Devices and methods of communication
By allowing the terminal device to transmit functionality-related information based on historical data post-configuration, the solution ensures timely and valid model inference, overcoming the lack of historical data after initial or reconfiguration.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- NEC CORP
- Filing Date
- 2024-11-04
- Publication Date
- 2026-05-07
AI Technical Summary
Current technologies face challenges in model inference for time domain prediction as there is no historical information available immediately after an initial or reconfiguration, necessitating a defined UE behavior during this period.
The terminal device is configured to transmit information related to functionality either based on historical data or without it, depending on the availability of historical information, ensuring timely model inference post-configuration.
This approach enables immediate and valid model inference results post-configuration by utilizing historical data when available, addressing the gap in existing technologies.
Smart Images

Figure CN2024129705_07052026_PF_FP_ABST
Abstract
Description
DEVICES AND METHODS OF COMMUNICATIONTECHNICAL FIELD
[0001] Embodiments of the present disclosure generally relate to the field of telecommunication, and in particular, to methods, devices and computer storage media of communication for a model inference.BACKGROUND
[0002] Currently, time domain beam prediction, channel status information (CSI) prediction, and CSI compression with time domain prediction aspects have been identified as use cases related to time domain prediction. Model inference for time domain prediction may give one or more results at one or more future time instances based on one or more historical information. However, there is no historical information that can be used for the model inference within a period of time after an initial configuration or reconfiguration of the model inference, and user equipment (UE) behavior needs to be defined for this period of time.SUMMARY
[0003] In general, embodiments of the present disclosure provide methods, devices and computer storage media of communication for a model inference.
[0004] In a first aspect, there is provided a terminal device. The terminal device comprises a processor configured to cause the terminal device to: receive, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; transmit the first information to the network device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; and perform, before the first occasion, a first operation comprising one of the following: transmitting second information related to the functionality to the network device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, or skipping the transmission of the first information.
[0005] In a second aspect, there is provided a terminal device. The terminal device comprises a processor configured to cause the terminal device to: receive, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; and in accordance with a determination that a third configuration indicating an activation of the first inference is received, transmit, to the network device, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.
[0006] In a third aspect, there is provided a network device. The network device comprises a processor configured to cause the network device to: transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; receive the first information from the terminal device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; and perform, before the first occasion, a second operation comprising one of the following: receiving second information related to the functionality from the terminal device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, or skipping reception of the first information.
[0007] In a fourth aspect, there is provided a network device. The network device comprises a processor configured to cause the network device to: transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; and in accordance with a determination that a third configuration indicating an activation of the first inference is transmitted, receive the first information from the terminal device based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.
[0008] In a fifth aspect, there is provided a method of communication. The method comprises: receiving, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; transmitting the first information to the network device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; and performing, before the first occasion, a first operation comprising one of the following: transmitting second information related to the functionality to the network device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, or skipping the transmission of the first information.
[0009] In a sixth aspect, there is provided a method of communication. The method comprises: receiving, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; and in accordance with a determination that a third configuration indicating an activation of the first inference is received, transmitting the first information to the network device based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.
[0010] In a seventh aspect, there is provided a method of communication. The method comprises: transmitting, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; receiving the first information from the terminal device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; and performing, before the first occasion, a second operation comprising one of the following: receiving second information related to the functionality from the terminal device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, or skipping reception of the first information.
[0011] In an eighth aspect, there is provided a method of communication. The method comprises: transmitting, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; and in accordance with a determination that a third configuration indicating an activation of the first inference is transmitted, receiving, from the terminal device, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.
[0012] In a ninth aspect, there is provided a computer readable medium having instructions stored thereon. The instructions, when executed on at least one processor, cause the at least one processor to perform the method according to any of the fifth to eighth aspects of the present disclosure.
[0013] Other features of the present disclosure will become easily comprehensible through the following description.BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Through the more detailed description of some embodiments of the present disclosure in the accompanying drawings, the above and other objects, features and advantages of the present disclosure will become more apparent, wherein:
[0015] FIG. 1 illustrates an example communication network in which some embodiments of the present disclosure can be implemented;
[0016] FIG. 2A illustrates a schematic diagram illustrating an example inference procedure for beam management (BM) in which some embodiments of the present disclosure can be implemented;
[0017] FIG. 2B illustrates a schematic diagram illustrating an example inference procedure for CSI prediction in which some embodiments of the present disclosure can be implemented;
[0018] FIG. 2C illustrates a schematic diagram illustrating an example inference procedure for CSI compression in which some embodiments of the present disclosure can be implemented;
[0019] FIG. 2D illustrates a schematic diagram illustrating an example scenario of a reporting for model inference in which some embodiments of the present disclosure can be implemented;
[0020] FIG. 3 illustrates a signaling chart illustrating an example process of communication for model inference according to some embodiments of the present disclosure;
[0021] FIG. 4A illustrates a schematic diagram illustrating an example determination of a prediction window and an observation window according to some embodiments of the present disclosure;
[0022] FIG. 4B illustrates a schematic diagram illustrating another example determination of a prediction window and an observation window according to some embodiments of the present disclosure;
[0023] FIG. 4C illustrates a schematic diagram illustrating an example reporting for model inference according to some embodiments of the present disclosure;
[0024] FIG. 5 illustrates a signaling chart illustrating another example process of communication for model inference according to some embodiments of the present disclosure;
[0025] FIG. 6 illustrates a schematic diagram illustrating another example reporting for model inference according to some embodiments of the present disclosure;
[0026] FIG. 7 illustrates a flowchart illustrating an example method of communication implemented at a terminal device in accordance with some embodiments of the present disclosure;
[0027] FIG. 8 illustrates a flowchart illustrating another example method of communication implemented at a terminal device in accordance with some embodiments of the present disclosure;
[0028] FIG. 9 illustrates a flowchart illustrating an example method of communication implemented at a network device in accordance with some embodiments of the present disclosure;
[0029] FIG. 10 illustrates a flowchart illustrating another example method of communication implemented at a network device in accordance with some embodiments of the present disclosure; and
[0030] FIG. 11 is a simplified block diagram of a device that is suitable for implementing embodiments of the present disclosure.
[0031] Throughout the drawings, the same or similar reference numerals represent the same or similar element.DETAILED DESCRIPTION
[0032] Principle of the present disclosure will now be described with reference to some embodiments. It is to be understood that these embodiments are described only for the purpose of illustration and help those skilled in the art to understand and implement the present disclosure, without suggesting any limitations as to the scope of the disclosure. The disclosure described herein can be implemented in various manners other than the ones described below.
[0033] In the following description and claims, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skills in the art to which this disclosure belongs.
[0034] As used herein, the term ‘terminal device’ refers to any device having wireless or wired communication capabilities. Examples of the terminal device include, but not limited to, user equipment (UE) , personal computers, desktops, mobile phones, cellular phones, smart phones, personal digital assistants (PDAs) , portable computers, tablets, wearable devices, Internet of things (IoT) devices, ultra-reliable and low latency communications (URLLC) devices, Internet of everything (IoE) devices, machine type communication (MTC) devices, device on vehicle for V2X communication where X means pedestrian, vehicle, or infrastructure / network, devices for integrated access and backhaul (IAB) , small data transmission (SDT) , mobility, multicast and broadcast services (MBS) , positioning, dynamic / flexible duplex in commercial networks, reduced capability (RedCap) , space borne vehicles or air borne vehicles in non-terrestrial networks (NTN) including satellites and high altitude platforms (HAPs) encompassing unmanned aircraft systems (UAS) , extended reality (XR) devices including different types of realities such as augmented reality (AR) , mixed reality (MR) and virtual reality (VR) , the unmanned aerial vehicle (UAV) commonly known as a drone which is an aircraft without any human pilot, devices on high speed train (HST) , or image capture devices such as digital cameras, sensors, gaming devices, music storage and playback appliances, or Internet appliances enabling wireless or wired Internet access and browsing and the like. The ‘terminal device’ can further has ‘multicast / broadcast’ feature, to support public safety and mission critical, V2X applications, transparent IPv4 / IPv6 multicast delivery, IPTV, smart TV, radio services, software delivery over wireless, group communications and IoT applications. It may also incorporate one or multiple subscriber identity module (SIM) as known as multi-SIM. The term ‘terminal device’ can be used interchangeably with a UE, a mobile station, a subscriber station, a mobile terminal, a user terminal or a wireless device.
[0035] The term ‘network device’ refers to a device which is capable of providing or hosting a cell or coverage where terminal devices can communicate. Examples of a network device include, but not limited to, a Node B (NodeB or NB) , an evolved NodeB (eNodeB or eNB) , a next generation NodeB (gNB) , a transmission reception point (TRP) , a remote radio unit (RRU) , a radio head (RH) , a remote radio head (RRH) , an IAB node, a low power node such as a femto node, a pico node, a reconfigurable intelligent surface (RIS) , network-controlled Repeaters, and the like.
[0036] The terminal device or the network device may have AI / ML capability. It generally includes a model which has been trained from numerous collected data for a specific function, and can be used to predict some information.
[0037] The terminal or the network device may work on several frequency ranges, e.g. FR1 (410 MHz to 7125 MHz) , FR2 (24.25GHz to 71GHz) , frequency band larger than 100GHz as well as Tera Hertz (THz) . It can further work on licensed / unlicensed / shared spectrum. The terminal device may have more than one connections with the network devices under MR-DC application scenario. The terminal device or the network device can work on full duplex, flexible duplex and cross division duplex modes.
[0038] The network device may have the function of network energy saving, self-organizing networks (SON) / minimization of drive tests (MDT) . The terminal device may have the function of power saving.
[0039] The embodiments of the present disclosure may be performed in test equipment, e.g. signal generator, signal analyzer, spectrum analyzer, network analyzer, test terminal device, test network device, channel emulator.
[0040] In one embodiment, the terminal device may be connected with a first network device and a second network device. One of the first network device and the second network device may be a master node and the other one may be a secondary node. The first network device and the second network device may use different radio access technologies (RATs) . In one embodiment, the first network device may be a first RAT device and the second network device may be a second RAT device. In one embodiment, the first RAT device is eNB and the second RAT device is gNB. Information related with different RATs may be transmitted to the terminal device from at least one of the first network device or the second network device. In one embodiment, information A may be transmitted to the terminal device from the first network device and information B may be transmitted to the terminal device from the second network device directly or via the first network device. In one embodiment, information related with configuration for the terminal device configured by the second network device may be transmitted from the second network device via the first network device. Information related with reconfiguration for the terminal device configured by the second network device may be transmitted to the terminal device from the second network device directly or via the first network device.
