE described.

[0263] It should be noted that the methods described herein describe possible implementations, and that the operations and the steps can be rearranged or otherwise modified and that other implementations are possible.

[0264] The description herein is presented to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other variations without departing from the scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A user equipment (UE) for wireless communication, comprising: At least one memory; and At least one processor, coupled to and configured with the at least one memory, to enable the UE to: Transmitting a first signaling instruction over a physical channel to a network device to identify a first training dataset report corresponding to a training dataset of a machine learning or artificial intelligence algorithm, the training dataset comprising multiple data points and partitioned into multiple dataset groups, each containing one or more of the multiple data points, each of the multiple dataset groups being associated with a first label and a second label, the first label corresponding to a time or time-domain related parameter, and the second label being at least one of a weight or value associated with a characteristic of the dataset. Update the second tag after the first signaling is transmitted; The training dataset is updated based on at least one of the first label or the second label by updating a subset of the values ​​of the second label of the plurality of dataset groups, removing a dataset group from the plurality of dataset groups, or adding a new dataset group to the dataset; and On the physical channel, a second signaling is transmitted to the network equipment, indicating a second training dataset report containing updated information corresponding to the updated training dataset.

2. The UE according to claim 1, wherein the physical channel is an uplink channel.

3. The UE according to claim 1, wherein the limit of the number of the plurality of dataset groups is the maximum number of dataset groups.

4. The UE according to claim 1, wherein the time or time-domain related parameter is at least one of a timestamp or a duration.

5. The UE according to claim 1, wherein the first tag is one of the following: This includes the duration of a parameter corresponding to at least one of the start time, or time interval, and time periodicity; A timestamp corresponding to either the transmission time of the data point in the dataset group or the collection time of the data point in the dataset group; or Its combination.

6. The UE of claim 1, wherein the weights of the dataset group are selected from a codebook of values ​​associated with the weights.

7. The UE of claim 1, wherein the weights of the dataset group are based on event updates, and the event: The configuration used to update the dataset is either a periodic or semi-persistent event; Triggered by at least one of the network configuration signal, downlink control information, or media access control element MAC-CE signal; Or a combination thereof.

8. The UE of claim 1, wherein the dataset group associated with a timestamp corresponding to an earlier value is replaced with a dataset group associated with a timestamp corresponding to a more recent value.

9. The UE of claim 1, wherein the dataset point is associated with a dataset group in the plurality of dataset groups based on one or more characteristics of the dataset point.

10. The UE of claim 9, wherein the characteristic among the one or more characteristics of the data point is an observable characteristic derived via at least one of the following: The deterministic formula for the values ​​of the dataset points; Transformed variants of the values ​​of the dataset points based on the transformation operation; or The normalization of the value of the dataset point relative to one or more values ​​of other dataset points.

11. The UE of claim 9, wherein one or more characteristics of the data set point are unobservable characteristics, corresponding to at least one of the following: Parameters for identifying whether the data points in the dataset are classified as outliers or normal points; The statistical correlation parameter corresponding to the approximate distribution associated with the dataset; or The parameters corresponding to the power delay profile that corresponds to the approximate distribution associated with the dataset.

12. The UE of claim 1, wherein the training dataset corresponds to at least one of channel state information (CSI), precoding information, or beam-based information, and wherein the first dataset group of the plurality of dataset groups is at least one of the following: The data is associated with one or more of the following: a timestamp corresponding to the collection time of the CSI; a signaling data time corresponding to the CSI; a valid time interval of the first dataset group; or a value corresponding to one or more of the following: a weight or a probability of occurrence. A larger value corresponds to a stronger correlation between the dataset points of the first dataset group and the CSI, and a smaller value corresponds to a weaker correlation between the dataset points of the first dataset group and the CSI. The CSI is classified based on the ratio of the maximum to the minimum of the singular values ​​of the channel taps, channel matrix, or precoding matrix, the observable characteristics of one or more of the power delay profiles associated with the CSI, or the unobservable characteristics of the CSI based on the observable characteristics.

13. The UE of claim 12, wherein the observable characteristic is relating to the number of dominant base indices based on the transformed frequency domain, one or more of the flags indicating whether the channel associated with the CSI corresponds to a line-of-sight LoS or a non-line-of-sight NLoS channel, and the dominant base index corresponds to an index having a minimum power threshold.