[0041] As used herein, the singular forms ‘a’ , ‘an’ and ‘the’ are intended to include the plural forms as well, unless the context clearly indicates otherwise. The term ‘includes’ and its variants are to be read as open terms that mean ‘includes, but is not limited to. ’ The term ‘based on’ is to be read as ‘at least in part based on. ’ The term ‘one embodiment’ and ‘an embodiment’ are to be read as ‘at least one embodiment. ’ The term ‘another embodiment’ is to be read as ‘at least one other embodiment. ’ The terms ‘first, ’ ‘second, ’ and the like may refer to different or same objects. The term ‘and / or’ indicates that there may be three relationships. For example, A and / or B may indicate cases includes ‘only A’ , ‘both A and B’ , and ‘only B’ . The term ‘at least one of the following items’ or a similar expression thereof refers to any combination of these items, including any combination of a single item or a plurality of items. For example, the term ‘at least one of A, B, or C’ may represent A, B, C, ‘A and B’ , ‘A and C’ , ‘B and C’ , or ‘A, B and C’ . The term ‘aset of’ may be interchangeably used with ‘one or more’ . Other definitions, explicit and implicit, may be included below.
[0042] In some examples, values, procedures, or apparatus are referred to as ‘best, ’ ‘lowest, ’ ‘highest, ’ ‘minimum, ’ ‘maximum, ’ or the like. It will be appreciated that such descriptions are intended to indicate that a selection among many used functional alternatives can be made, and such selections need not be better, smaller, higher, or otherwise preferable to other selections.
[0043] Currently, there are three use cases related to time domain prediction: time domain beam prediction (also referred to as BM-Case2 herein) , CSI prediction, and CSI compression with temporal domain aspects. For the beam and CSI prediction, one-sided model (i.e., network (NW) -sided or UE-sided model) is assumed. Embodiments of the present disclosure mainly focus on UE-sided model inference. For the CSI compression, two-sided model is assumed with a UE part as a CSI generation part and a NW part as a CSI reconstruction part.
[0044] Usually, a periodic report is configured for reporting one or more prediction results. A model inference is based on one or more measurement results at one or more historical time instances. Immediately after an initial configuration or a reconfiguration of the model inference, UE cannot perform any prediction since there is no historical information that can be used for the model inference. UE behavior may need to be defined for this stage.
[0045] Embodiments of the present disclosure provide solutions of communication for a model inference. In one aspect, a network device may transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality. The first information is associated with a first inference related to the functionality using historical information. The terminal device may transmit the first information to the network device from a first occasion that is an earliest occasion after a first number of time units since first reference time. The first number of time units is based on at least time required for collecting the historical information for the first inference. The terminal device may perform, before the first occasion, a first operation comprising one of the following: transmitting second information related to the functionality to the network device, or skipping the transmission of the first information. The second information is associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information. In this way, an issue that UE cannot have prediction results immediately after an initial configuration or a reconfiguration of a model inference may be solved.
[0046] In another aspect, a network device may transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality. The first information is associated with a first inference related to the functionality using historical information. In accordance with a determination that a third configuration indicating an activation of the first inference is received, the terminal device may transmit, to the network device, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time. The first number of time units is based on at least time required for collecting the historical information for the first inference. In this way, a process to get the first valid report for a model inference may be accelerated.
[0047] For convenience, definitions of some terms in the present disclosure may be listed as below.
[0048] · AI / ML Model: a data driven algorithm that applies AI / ML techniques to generate a set of outputs based on a set of inputs.
[0049] · AI / ML model delivery: a generic term referring to delivery of an AI / ML model from one entity to another entity in any manner. An entity may mean a network node / function (e.g., gNB, location management function (LMF) , etc. ) , UE, proprietary server, etc.
[0050] · AI / ML model inference: a process of using a trained AI / ML model to produce a set of outputs based on a set of inputs.
[0051] · AI / ML model testing: a subprocess of training, to evaluate performance of a final AI / ML model using a dataset different from one used for model training and validation. Differently from AI / ML model validation, testing does not assume subsequent tuning of the model.
[0052] · AI / ML model training: a process to train an AI / ML model (e.g., by learning an input / output relationship) in a data driven manner and obtain the trained AI / ML model for inference.
[0053] · AI / ML model transfer: a delivery of an AI / ML model over an air interface in a manner that is not transparent to the third generation partnership project (3GPP) signalling, either parameters of a model structure known at a receiving end or a new model with parameters. The delivery may contain a full model or a partial model.
[0054] · AI / ML model validation: a subprocess of training, to evaluate quality of an AI / ML model using a dataset different from one used for model training, that helps selecting model parameters that generalize beyond a dataset used for model training.
[0055] · data collection: a process of collecting data by a network node, management entity, or UE for the purpose of AI / ML model training, data analytics and inference.
[0056] · federated learning / federated training: a machine learning technique that trains an AI / ML model across multiple decentralized edge nodes (e.g., UEs, gNBs) each performing local model training using local data samples. The technique requires multiple interactions of the model, but no exchange of local data samples.
[0057] · functionality: the term ‘functionality’ may refer to UE-capability information / parameters e.g., AI / ML-specific feature groups (FGs) . This interpretation may be suitable when ‘supported functionalities’ is used. The term ‘functionality’ may refer to an information element (IE) ‘CSI-ReportConfig’ for inference configuration or a set of inference related parameters or information / parameters indicated by UE. This interpretation may be suitable when ‘applicable functionalities’ is used. The term ‘functionality’ may refer to configurations based on CSI framework. This interpretation may be suitable when ‘activated functionalities’ is used. Therefore, meaning and granularity of the term ‘functionality’ for applicable functionalities, activated functionalities and supported functionalities may or may not be the same.
[0058] · functionality identification: a process / method of identifying an AI / ML functionality for the common understanding between the NW and the UE. Note: Information regarding the AI / ML functionality may be shared during functionality identification. Where AI / ML functionality resides depends on the specific use cases and sub use cases.
[0059] · model activation: enable an AI / ML model for a specific AI / ML-enabled feature. In some embodiments, model activation is equivalent to activation of configurations for model inference.
[0060] · model deactivation: disable an AI / ML model for a specific AI / ML-enabled feature. In some embodiments, model deactivation is equivalent to deactivation of configurations for model inference.
[0061] · model download: Model transfer from the network to UE.
[0062] · model identification: a process / method of identifying an AI / ML model for the common understanding between the NW and the UE. The process / method of model identification may or may not be applicable. Information regarding the AI / ML model may be shared during model identification.
[0063] · model monitoring: a procedure that monitors inference performance of an AI / ML model. In some embodiments, model monitoring is equivalent to performance monitoring, or actions according to configurations for performance monitoring.
[0064] · model parameter update: a process of updating model parameters of a model.
[0065] · model selection: a process of selecting an AI / ML model for activation among multiple models for the same AI / ML enabled feature. Model selection may or may not be carried out simultaneously with model activation. In some embodiments, model selection is equivalent to selection of configurations for model inference.
[0066] · model switching: a process of deactivating a currently active AI / ML model and activating a different AI / ML model for a specific AI / ML-enabled feature. In some embodiments, model switching is equivalent to deactivating a currently active configuration for AI / ML model and activating a different configuration for AI / ML model for a specific AI / ML-enabled feature.
[0067] · model update: a process of updating model parameters and / or model structure of a model.
[0068] · model upload: model transfer from UE to the network.
[0069] · NW-side (AI / ML) model: an AI / ML Model whose inference is performed entirely at the network.
[0070] · offline field data: data collected from field and used for offline training of an AI / ML model.
[0071] · offline training: an AI / ML training process where the model is trained based on collected dataset, and where the trained model is later used or delivered for inference. Note: This definition only serves as a guidance. There may be cases that may not exactly conform to this definition but could still be categorized as offline training by commonly accepted conventions. In some embodiments, model training is equivalent to actions according to configurations for model training.
[0072] · online field data: data collected from field and used for online training of the AI / ML model.
[0073] · online training: an AI / ML training process where the model being used for inference) is (typically continuously) trained in (near) real-time with the arrival of new training samples. The notion of (near) real-time vs. non real-time is context-dependent and is relative to the inference time-scale. This definition only serves as a guidance. There may be cases that may not exactly conform to this definition but could still be categorized as online training by commonly accepted conventions. Fine-tuning / re-training may be done via online or offline training.
[0074] · reinforcement learning (RL) : a process of training an AI / ML model from an input (also known as a state) and a feedback signal (also known as a reward) resulting from the model’s output (also known as an action) in an environment the model is interacting with.
[0075] · semi-supervised learning: a process of training a model with a mix of labelled data and unlabelled data.
[0076] · supervised learning: a process of training a model from input and its corresponding labels.
[0077] · two-sided (AI / ML) model: a paired AI / ML model (s) over which joint inference is performed, where joint inference comprises AI / ML Inference whose inference is performed jointly across the UE and the network, i.e, the first part of inference is firstly performed by UE and then the remaining part is performed by gNB, or vice versa.
[0078] · UE-side (AI / ML) model: an AI / ML Model whose inference is performed entirely at the UE.
[0079] · unsupervised learning: a process of training a model without labelled data.
[0080] · proprietary-format models: ML models of vendor- / device-specific proprietary format, from 3GPP perspective. They are not mutually recognizable across vendors and hide model design information from other vendors when shared. Note: An example is a device-specific binary executable format.
[0081] · open-format models: ML models of specified format that are mutually recognizable across vendors and allow interoperability, from 3GPP perspective. They are mutually recognizable between vendors and do not hide model design information from other vendors when shared.
[0082] In the context of the present disclosure, the terms ‘model’ , ‘functionality’ , ‘model / functionality’ , ‘inference configuration’ , ‘a set of inference related parameters’ and ’AI / ML related information / parameters indicated by UE’ may be used interchangeably. The terms ‘feature’ and ‘feature group’ may be used interchangeably. The terms ‘model’ and ‘model group’ may be used interchangeably. The terms ‘functionality’ , ‘functionality group’ , ‘functionality set’ may be used interchangeably. The terms ‘ID’ , ‘index’ , ‘indicator’ and ‘identifier’ may be used interchangeably.
[0083] In the context of the present disclosure, the term ‘NW’ herein may refer to ‘operations, administration and maintenance (OAM) ’ , ‘server’ , or ‘advanced mobile location (AML) / LMF’ . The term ‘occasion’ herein may be interchangeably used with ‘report occasion’ . The term ‘quantization’ may be interchangeably used with ‘quantizer’ . The term ‘dequantization’ may be interchangeably used with ‘inverse quantizer’ .
[0084] In the context of the present disclosure, the term ‘conditions’ may refer to configurations supported indicated via UE capability reporting, related to model training, model inference, performance monitoring, validation procedure, fallback, of an AI / ML model / functionality or a group of models / functionalities. The term ‘additional conditions’ may refer to, e.g., application conditions, scenarios, sites, datasets, cell identity (ID) , timestamp, beam shape, etc. For an AI / ML-enabled feature / feature group (FG) , additional conditions refer to any aspects that are assumed for a training of a model but are not a part of UE capability for the AI / ML-enabled feature / FG. It does not imply that additional conditions are necessarily specified. Additional conditions can be divided into two categories: NW-side additional conditions and UE-side additional conditions. The term ‘applicable conditions’ may refer to signal-noise ratio (SNR) , line-of-sight (LOS) / non-line-of-sight (NLOS) , channel conditions, etc. The term ‘UE’ internal conditions’ may refer to, e.g., memory, battery, and other hardware limitations on functionality / model operations. The term ‘complexity / processing capability’ may refer to Tera operations per second (TOPs) , floating point operations (FLOPs) , multiply accumulate operations (MACs) , or the number of parameters, and / or a size.
[0085] In the context of the present disclosure, the terms ‘beam’ , ‘precoder’ , ‘precoding’ , ‘precoding matrix’ , ‘spatial relation information’ , ‘spatial relation info’ , ‘precoding information’ , ‘precoding information and number of layers’ , ‘precoding matrix indicator (PMI) ’ , ‘precoding matrix indicator’ , ‘transmission precoding matrix indication’ , ‘precoding matrix indication’ , ‘transmission configuration indication (TCI) state’ , ‘uplink (UL) TCI state’ , ‘joint TCI state’ , ‘transmission configuration indicator’ , ‘quasi co-location (QCL) ’ , ‘quasi-co-location’ , ‘QCL parameter’ , ‘QCL assumption’ , ‘QCL relationship’ and ‘spatial relation’ can be used interchangeably. A beam may refer to a downlink (DL) beam, UL beam, transmitting (Tx) beam, receiving (Rx) beam, beam pair, reference signal (RS) resource, RS resource set, antenna port, antenna port group, antenna element (s) , antenna array (s) , or beam group.
[0086] The terms ‘reference signal received power (RSRP) ’ , ‘layer 1 (L1) -RSRP’ , ‘layer 3 (L3) -RSRP’ , ‘filtered RSRP’ herein may be used interchangeably. If ‘RSRP’ is used as a beam quality metric, methods are readily extended to other metrics like ‘signal to interference and noise ratio (SINR) ’ , ‘reference signal received quality (RSRQ) ’ , ‘received signal strength indicator (RSSI) ’ , etc. The term ‘CSI-RS for beam management’ and ‘CSI-RS configured in a resource set with higher layer parameter ‘repetition” may be used interchangeably. The term ‘CSI-RS for channel acquisition’ and ‘CSI-RS configured in a resource set without higher layer parameter ‘repetition’ and without higher layer parameter ‘trs-info” may be used interchangeably.
[0087] In the context of the present disclosure, the term ‘historical measurement result’ may refer to one of the following options:
[0088] · Option 1: based on number of measurements (denoted as Pt) , number of RSs (denoted as Mt) and prediction time (denoted as T2) . T2 means a time duration for prediction. Mt means the number of time instances for measurement as AI / ML inputs with a periodicity of Tper. Pt means the number of time instance (s) for prediction with a periodicity of Tper in T2.
[0089] · Option 2: based on a periodicity (denoted as T) of required reference signals for measurements. For every T=Y ms, reference signals for measurements are needed.
[0090] · Option 3: based on times (denoted as Z) of a given minimal periodicity Tper of reference signals for measurements. UE may measure reference signals for model inputs every Z times of Tper.
[0091] · Option 4: based on an observation window (e.g., number / distance) , e.g., 5 / 5ms (e.g., 5 times observations and 5ms between two observations, or 5 times observations in 5ms) , or 10 / 5ms (e.g., 10 times observations and 5ms between two observations, or 10 times observations in 5ms) .
[0092] It is to be noted that the term ‘periodicity of Tper’ , ‘T’ , or ‘distance’ mentioned above may be called as an interval or time interval between two historical measurements, or between two measurements for historical results. The time duration related to Mt×Tper, Y ×Tper, or ‘distance’ , or ‘observation window’ mentioned above may be called as a measurement window for historical measurement results.
[0093] In the context of the present disclosure, the term ‘future time instance’ may refer to one of the following options:
[0094] · Option 1: N future time instance (s) that is based on an output of AI / ML model inference. The future time instance may include information about a timestamp, or an interval between two joint time instances.
[0095] · Option 2: based on number of measurements (denoted as Pt) , number of RSs (denoted as Mt) and prediction time (denoted as T2) . T2 means a time duration for beam prediction. Mt means the number of time instances for measurement as AI / ML inputs with a periodicity of Tper. Pt means the number of time instance (s) for prediction with a periodicity of Tper in T2.
[0096] · Option 3: based on a periodicity (denoted as T) of required reference signals for measurements. For every T=Y ms, reference signals for measurements are needed.
[0097] · Option 4: based on times (denoted as Z) of a given minimal periodicity Tper of reference signals for measurements. UE may measure reference signals for model inputs every Z times of Tper. Prediction time is defined as the time from each measurement instance to the latest prediction instance before the next measurement instance.
[0098] · Option 5: based on a prediction window (e.g., number / distance between prediction instances / distance from the last observation instance to the first or starting prediction instance) , e.g., 1 / 5ms / 5ms.
[0099] It is to be noted that the term ‘periodicity of Tper’ , ‘distance’ mentioned above may be called as an interval or time interval between two future time instances. The time duration related to Pt×Tper, ‘Z’ , ‘distance’ , number*distance may be called as prediction window for future time instances.
[0100] As used herein, a model may be equivalent to at least one of the following: an AI / ML model, a ML model, an AI model, a data-driven, a data processing model, an algorithm, a functionality, a procedure, a process, an entity, a function, a feature, a feature group, a model identifier (ID) , an ID, a functionality ID, a configuration ID, a scenario ID, a site ID, a dataset ID, a set of AI / ML related parameters, or an ID of a set of AI / ML related parameters. As a result, the above terms may be used interchangeably.
[0101] In some embodiments, the model may be represented by or associated with a channel, a resource, a resource set, a RS resource, a RS resource set, a RS port, a set of RS ports, a RS port ID, or a set of RS port IDs.
[0102] In some embodiments, the model may comprise a set of weights values that may be learned during training, for example for a specific architecture or configuration, where a set of weights values may also be called a parameter set.
[0103] In some embodiments, the model may be used to predict a target cell, or measurements of a set of beams of a set of candidate cells in future based on at least historical measurements (e.g., layer 1 (L1) -reference signal received power (RSRP) , or L1-signal to interference plus noise ratio (SINR) ) of a set of beams of a set of candidate cells.
[0104] In some embodiments, an input of the ML model (i.e., AI input) may refer to the input of a model and indicate data inputted into the model, which may be equivalent to data.
[0105] In some embodiments, an output of ML model (i.e., AI output) may refers to the output of a model and indicate result (s) outputted by the model, which is equivalent to label / data.
[0106] In some embodiments, a ground truth label of data (or ground-truth label) for monitoring or training the ML model (i.e., AI output) may refers to the authoritative, accepted data, or true answer or outcome for AI / ML model.
[0107] In some embodiments, ‘ground truth’ , ‘ground truth label’ , ‘ground truth label of data’ , ‘input label’ , ‘input data’ and ‘data’ can be used interchangeably.
[0108] In some embodiments, the ground truth can be interpreted as actual / factual (i.e. actual / factual measured) data / values / results / collections / parameters, which can be used as reference, compared to prediction or inference.
[0109] Principles and implementations of the present disclosure will be described in detail below with reference to the figures.
[0110] EXAMPLE OF COMMUNICATION NETWORK
[0111] FIG. 1 illustrates an example communication network 100 in which embodiments of the present disclosure can be implemented. As shown in Fig. 1, the communication network 100 includes a terminal device 110 and a network device 120 served by the terminal device 110.
[0112] As shown in FIG. 1, the terminal device 110 may have a plurality of beams, and the network device 120 may have a plurality of beams. A channel (or called as a sub-channel) may be formed between one of the beams of the terminal device 110 and one of the beams of the network device 120. The terminal device 110 may transmit information to the network device 120 or receive information from the network device 120 via one or more of the sub-channels.
[0113] It is to be understood that the number of devices and beams in FIG. 1 is given for the purpose of illustration without suggesting any limitations to the present disclosure. The communication network 100 may include any suitable number of network devices and / or terminal devices and / or beams adapted for implementing implementations of the present disclosure.
[0114] The communications in the communication network 100 may conform to any suitable standards including, but not limited to, global system for mobile communications (GSM) , long term evolution (LTE) , LTE-evolution, LTE-advanced (LTE-A) , new radio (NR) , wideband code division multiple access (WCDMA) , code division multiple access (CDMA) , GSM EDGE radio access network (GERAN) , machine type communication (MTC) and the like. The embodiments of the present disclosure may be performed according to any generation communication protocols either currently known or to be developed in the future. Examples of the communication protocols include, but not limited to, the first generation (1G) , the second generation (2G) , 2.5G, 2.75G, the third generation (3G) , the fourth generation (4G) , 4.5G, the fifth generation (5G) communication protocols, 5.5G, 5G-advanced networks, or the sixth generation (6G) networks.
[0115] Communication in a direction from the terminal device 110 towards the network device 120 is referred to as UL communication, while communication in a reverse direction from the network device 120 towards the terminal device 110 is referred to as DL communication. A wireless communication channel may comprise a physical uplink control channel (PUCCH) , a physical uplink shared channel (PUSCH) , a physical random-access channel (PRACH) , a physical downlink control channel (PDCCH) , a physical downlink shared channel (PDSCH) and a physical broadcast channel (PBCH) .