14. The UE of claim 1, wherein the training dataset corresponds to location information, and wherein the first dataset group of the plurality of dataset groups is at least one of the following: The data is associated with one or more of the following: a timestamp corresponding to the collection time of the location information; a signaling data time corresponding to the location information; a valid time interval of the first dataset group; or a value corresponding to one or more of the following: a weight or a probability of occurrence. A larger value corresponds to a stronger correlation between the dataset points of the first dataset group and the actual location, and a smaller value corresponds to a weaker correlation between the dataset points of the first dataset group and the actual location. The location is classified based on observable characteristics of the location, including one or more of the angle of arrival, departure angle, round-trip time, or time difference of arrival, or based on unobservable characteristics of the actual location.

15. The UE of claim 14, wherein the observable characteristic is one or more of the values ​​based on the angle of arrival, the departure angle, the round-trip time, or the time difference of arrival, and the channel associated with the positioning information is an indicator of whether it corresponds to an indoor or outdoor UE.

16. The UE of claim 1, wherein the training dataset corresponds to mobility information, and wherein the first dataset group of the plurality of dataset groups is at least one of the following: The data is associated with one or more of the following: a timestamp corresponding to the collection time of the mobility information or cell association information; a signaling data time corresponding to the cell association; a valid time interval of the first dataset group; or a value corresponding to one or more of the weights or probabilities of occurrence, wherein a larger value corresponds to a stronger correlation between the dataset points of the first dataset group and the heuristic cell association or selection, and a smaller value corresponds to a weaker correlation between the dataset points of the first dataset group and the heuristic cell association or selection; or The UE mobility is classified based on observable characteristics of the mobility information, which include one or more of the following: Reference Signal Received Power (RSRP), Signal-to-Interference-Noise Ratio (SINR), Beam-based information, and Channel State Information (CSI), or unobservable characteristics of the mobility information based on observable characteristics.

17. The UE of claim 16, wherein the observable characteristic is an indicator of whether the UE is associated with an optimal cell based on one or more of the RSRP value, the SINR value, beam-based information, or CSI.

18. A base station for wireless communication, comprising: At least one memory; and At least one processor, coupled to and configured with the at least one memory, to enable the base station to: Transmitting a first signaling instruction to a user equipment (UE) on a physical channel to identify a first training dataset report corresponding to a training dataset of a machine learning or artificial intelligence algorithm, the training dataset comprising multiple data points and partitioned into multiple dataset groups, each containing one or more of the multiple data points, each of the multiple dataset groups being associated with a first label and a second label, the first label corresponding to a time or time-domain related parameter, and the second label being at least one of a weight or value associated with a characteristic of the dataset; Update the second tag after the first signaling is transmitted; The training dataset is updated based on at least one of the first label or the second label by updating a subset of the values ​​of the second label of the plurality of dataset groups, removing a dataset group from the plurality of dataset groups, or adding a new dataset group to the dataset; and On the physical channel, a second signaling is transmitted to the UE indicating a second training dataset report containing updated information corresponding to the updated training dataset.

19. A processor for wireless communication, comprising: At least one controller, coupled to at least one memory and configured to enable the processor to: Transmitting a first signaling instruction over a physical channel to a network device to identify a first training dataset report corresponding to a training dataset of a machine learning or artificial intelligence algorithm, the training dataset comprising multiple data points and partitioned into multiple dataset groups, each containing one or more of the multiple data points, each of the multiple dataset groups being associated with a first label and a second label, the first label corresponding to a time or time-domain related parameter, and the second label being at least one of a weight or value associated with a characteristic of the dataset. Update the second tag after the first signaling is transmitted; The training dataset is updated based on at least one of the first label or the second label by updating a subset of the values ​​of the second label of the plurality of dataset groups, removing a dataset group from the plurality of dataset groups, or adding a new dataset group to the dataset; and On the physical channel, a second signaling is transmitted to the network equipment, indicating a second training dataset report containing updated information corresponding to the updated training dataset.

20. A method performed by a user equipment (UE), the method comprising: Transmitting a first signaling instruction over a physical channel to a network device to identify a first training dataset report corresponding to a training dataset of a machine learning or artificial intelligence algorithm, the training dataset comprising multiple data points and partitioned into multiple dataset groups, each containing one or more of the multiple data points, each of the multiple dataset groups being associated with a first label and a second label, the first label corresponding to a time or time-domain related parameter, and the second label being at least one of a weight or value associated with a characteristic of the dataset. Update the second tag after the first signaling is transmitted; The training dataset is updated based on at least one of the first label or the second label by updating a subset of the values ​​of the second label of the plurality of dataset groups, removing a dataset group from the plurality of dataset groups, or adding a new dataset group to the dataset; and On the physical channel, a second signaling is transmitted to the network equipment, indicating a second training dataset report containing updated information corresponding to the updated training dataset.

Images