[0116] FIG. 2A illustrates a schematic diagram 200A illustrating an example inference procedure for BM in which some embodiments of the present disclosure can be implemented. As shown in FIG. 2A, for an inference procedure for BM, measurements based on a set of beams (also referred to as Set B herein) are used as a model input of an AI / ML model 210 to predict information of another set of beams (also referred to as Set A herein) . In addition, beam ID information may be also provided as an input to an AI / ML model. Based on model output (e.g., probability of each beam in Set A to be the Top-1 beam, predicted L1-RSRP) of the AI / ML model 210, Top-1 / Top-N beam (s) among Set A of beams can be predicted and / or potentially with predicted L1-RSRPs (depending on a labeling) . For BM-Case 1, the measurements of Set B are used as a model input to predict Top-1 / Top-N beams from Set A. For BM-Case 2, the measurements of Set B at historic time instance (s) are used as a model input for temporal DL beam prediction of beams from Set A. The case that Set A and Set B are different (Set B is not a subset of Set A) , and Set B is a subset of Set A for both BM-Case1 and BM-Case2, and the case that Set A and Set B are the same for BM-Case2 are considered. Performance of DL Tx beam prediction and DL Tx-Rx beam pair prediction may be evaluated.
[0117] For both BM-Case 1 and BM-Case 2, a terminal device may report a prediction result to NW based on an output of a UE-sided model, or NW may predict the Top-1 / Top-N beam (s) based on the reported measurements of Set B for a NW-sided model. BM-Case 2 may also be called as a time domain beam prediction herein.
[0118] FIG. 2B illustrates a schematic diagram 200B illustrating an example inference procedure for CSI prediction in which some embodiments of the present disclosure can be implemented. In this example, a time domain CSI prediction using a UE-sided model is used. As shown in FIG. 2B, historical CSI information is used as a model input of a CSI prediction model 220, and predicted CSI is obtained as a model output of the CSI prediction model 220. For example, the model inputs may be K (K≥1) historical CSI information, which may be in following form (s) : a precoding matrix or raw channel matrix in spatial-frequency domain using angular-delay domain projection; or a rank indication (RI) , a precoding matrix indicator (PMI) , or a channel quality indicator (CQI) . The model outputs may be predicted CSI of F (F≥1) future time instances, which may be in following form (s) : a precoding matrix or raw channel matrix in spatial-frequency domain using angular-delay domain projection; or RI, PMI, or CQI. For generating the model input of the CSI prediction model 220, some further pre-processing on a measured channel may be applied. For the model output of the CSI prediction model 220, some further post-processing may also be applied.
[0119] FIG. 2C illustrates a schematic diagram 200C illustrating an example inference procedure for CSI compression in which some embodiments of the present disclosure can be implemented. In this example, a time-spatial-frequency domain CSI compression using two-sided model is used. As shown in FIG. 2C, a CSI generation model (also referred to as a CSI generation part or an encoder) 230 (i.e., UE-sided model) and a CSI reconstruction model (also referred to as a CSI reconstruction part or a decoder) 240 (i.e., NW-sided model) may be applied. The CSI generation model 230 may comprise an inference 231 for CSI generation and quantization 232, and the CSI reconstruction model 240 may comprise dequantization 241 and an inference 242 for CSI reconstruction.
[0120] For generating an input of the CSI generation model 230, some further pre-processing on a measured channel may be applied. For an output of the CSI reconstruction model 240, some further post-processing may also be applied. Besides CSI feedback as an output of the quantization 232, there may also be other CSI / PMI related information transmitted. It is to be noted that there may be other examples of the CSI generation model 230 and / or the CSI reconstruction model 240, e.g., merging the quantization 232 into the inference 231 for CSI generation, and / or merging the dequantization 241 into the inference 242 for CSI reconstruction.
[0121] In some embodiments, the input of the CSI generation model 230 may comprise measured CSI (also referred to as present CSI herein) . Accordingly, an output of the CSI generation model 230 may comprise compressed present CSI without historical CSI information. In some embodiments, the input of the CSI generation model 230 may comprise the measured CSI and historical CSI information. Accordingly, the output of the CSI generation model 230 may comprise compressed present CSI with historical CSI information. In some embodiments, the input of the CSI generation model 230 may comprise predicted CSI (also referred to as future CSI herein) for a set of future time instances. Accordingly, the output of the CSI generation model 230 may comprise compressed future CSI without historical CSI information. In some embodiments, the input of the CSI generation model 230 may comprise the predicted CSI and the historical CSI information. Accordingly, the output of the CSI generation model 230 may comprise compressed future CSI with historical CSI information.
[0122] For example, a CSI type for the input of the CSI generation model 230 may comprise raw channel matrix, eigenvector (s) of the raw channel matrix, feedback CSI information, etc. The input of the CSI generation model 230 may also comprise assumptions on an observation window, i.e., number / time distance of historic CSI / channel measurements.
[0123] For example, a CSI type for the output the CSI reconstruction model 240 may comprise channel matrix, eigenvector (s) , feedback CSI information, etc. The output the CSI reconstruction model 240 may also comprise assumptions on a prediction window, i.e., number / time distance of predicted CSI / channel.
[0124] Time domain aspects may be considered for the CSI compression. In one aspect, whether historical CSI information is used in the CSI compression may be considered. For example, a recurrent spatial-frequency-time encoder / decoder may be used for the CSI generation model 230 / the CSI reconstruction model 240. Alternatively, a spatial-frequency encoder / decoder and a recurrent / differential quantization may be used for the CSI generation model 230 / the CSI reconstruction model 240. In another aspect, whether future CSI is involved in the CSI compression may be considered. Prediction of the future CSI may be a separate step or jointly with CSI compression. For example, the future CSI may be obtained by AI / ML, e.g., CSI prediction. Alternatively, the future CSI may be obtained by non-AI / ML prediction, e.g., auto-regression (AR) channel.
[0125] For these inference procedures as described in FIGs. 2A to 2C, a terminal device may be configured to report an inference result to NW, e.g., in case of a UE-sided model. However, for the first few configured occasions, the terminal device cannot have prediction results. Details will be described in connection with FIG. 2D.
[0126] FIG. 2D illustrates a schematic diagram 200D illustrating an example scenario of a reporting for model inference in which some embodiments of the present disclosure can be implemented. As shown in FIG. 2D, the terminal device may receive a configuration for the reporting of the model inference at a timing T1. There are no prediction results for a set of report occasions 251, and prediction results are available for a set of report occasions 252 subsequent to the set of report occasions 251.
[0127] In this case, a report behavior of the terminal device for the set of report occasions 251 is unclear. For example, it is unclear whether the terminal device needs to report or what the terminal device can report. Further, NW scheduling is unclear since NW may have no valid beam / CSI report for the set of report occasions 251.
[0128] In view of this, embodiments of the present disclosure provide solutions for a model inference so as to overcome the above and other potential issues. The solutions will be described in detail with reference to FIGs. 3 to 6 below.
[0129] EXAMPLE IMPLEMENTATION OF REPORTING FOR MODEL INFERENCE
[0130] Embodiments of the present disclosure provide a solution of a reporting for the model inference. The solution will be described in connection with FIG. 3 below.
[0131] FIG. 3 illustrates a signaling chart illustrating an example process 300 of communication for model inference according to some embodiments of the present disclosure. For the purpose of discussion, the process 300 will be described with reference to FIG. 1. The process 300 may involve the terminal device 110 and the network device 120 as illustrated in FIG. 1. It is to be understood that the steps and the order of the steps in FIG. 3 are merely for illustration, and not for limitation. For example, the order of the steps may be changed. Some of the steps may be omitted or any other suitable additional steps may be added.
[0132] As shown in FIG. 3, at step 310, the terminal device 110 may receive, from the network device 120, a configuration (also referred to as a first configuration or an inference configuration herein) indicating a transmission or report of information (also referred to as first information herein) related to a functionality. The first information is associated with an inference (also referred to as a first inference herein) related to the functionality using historical information.
[0133] In some embodiments, as shown in step 311, the network device 120 may transmit, to the terminal device 110, a message for enquiring capability of the terminal device 110 to initiate a procedure of reporting one or more supported functionalities by the terminal device 110. For example, the message may be a UE capability enquiry message or any other suitable messages.
[0134] As shown in step 312, the terminal device 110 may transmit, to the network device 120, a message indicating one or more supported functionalities at the terminal device 110. For example, the message may be a UE capability information message or any other suitable messages.
[0135] As shown in step 313, the network device 120 may transmit one or more configurations to the terminal device 110. For example, the network device 120 may transmit a radio resource control (RRC) reconfiguration message comprising the one or more configurations. In some embodiments, the one or more configurations may comprise information indicating that the terminal device 110 is allowed to do a UE assistance information (UAI) reporting via an IE ‘OtherConfig’ . In some embodiments, the one or more configurations may comprise NW-side additional conditions. In some embodiments, the one or more configurations may comprise a configuration (e.g., the inference configuration) of the one or more supported functionalities.
[0136] In some embodiments, the network device 120 may configure one or more CSI report configurations (e.g., IE ‘CSI-ReportConfig’ ) for the inference configuration. An associated ID of each CSI report configuration in one or more CSI report configurations may be configured in CSI framework.
[0137] In some embodiments, the network device 120 may configure one set or multiple sets of inference related parameters. It is to be noted that the set of inference related parameters is not configured by the IE ‘CSI-ReportConfig’ .
[0138] In some embodiments, the set of inference related parameters may at least comprise information related to Set A, information related to Set B, and information related to report content. In some embodiments for BM-Case 2, the set of inference related parameters may comprise information related to one or more time instances for measurements and information related to one or more time instances for prediction.
[0139] In some embodiments, one or more associated IDs may be configured. In some embodiments, the one or more associated IDs may be part of one set of the inference related parameters. In some embodiments, the one or more associated IDs may be independently from the one set of the inference related parameters.
[0140] In some embodiments, the one or more associated IDs may be provided to the terminal device 110, e.g., via a RRC parameter.
[0141] As shown in step 314, the terminal device 110 may report one or more applicable functionalities to the network device 120. In some embodiments, the terminal device 110 may report the one or more applicable functionalities if the terminal device 110 is configured to provide the one or more applicable functionalities and the one or more applicable functionalities are changed via UAI. In some embodiments, the terminal device 110 may report the one or more applicable functionalities as a response to NW-side additional conditions requesting applicable functionality reporting, and / or other network configurations (e.g., the inference configuration) in the step 313, via a signaling such as UAI or via a RRC reconfiguration complete message.
[0142] In some embodiments, the terminal device 110 may report applicability of the above one or more CSI report configurations. In some embodiments, the terminal device 110 may report applicability of the above one or multiple sets of inference related parameters. In some embodiments, the terminal device 110 may report the applicable one or multiple sets of inference related parameters. In some embodiments, the set of inference related parameters may at least comprise information related to Set A, information related to Set B, and information related to report content. In some embodiments for BM-Case 2, the set of inference related parameters may comprise information related to one or more time instances for measurements and information related to one or more time instances for prediction.
[0143] It is to be noted that if a reporting of the inference related parameters is not supported, only the applicability of the above one or more CSI report configurations is reported or no information is reported in the step 314. In some embodiments, one or more associated IDs may be configured. In some embodiments, the one or more associated IDs may be part of one set of the inference related parameters. In some embodiments, the one or more associated IDs may be independently from the one set of the inference related parameters.
[0144] In some embodiments, as shown in step 315, after the reporting of the one or more applicable functionalities by the terminal device 110, the network device 120 may configure or update the inference configuration (e.g., if the inference configuration is already provided in the step 313) . In some embodiments, an applicable functionality may be activated by receiving its inference configuration when the applicable functionality is provided in the step 315. In other words, if the inference configuration for a functionality is received, the terminal device 110 may consider that the functionality is activated. Accordingly, if the inference configuration for a functionality is transmitted, the network device 120 may consider that the functionality is activated.
[0145] In some embodiments, the network device 120 may also configure or update model training and performance monitoring configurations. In some embodiments, an initial state of an applicable functionality may be an activated state or a deactivated state. In some embodiments, an initial state of the inference configuration in the step 313 may be an activated state or a deactivated state.
[0146] In some embodiments, if an initial state of an applicable functionality and / or an inference configuration related to the applicable functionality is the deactivated state, an additional L1 / L2 signaling may be used for activation or deactivation of the applicable functionality and / or the inference configuration related to the applicable functionality.
[0147] In some embodiments, one or multiple applicable functionalities may be activated at the same time, based on capability of the terminal device 110 or UAI report.
[0148] It can be seen that the network device 120 may provide the inference configuration (i.e., the first configuration) in at least one of the step 313 or the step 315. In some embodiments, the first configuration may comprise at least a first report configuration and a first resource configuration.
[0149] In some embodiments, the first report configuration may comprise a first report quantity. In some embodiments, the first report quantity may comprise a prediction result, e.g., predicted beam or predicted CSI. In some embodiments for BM-case 2, the first report quantity may be one or configured number of predicted synchronization signal and physical broadcast channel block (SSB) resource indicator (SSBRI) / CSI-RS resource indicator (CRI) and / or predicted RSRP associated with the SSBRI / CRI. In some embodiments for CSI prediction, the first report quantity may be one or configured number of predicted SSBRI / CRI and / or predicted RI / PMI / CQI. In some embodiments for CSI compression, the first report quantity may comprise an inference result, e.g., compressed present or future CSI (e.g., compressed present or future CSI with historical CSI information) which is obtained by using historical CSI information.
[0150] In some embodiments, the first report configuration may comprise time domain information for the report of the first information. In some embodiments, the time domain information may comprise a periodicity (e.g., denoted as P_report) and an offset (e.g., denoted as Offset_report) .
[0151] It is to be noted that the first report quantity and the time domain information may adopt any other suitable forms, and the first report configuration may comprise any other suitable information or information combinations.
[0152] In some embodiments, the first resource configuration may comprise one or more RS resource sets and RS resources for measurement, such as channel status information-reference signal (CSI-RS) , SSB, etc. In some embodiments for BM-case 2, Set B and Set A are configured as two RS resource sets. In some embodiments for CSI prediction, one or more CSI-RS resources are configured.
[0153] In some embodiments, the first resource configuration may comprise time domain information of resources, such as, periodicity (e.g., denoted as P_RS) and offset (e.g., denoted as Offset_RS) . In some embodiments, P_report = P_RS. In some embodiments, P_report is an integer multiple of P_RS. In some embodiments, Offset_report > Offset_RS.
[0154] An example configuration for CSI compression is described in Table 1 below.
[0155] Table 1
[0156] For case 0, the terminal device 110 reports compressed present CSI without historical CSI information. For case 1, the terminal device 110 reports compressed present CSI with historical CSI information. For case 2, the terminal device 110 reports compressed present CSI with historical CSI information. For case 3, the terminal device 110 reports compressed future CSI with historical CSI information. For case 4, the terminal device 110 reports compressed future CSI with historical CSI information. For case 5, the terminal device 110 reports compressed present CSI without historical CSI information.
[0157] In some embodiments, one or more additional configurations for additional model inference related information may also be provided in the inference configuration, such as a configuration for input related information and / or a configuration for output related information. In some embodiments, the input related information may indicate an observation window (e.g., denoted as T_observation) . In some embodiments, the observation window may be associated with number (e.g., denoted as Mt) of historical time instances within the observation window and a time interval (e.g., denoted as T_hist) between two neighbor historical time instances. In some embodiments, the observation window may be represented or configured in the form of Mt and T_hist. In some embodiments, both T_observation and {Mt, T_hist} are provided. In some embodiments, T_observation ≥ (Mt -k) × T_hist, where k is an integer greater than or equal to zero. In some embodiments, T_hist = P_RS. In some embodiments, T_hist is an integer multiple of P_RS. In some embodiments, Mt historical time instances are consecutive. In some embodiment, Mt historical time instances are not consecutive, and an indication of Mt historical time instances in a larger set of historical time instances is needed.
[0158] In some embodiments, the output related information may indicate a prediction window (e.g., denoted as T_prediction) . In some embodiments, the prediction window may be associated with number (e.g., denoted as N) of future time instances within the prediction window and a time interval (e.g., denoted as T_future) between two neighbor future time instances. In some embodiments, the prediction window may be represented or configured in the form of N and T_future. In some embodiments, both T_prediction and {N, T_future} are provided. In some embodiments, T_prediction ≤ (N -l) × T_future, where l is an integer greater than or equal to zero. In some embodiments, T_future = P_report. In some embodiments, T_future is an integer multiple of P_report. In some embodiments, N future time instances are consecutive. In some embodiment, N future time instances are not consecutive, and an indication of N future time instances in a larger set of future time instances is needed.
[0159] In some embodiments, T_observation ≥ T_prediction. In some embodiments, T_observation is an integer multiple of T_prediction. In some embodiments, Mt ≥ N. In some embodiments, Mt is an integer multiple of N.
[0160] In some embodiments where the functionality is CSI compression, the additional model inference related information may also comprise information (e.g., size, dimension) related to an encoder / decoder, such as a latent vector, a bit vector, or number of compressed bits. In some embodiments where the functionality is CSI compression, the additional model inference related information may also comprise information related to a quantizer and / or an inverse quantizer, such as a type (e.g., scalar, vector, recurrent, independent, recurrent neural network (RNN) -based, long short term memory (LTSM) -based) , or based codebook, or differential quantization codebook.
[0161] In some embodiments, the additional model inference related information such as the input related information and / or output related information may be associated with a functionality or model that the terminal device 110 supports. In this case, these additional model inference related information may not be explicitly configured since it is already known at the terminal device 110.
[0162] Both the observation window and the prediction window are in the form of a time duration, and thus starting time points and an ending time points of the observation window and the prediction window need to be defined. Some example embodiments for definition of the starting and ending time points will be described in connection with FIGs. 4A and 4B below.
[0163] In some embodiments, the ending time point of the observation window may be the same as the starting time point of the prediction window configured for the first reference. In other words, the start time point of the i-th prediction window and the end time point of the i-th observation window is the same, and called as the i-th reference timing (e.g., denoted as t_ref2_i) . FIG. 4A illustrates a schematic diagram 400A illustrating an example determination of a prediction window and an observation window according to some embodiments of the present disclosure. As shown in FIG. 4A, T_observation denotes an observation window, T_prediction denotes a prediction window, and t_ref2_i denotes an ending time point of the observation window and a starting time point of the prediction window. The terminal device 110 may use measurement results in a time window (t_ref2_i -T_observation, t_ref2_i) to predict a beam or CSI in a time window (t_ref2_i, t_ref2_i +T_prediction) .
[0164] In some embodiments, the ending time point of the observation window or the starting time point of the prediction window may be based on a configured time of the transmission of the first information associated with the prediction window or the observation window. In other words, the i-th reference timing t_ref2_i may be based on the configured time of the i-th report. For example, as shown in FIG. 4A, the i-th reference timing t_ref2_i may be based on X1 time units before the configured time of the i-th report.
[0165] In the context of the present disclosure, a configured time of a report may mean that for a periodic or semi-persistent CSI report, a periodicity TCSI (measured in slots) and a slot offset Toffset are configured by a higher layer parameter ‘reportSlotConfig’ . Unless specified otherwise, the terminal device 110 may transmit the CSI report in frames with SFN nf and slot number within the frame satisfying where μ is a subcarrier spacing (SCS) configuration of a UL bandwidth part (BWP) the CSI report is transmitted on. In some embodiments, the periodicity TCSI (measured in slots) and the slot offset TCSI may be P_report and Offset_report respectively.
[0166] In some embodiments, the ending time point of the observation window or the starting time point of the prediction window may be based on a configured time of a latest reference signal before the configured time of the transmission of the first information associated with the prediction window or the observation window. In other words, the ending time point of the i-th observation window or the starting time point of the i-th prediction window may be based on the configured time of the latest RS before the configured time of the i-th report. For example, as shown in FIG. 4A, the ending time point of the i-th observation window or the starting time point of the i-th prediction window may be based on X2 time units after the configured time of the latest RS before the configured time of the i-th report. The configured time of the latest RS may be X3 time units before the configured time of the i-th report.
[0167] In some alternative embodiments, the ending time point of the observation window may be different from the starting time point of the prediction window, and a time interval between the ending time point of the observation window and the starting time point of the prediction window may depend on at least time used for the first inference. FIG. 4B illustrates a schematic diagram 400B illustrating another example determination of a prediction window and an observation window according to some embodiments of the present disclosure. As shown in FIG. 4B, a reference timing t_ref3_i corresponds to an ending time point of an observation window T_observation, and a reference timing t_ref4_i corresponds to a starting time point of a prediction window T_prediction. The terminal device 110 may use measurement results in a time window (t_ref3_i -T_observation, t_ref3_i) to predict a beam / CSI in a time window (t_ref4_i, t_ref4_i + T_prediction) .
[0168] In some embodiments, the reference timing t_ref3_i may be implemented as that for the reference timing t_ref2_i. For example, the reference timing t_ref3_i may be based on the configured time of the i-th report. In some embodiments, the reference timing t_ref4_i may also be implemented as that for the reference timing t_ref2_i. For example, the reference timing t_ref4_i may be based on the configured time of the latest RS before the configured time of the i-th report. Other details are not repeated here for conciseness.
[0169] As shown in FIG. 4B, a required separation between the reference timings t_ref3_i and t_ref4_i may be X4 time units. X4 may depend on required time for the first inference of the functionality.
[0170] Continuing to refer to FIG. 3, at step 320, the terminal device 110 may transmit the first information to the network device 120 from a first occasion (i.e., at and after the first occasion) . The first occasion is the earliest occasion after a first number (denoted as X herein) of time units since first reference time. The first number of time units are based on at least time required for collecting the historical information for the first inference. For example, the first occasion may be the first slot (after X time units since the first reference time) satisfying Accordingly, the network device 120 may receive the first information from the terminal device 110 from the first occasion.
[0171] In some embodiments, the first reference time may be based on a time in which the inference configuration is received or activated or triggered. For example, the first reference time may be a time of reception of a RRC reconfiguration message comprising the inference configuration (e.g., in the step 313) .
[0172] In some embodiments, the first reference time may be based on a time in which an indication of activating the inference configuration is received. For example, the first reference time may be a time of reception of a MAC activation command for activating the inference configuration (e.g., in the step 315) .
[0173] In some embodiments, the first reference time may be based on a time in which an indication of triggering a report based on the inference configuration is received. For example, the first reference time may be a time of reception of a PDCCH (e.g., with a DCI format containing a CSI request filed) for triggering a report based on the inference configuration.
[0174] In some embodiments, the first reference time may be based on a time in which a response to the reception of the inference configuration is transmitted. For example, the first reference time may be a slot that a RRC reconfiguration completion message has been transmitted in response to the RRC reconfiguration message comprising the inference configuration. In other words, the first reference time may be a slot that the inference configuration has been acknowledged.
[0175] In some embodiments, the first reference time may be based on a time in which an acknowledgement to the reception of the indication of activating the inference configuration is transmitted. For example, the first reference time may be a slot that a hybrid automatic repeat request acknowledgement (HARQ-ACK) to the MAC activation command is transmitted.
[0176] In some embodiments, the first reference time may be based on the last report of the first information.
[0177] It is to be noted that a slot is used as an example of a time unit for purpose of illustration, and any other suitable time units may also be feasible, such as a frame, subframe, symbol, or ms. The time unit is related to SCS.
[0178] In some embodiments, the first number of time units (i.e., X time units) may be larger than or equal to a time length of the observation window configured for the collection of the historical information. For example, X time units may be larger than or equal to Mt × T_RS, or T_observation. In another example, X time units may be larger than or equal to max (Mt × T_RS, T_observation) , or min (Mt × T_RS, T_observation) if all of Mt, T_RS, and T_observation are provided.
[0179] In some embodiments, the first number of time units may be larger than or equal to time (e.g., denoted as T_modelInference) used for the first inference.
[0180] In some embodiments, the first number of time units may be larger than or equal to time used for validation or monitoring of the first inference. In some embodiment, convergence time may be required to achieve stable prediction.
[0181] In some embodiments, the first number of time units may be larger than or equal to time (e.g., denoted as T_preparation) used for preparing a UL channel to carry the transmission of the first information (i.e., carry the report) .
[0182] In some embodiments, the first number of time units may be larger than or equal to time used for processing a signaling carrying the inference configuration. For example, the signaling may be a RRC signaling or a medium access control (MAC) control element (CE) .
[0183] In some embodiments, the first number of time units may be larger than or equal to time used for a model transfer. In some embodiments, the first number of time units may be larger than or equal to time used for reporting one or more applicable functionalities. In some embodiments, the first number of time units may be larger than or equal to time used for reporting one or more activated functionalities.
[0184] In some embodiments, the first number of time units may be the sum of one or more of the above time lengths. In some embodiments, the above time lengths may be configured by the network device 120. In some embodiments, the above time lengths may be predefined. In some embodiments, the above time lengths may be based on capability of the terminal device 110. In some embodiments, the above time lengths may be based on a model for the functionality.
[0185] In some embodiments, if the first information is transmitted, the terminal device 110 may consider that the functionality is activated or the activation of the functionality is successful. For example, after the first valid report for an inference of a functionality, the functionality is considered as activated or the activation of the functionality is considered as successful. Accordingly, if the first information is received, the network device 120 may consider that the functionality is activated or the activation of the functionality is successful.
[0186] In some embodiments, the terminal device 110 may report the first information which may be a prediction or inference result as configured, e.g., by the first report configuration. In some embodiments where the functionality is BM, the first information may comprise a set of predicted results for a first set of beams (i.e., Set A) in a set of future time instances. For example, the terminal device 110 may report RSRP associated with a SSBRI or CRI (e.g., denoted as ssbri / cri-rsrp) for the set of future time instances based on the Set A configuration. For example, a report format may include a field for SSBRI / CRI whose bit width is determined by ceil (log2 (N_resourcesInSetA) ) , where N_resourcesInSetA denotes number of resources for the Set A.
[0187] In some embodiments where the functionality is CSI prediction, the first information may comprise a set of predicted CSI for a set of future time instances. For example, the terminal device 110 may report predicted RI or PMI or CQI for the set of future time instances.
[0188] In some embodiments where the functionality is CSI compression, the first information may comprise compressed CSI (also referred to as first compressed CSI herein) with historical CSI information. In some embodiments, the first compressed CSI may be determined based on a first encoder or quantization, e.g., a first type of encoder or quantization. For example, the first encoder or quantization may use a time-spatial-frequency encoder, use a time-spatial-frequency quantizer, use a differential quantization codebook, use a vector (also referred to as a back state vector herein) based on historical CSI information, or any other similar ways. Accordingly, the network device 120 may need to do CSI reconstruction using historical CSI information. In some embodiments, the CSI reconstruction may be implemented by using a first decoder or inverse quantization, e.g., a first type of decoder or inverse quantization. For example, the first decoder or inverse quantization may use a time-spatial-frequency decoder, use a time-spatial-frequency inverse quantizer, use the differential quantization codebook, use the back state vector based on historical CSI information, or any other similar ways.
[0189] In some embodiments, the report of the first information may be based on only the most recent measurement. Alternatively, the report of the first information may be based on more than one most recent measurements.
[0190] As shown in step 330, the terminal device 110 may perform, before the first occasion, an operation (also referred to as a first operation herein) related to the reporting or transmission for the first inference. In some embodiments, as shown in step 331, the first operation may comprise that the terminal device 110 transmits second information related to the functionality to the network device 120 before the first occasion.
[0191] In other words, before the first occasion, the terminal device 110 may report the second information, instead of reporting the first information as configured. After the first occasion, the terminal device 110 may report the first information as configured. FIG. 4C illustrates a schematic diagram 400C illustrating an example reporting for model inference according to some embodiments of the present disclosure. As shown in FIG. 4C, the first reference time is a timing T2, and the first occasion is an occasion Z. A time interval between the occasion Z and the timing T2 is X time units. The terminal device 110 may report the second information at one or more report occasions before the occasion Z. After the occasion Z, prediction results are available. The terminal device 110 may report the first information at one or more report occasions after the occasion Z. In some embodiments, the terminal device 110 may report the first information at one or more report occasions from the occasion Z.
[0192] In some embodiments, the second information may be associated with a measurement related to the functionality. In some embodiments where the functionality is BM, the second information may comprise a set of measured results for a second set of beams (i.e., Set B) . For example, the terminal device 110 may report RSRP associated with a SSBRI or CRI based on one or more measurements for the Set B. For example, a report format may include a field for SSBRI / CRI whose bit width is determined by ceil (log2 (N_resourcesInSetB) ) , where N_resourcesInSetB denotes number of resources for the Set B.
[0193] In some embodiments where the functionality is CSI prediction, the second information may comprise a set of measured CSI. For example, a report quantity for the set of measured CSI may be measured RI or PMI or CQI.
[0194] In some embodiments, the second information may be associated with a second inference related to the functionality without using the historical information. In some embodiments where the functionality is CSI compression, the second information may comprise compressed CSI (also referred to as second compressed CSI herein) without the historical CSI information. For example, the second compressed CSI may be compressed present CSI without historical CSI information or compressed future CSI without historical CSI information.
[0195] In some embodiments, the second compressed CSI may be determined based on a second encoder or quantization (e.g., a second type of encoder or quantization) and an initial setting of a vector (e.g., the back state vector) . For example, the second encoder or quantization may not use the time-spatial-frequency encoder, not use the time-spatial-frequency quantizer, not use the differential quantization codebook, not use the vector based on historical CSI information, or any other similar ways. In other words, the second encoder or quantization may use a spatial-frequency encoder, use a spatial-frequency quantizer, or any other similar ways.
[0196] Accordingly, the network device 120 may need to do CSI reconstruction without using the historical CSI information. In some embodiments, the CSI reconstruction may be implemented by using a second decoder or inverse quantization, e.g., a second type of decoder or inverse quantization. For example, the second decoder or inverse quantization may not use the time-spatial-frequency decoder, not use the time-spatial-frequency inverse quantizer, not use the differential quantization codebook, not use the back state vector based on historical CSI information, or any other similar ways. In other words, the second decoder or inverse quantization may use a spatial-frequency decoder, use a spatial-frequency inverse quantizer, or any other similar ways.
[0197] In some embodiments, the initial setting of the vector may be done by assuming an initial state of the vector as a special vector, e.g., all 0 vector, all 1 vector, random vector, etc. Alternatively, the initial setting of the vector may be a given vector configured by the network device 120.
[0198] In some alternative or additional embodiments, if the functionality is CSI compression, the second information may comprise ground truth CSI. The ground truth CSI is channel state information without any AI / ML compression, e.g., raw channel matrix, or codebook based PMI, including type I codebook, type II codebook, or port selection codebook. It is useful for the network device 120 to use the ground truth CSI as the historical CSI information.
[0199] In some embodiments, the terminal device 110 may inform the network device 120 whether the historical CSI information is used for CSI compression or not. In some embodiments, the terminal device 110 may inform the network device 120 whether the reported compressed CSI is the compressed present CSI or the compressed future CSI.
[0200] In some embodiments, the report of the second information may be based on only the most recent measurement. Alternatively, the report of the second information may be based on more than one most recent measurements.
[0201] In some embodiments, the terminal device 110 may receive, from the network device 120, a configuration (also referred to as a second configuration herein) indicating the transmission of the second information related to the functionality. That is, the terminal device 110 may be provided with two configurations for report, i.e., the first configuration for reporting the first information and the second configuration for reporting the second information.
[0202] In some embodiments, the network device 120 may provide the second configuration and an association between the first and second configurations. In some embodiments, the second configuration may at least comprise a second report configuration and a second resource configuration.
[0203] In some embodiments, the second report configuration may comprise a second report quantity. In some embodiments, the second report quantity may comprise a measured result, e.g., measured beam or measured CSI. In some embodiments for BM-case 2, the second report quantity may be measured RSRP associated with a SSBRI, or measured RSRP associated with a CRI, or any other suitable quantities. In some embodiments for CSI prediction, the second report quantity may be measured RI, measured PMI, measured CQI, or any other suitable quantities. In some embodiments for CSI compression, the second report quantity may comprise an inference result, e.g., compressed future CSI without historical CSI information, compressed present CSI without historical CSI information, or ground truth CSI.
[0204] In some embodiments, the second report configuration may comprise time domain information for the report of the second information. In some embodiments, the time domain information may comprise a periodicity (e.g., denoted as P_report_2) and an offset (e.g., denoted as Offset_report_2) .
[0205] It is to be noted that the second report quantity and the time domain information may adopt any other suitable forms, and the second report configuration may comprise any other suitable information or information combinations.
[0206] In some embodiments, the second resource configuration may comprise one or more RS resource sets and RS resources for measurement, such as CSI-RS, SSB, etc. In some embodiments for BM-case 2, Set B is configured as one RS resource set, and Set A is not needed. In some embodiments for CSI prediction, one or more CSI-RS resources are configured.
[0207] In some embodiments, the second resource configuration may comprise time domain information of resources, such as, periodicity (e.g., denoted as P_RS_2) and offset (e.g., denoted as Offset_RS_2) . In some embodiments, P_report_2 = P_RS_2. In some embodiments, P_report_2 is an integer multiple of P_RS_2. In some embodiments, Offset_report_2 > Offset_RS_2.
[0208] In some embodiments, the first configuration is invalid and the second configuration is valid before the first occasion. The terminal device 110 may report the second information according to the second configuration before the first occasion. That is, before the first occasion, the first configuration may be considered as deactivated.
[0209] In some embodiments, the first configuration is valid and the second configuration is invalid after the first occasion. The terminal device 110 may report the first information according to the first configuration from the first occasion. That is, from the first occasion, the second configuration may be considered as deactivated.
[0210] In some embodiments, the network device 120 may provide two report quantities in the report configuration, and the two report quantities are configured for the first and second information respectively. In some embodiments, the first and second configurations may be only different in a report quantity and the rest information is the same. In some embodiments, the network device 120 may provide the second configuration in the step 313 and the first configuration in the step 315.
[0211] In some embodiments, as shown in step 332, the first operation may comprise that the terminal device 110 skips the transmission of the first information before the first occasion. That is, no report is required before the first occasion. In some embodiments, the terminal device 110 may transmit a CSI report only if receiving at least Mt consecutive CSI-RS transmission occasions for each CSI-RS resource in the corresponding CSI-RS resource set for channel measurement (e.g., Set B) , and drop the CSI report otherwise.
[0212] Accordingly, the network device 120 may perform, before the first occasion, an operation (also referred to as a second operation herein) related to reception for the first inference. In some embodiments, the second operation may comprise that the network device 120 receives the second information from the terminal device 110 before the first occasion. In some embodiments, the second operation may comprise that the network device 120 skips reception of the first information before the first occasion.
[0213] So far, a solution of a reporting for model inference is described in connection with the process 300. With the solution, an issue that UE cannot have prediction results immediately after an initial configuration or a reconfiguration of a model inference may be solved. It is to be noted that operations or steps described in the process 300 may be carried out separately or in any suitable combinations.
[0214] EXAMPLE IMPLEMENTATION OF FAST ACTIVATION OF MODEL INFERENCE
[0215] If an observation window required is large, or a prediction window is large, NW may not have useful information for scheduling after configuring a model inference, especially in the process 300 if UE does not provide any report.
[0216] Embodiments of the present disclosure provide a solution of fast activation of a model inference. The fast activation may be used to reduce a time length from the first reference time (e.g., a timing of reception of the inference configuration) to the first occasion (i.e., a timing of a reporting of the first valid inference result) to a value (e.g., Y) much smaller than X described in FIG. 4C. The solution will be described in connection with FIG. 5 below.
[0217] FIG. 5 illustrates a signaling chart illustrating another example process 500 of communication for model inference according to some embodiments of the present disclosure. For the purpose of discussion, the process 500 will be described with reference to FIG. 1. The process 500 may involve the terminal device 110 and the network device 120 as illustrated in FIG. 1. It is to be understood that the steps and the order of the steps in FIG. 3 are merely for illustration, and not for limitation. For example, the order of the steps may be changed. Some of the steps may be omitted or any other suitable additional steps may be added.
[0218] As shown in FIG. 5, at step 510, the terminal device 110 may receive, from the network device 120, a configuration (i.e., the first configuration) indicating a transmission or report of information (i.e., the first information) related to a functionality. The first information is associated with an inference (i.e., the first inference) related to the functionality using historical information.
[0219] In some embodiments where the functionality is BM, the first information may comprise a set of predicted results for a first set of beams (i.e., Set A) in a set of future time instances. In some embodiments where the functionality is CSI prediction, the first information may comprise a set of predicted CSI for a set of future time instances. In some embodiments where the functionality is CSI compression, the first information may comprise compressed CSI (i.e., the first compressed CSI) with historical CSI information.
[0220] It is to be noted that the step 510 is the same as the step 310, and thus other details are not repeated here for conciseness.
[0221] As shown in step 520, the terminal device 110 may also receive, from the network device 120, a configuration (also referred to as a third configuration herein) indicating an activation (or fast activation) of the first inference. In some embodiments, the terminal device 110 may transmit, to the network device 120, information (also referred to as capability information herein) indicating whether the terminal device 110 supports the activation of the first reference, e.g., via a UE capability information message or any other suitable messages. In some embodiments, if the information indicates that the terminal device 110 supports the activation of the first reference, the network device 120 may transmit the third configuration to the terminal device 110.
[0222] In some embodiments, the third configuration may comprise a dataset for the first inference. In some embodiments, the third configuration may comprise a burst of resources for data collection. In some embodiments, the burst of resources may comprise a set of persistent or semi-persistent resources. In some embodiments, the burst of resources may comprise a set of aperiodic resources.
[0223] As shown in step 530, the terminal device 110 may transmit the first information to the network device 120 based on the third configuration before a first occasion. Accordingly, the network device 120 may receive the first information before the first occasion. The first occasion is the earliest occasion after a first number of time units since a first reference time, and the first number of time units is based on at least time required for collecting the historical information for the first inference. Other details of the first occasion, the first number of time units and the first reference time are the same as that described in the process 300, and are not repeated here for conciseness.
[0224] In some embodiments, the functionality may be considered as activated upon reporting of the first valid inference result. The first valid inference result may be the first valid prediction or inference result obtained by using historical information.
[0225] As such, a fast activation for the first inference may be achieved. FIG. 6 illustrates a schematic diagram 600 illustrating another example reporting for model inference according to some embodiments of the present disclosure. As shown in FIG. 6, based on the burst of resources or dataset indicated in the third configuration, the terminal device 110 performs the first inference for the functionality, and a fast activation for the first inference is completed within Y time units since the reference timing T2. As such, the first information may be obtained before the occasion Z. That is, the first information may be obtained earlier. The fast activation may reduce a time length from the reference time T2 to the occasion Z from the X time units to the Y time units.
[0226] In some embodiments where the third configuration comprises the dataset for the first inference, before the first occasion, the terminal device 110 may derive the first information by the first inference based on the dataset. That is, the fast activation may be achieved by data (e.g., dataset or data samples) transfer. In this way, the terminal device 110 may be enabled to have model inference output before the first configured report occasion (i.e., the first occasion) .
[0227] In some embodiments, the dataset may comprise one or more previous results for the terminal device 110. In some embodiments, the dataset may comprise one or more measurement results from other terminal devices. In some embodiments, the dataset may comprise one or more measurement results from other terminal devices including one or more ground truth values. In some embodiments, the dataset may comprise one or more measurement results from the terminal device 110 including one or more logged results from one or more resources (e.g., SSB or other CSI-RS resources) .
[0228] In some embodiments, the network device 120 may transfer the dataset in the form of an exact format as model input required, e.g., ssbri / cri-rsrp for BM case, RI / PMI / CQI for CSI prediction case, etc. In some embodiments, the network device 120 may transfer the dataset in the form of statistic of a model input, e.g., distribution of the model input. In some embodiments, the network device 120 may transfer the dataset in the form of an indication to the terminal device 110 to enable use of previous results before the first configuration. In some embodiments, the network device 120 may transfer the dataset with an associated ID, a dataset ID, or a model ID.
[0229] In some embodiments where the third configuration comprises the burst of resources for data collection, before the first occasion, the terminal device 110 may perform a measurement based on the burst of resources and derive the first information by the first inference based on a result of the measurement. That is, the fast activation may be achieved by the burst of resources for measurements or for data collection. In this way, the terminal device 110 may also be enabled to have model inference output before the first configured report occasion (i.e., the first occasion) .
[0230] In some embodiments, the third configuration may comprise a report configuration (also referred to as a third report configuration herein) and resource configuration (also referred to as a third resource configuration herein) .
[0231] In some embodiments, the burst may be configured as persistent or semi-persistent resources, which has a smaller periodicity than the resources configured in the first configuration. In some embodiments, the configuration for the burst may be deactivated after Y time units upon the reception or ACK of the first configuration. In some embodiments, the configuration for the burst may be deactivated after the reporting of the first information according to the first configuration. In some embodiments, the configuration for the burst may be deactivated after the last RS before the reporting of the first information according to the first configuration. In some embodiments, the configuration for the burst may be deactivated after transmitting a message indicating the fast activation is completed.
[0232] In some embodiments, the burst may be configured as aperiodic resources. In some embodiments, the burst may be configured as repeated aperiodic resources. In some embodiments, the number of repetitions may be configured as K in a resource set for channel measurement. In some embodiments, the K resources may be triggered by the same triggering instance, and a separation between two consecutive resources may be m time units. In some embodiments, m may be configured by a higher layer parameter. In some embodiments, the K aperiodic resources may be transmitted following an order of resource IDs configured in the resource set. In some embodiments, the terminal device 110 may assume that an antenna port with the same port index of the K aperiodic resources is the same. In some embodiments, the terminal device 110 may report capability information comprising at least one of the following: the maximum value of K supported by the terminal device 110, or the minimum value of m supported by the terminal device 110.
[0233] In some embodiments, the burst may be configured with an aperiodic report. In some embodiments, the burst may be configured with no report. In some embodiments, the burst may be configured with the transmission of the message indicating that the fast activation is completed.
[0234] So far, a solution of fast activation for model inference is described in connection with the process 500. With the solution, a process for getting the first valid report for the model inference may be accelerated. It is to be understood that operations or steps described in the process 500 may be carried out separately or in any suitable combinations.
[0235] EXAMPLE IMPLEMENTATION OF METHODS
[0236] Corresponding to the above processes, embodiments of the present disclosure provide methods of communication implemented at terminal devices and network devices. These methods will be described below with reference to FIGs. 7 to 10.
[0237] FIG. 7 illustrates a flowchart illustrating an example method 700 of communication implemented at a terminal device in accordance with some embodiments of the present disclosure. For example, the method 700 may be performed at the terminal device 110 as shown in FIG. 1. For the purpose of discussion, in the following, the method 700 will be described with reference to FIG. 1. It is to be understood that the method 700 may include additional blocks not shown and / or may omit some blocks as shown, and the scope of the present disclosure is not limited in this regard.
[0238] At block 710, the terminal device 110 may receive, from the network device 120, a first configuration indicating a transmission of first information related to a functionality. The first information is associated with a first inference related to the functionality using historical information.
[0239] At block 720, the terminal device 110 may transmit the first information to the network device 120 from a first occasion that is an earliest occasion after a first number of time units since first reference time. The first number of time units is based on at least time required for collecting the historical information for the first inference.
[0240] In some embodiments, the first number of time units may be larger than or equal to a sum of one or more of the following: a time length of an observation window configured for the collection of the historical information; time used for the first inference; time used for validation or monitoring of the first inference; time used for preparing an uplink channel to carry the transmission of the first information; time used for processing a signaling carrying the first configuration; time used for a model transfer; or time used for reporting one or more applicable functionalities.
[0241] In some embodiments, the observation window may be associated with the following: number of historical time instances within the observation window, and a time interval between two neighbor historical time instances.
[0242] In some embodiments, an ending time point of the observation window may be the same as a starting time point of a prediction window configured for the first reference. In some embodiments, the ending time point of the observation window may be different from the starting time point of the prediction window, and a time interval between the ending time point of the observation window and the starting time point of the prediction window depends on at least time used for the first inference.
[0243] In some embodiments, the prediction window may be associated with the following: number of future time instances within the prediction window, and a time interval between two neighbor future time instances.
[0244] In some embodiments, the ending time point of the observation window or the starting time point of the prediction window may be based on one of the following: a configured time of the transmission of the first information associated with the prediction window or the observation window, or a configured time of a latest reference signal before the configured time of the transmission of the first information associated with the prediction window or the observation window.
[0245] In some embodiments, the first reference time may be based on one of the following: a time in which the first configuration is received; a time in which an indication of activating the first configuration is received; a time in which a response to the reception of the first configuration is transmitted; or a time in which an acknowledgement to the reception of the indication is transmitted.
[0246] In some embodiments, in accordance with a determination that the first configuration is received, the terminal device 110 may consider that the functionality is activated. In some embodiments, in accordance with a determination that the first information is transmitted, the terminal device 110 may consider that the activation of the functionality is successful.
[0247] At block 730, the terminal device 110 may perform, before the first occasion, a first operation comprising one of the following: transmitting second information related to the functionality to the network device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information; or skipping the transmission of the first information.
[0248] In some embodiments, the terminal device 110 may receive, from the network device 120, a second configuration indicating the transmission of the second information related to the functionality. In some embodiments, the first configuration is invalid and the second configuration is valid before the first occasion, and the first configuration is valid and the second configuration is invalid after the first occasion.
[0249] In some embodiments, the functionality may be beam management. In these embodiments, the first information may comprise a set of predicted results for a first set of beams in a set of future time instances, and the second information may comprise a set of measured results for a second set of beams.,
[0250] In some embodiments, the functionality may be CSI prediction. In these embodiments, the first information may comprise a set of predicted CSI for a set of future time instances, and the second information may comprise a set of measured CSI.
[0251] In some embodiments, the functionality may be CSI compression. In these embodiments, the first information may comprise first compressed CSI with historical CSI information, and the second information may comprise second compressed CSI without the historical CSI information. In some embodiments, the first compressed CSI may be determined based on a first encoder or quantization, and the second compressed CSI may be determined based on a second encoder or quantization and an initial setting of a vector.
[0252] With the method 700, an issue that UE cannot have prediction results immediately after an initial configuration or a reconfiguration of a model inference may be solved. It is to be understood that operations of the method 700 correspond to that described in connection with FIGs. 3 to 4C, and other details are omitted here for conciseness.
[0253] FIG. 8 illustrates a flowchart illustrating an example method 800 of communication implemented at a terminal device in accordance with some embodiments of the present disclosure. For example, the method 800 may be performed at the terminal device 110 as shown in FIG. 1. For the purpose of discussion, in the following, the method 800 will be described with reference to FIG. 1. It is to be understood that the method 800 may include additional blocks not shown and / or may omit some blocks as shown, and the scope of the present disclosure is not limited in this regard.
[0254] At block 810, the terminal device 110 may receive, from the network device 120, a first configuration indicating a transmission of first information related to a functionality. The first information is associated with a first inference related to the functionality using historical information.
[0255] At block 820, the terminal device 110 may determine that a third configuration indicating an activation of the first inference is received. In some embodiments, the terminal device 110 may transmit, to the network device 120, information indicating whether the terminal device supports the activation of the first reference.
[0256] In some embodiments, the third configuration may comprise a dataset for the first inference, or a burst of resources for data collection. In some embodiments, the burst of resources may comprise a set of persistent or semi-persistent resources, or a set of aperiodic resources.
[0257] At block 830, the terminal device 110 may transmit, to the network device 120, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time. The first number of time units is based on at least time required for collecting the historical information for the first inference.
[0258] In some embodiments, the functionality may be beam management, and the first information may comprise a set of predicted results for a first set of beams in a set of future time instances.
[0259] In some embodiments, the functionality may be CSI prediction, and the first information may comprise a set of predicted CSI for a set of future time instances.
[0260] In some embodiments, the functionality may be CSI compression, and the first information may comprise first compressed CSI with historical CSI information.
[0261] With the method 800, a process to get the first valid report for a model inference may be accelerated. It is to be understood that operations of the method 800 correspond to that described in connection with FIGs. 5 and 6, and other details are omitted here for conciseness.
[0262] FIG. 9 illustrates a flowchart illustrating an example method 900 of communication implemented at a network device in accordance with some embodiments of the present disclosure. For example, the method 900 may be performed at the network device 120 as shown in FIG. 1. For the purpose of discussion, in the following, the method 900 will be described with reference to FIG. 1. It is to be understood that the method 900 may include additional blocks not shown and / or may omit some blocks as shown, and the scope of the present disclosure is not limited in this regard.
[0263] At block 910, the network device 120 may transmit, to the terminal device 110, a first configuration indicating a transmission of first information related to a functionality. The first information is associated with a first inference related to the functionality using historical information.
[0264] At block 920, the network device 120 may receive the first information to the network device 120 from a first occasion that is an earliest occasion after a first number of time units since first reference time. The first number of time units is based on at least time required for collecting the historical information for the first inference.
[0265] In some embodiments, the first number of time units may be larger than or equal to a sum of one or more of the following: a time length of an observation window configured for the collection of the historical information; time used for the first inference; time used for validation or monitoring of the first inference; time used for preparing an uplink channel to carry the transmission of the first information; time used for processing a signaling carrying the first configuration; time used for a model transfer; or time used for reporting one or more applicable functionalities.
[0266] In some embodiments, the observation window may be associated with the following: number of historical time instances within the observation window, and a time interval between two neighbor historical time instances.
[0267] In some embodiments, an ending time point of the observation window may be the same as a starting time point of a prediction window configured for the first reference. In some embodiments, the ending time point of the observation window may be different from the starting time point of the prediction window, and a time interval between the ending time point of the observation window and the starting time point of the prediction window depends on at least time used for the first inference.
[0268] In some embodiments, the prediction window may be associated with the following: number of future time instances within the prediction window, and a time interval between two neighbor future time instances.
[0269] In some embodiments, the ending time point of the observation window or the starting time point of the prediction window may be based on one of the following: a configured time of the transmission of the first information associated with the prediction window or the observation window, or a configured time of a latest reference signal before the configured time of the transmission of the first information associated with the prediction window or the observation window.
[0270] In some embodiments, the first reference time may be based on one of the following: a time in which the first configuration is received; a time in which an indication of activating the first configuration is received; a time in which a response to the reception of the first configuration is transmitted; or a time in which an acknowledgement to the reception of the indication is transmitted.
[0271] In some embodiments, in accordance with a determination that the first configuration is transmitted, the network device 120 may consider that the functionality is activated. In some embodiments, in accordance with a determination that the first information is received, the network device 120 may consider that the activation of the functionality is successful.
[0272] At block 930, the network device 120 may perform, before the first occasion, a second operation comprising one of the following: receiving second information related to the functionality from the terminal device 110, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information; or skipping reception of the first information.
[0273] In some embodiments, the network device 120 may transmit, to the terminal device 110, a second configuration indicating the transmission of the second information related to the functionality. In some embodiments, the first configuration is invalid and the second configuration is valid before the first occasion, and the first configuration is valid and the second configuration is invalid after the first occasion.
[0274] In some embodiments, the functionality may be beam management. In these embodiments, the first information may comprise a set of predicted results for a first set of beams in a set of future time instances, and the second information may comprise a set of measured results for a second set of beams.,
[0275] In some embodiments, the functionality may be CSI prediction. In these embodiments, the first information may comprise a set of predicted CSI for a set of future time instances, and the second information may comprise a set of measured CSI.
[0276] In some embodiments, the functionality may be CSI compression. In these embodiments, the first information may comprise first compressed CSI with historical CSI information, and the second information may comprise second compressed CSI without the historical CSI information. In some embodiments, the first compressed CSI may be determined based on a first encoder or quantization, and the second compressed CSI may be determined based on a second encoder or quantization and an initial setting of a vector.
[0277] With the method 900, NW may get prediction results timely after an initial configuration or a reconfiguration of a model inference. It is to be understood that operations of the method 900 correspond to that described in connection with FIGs. 3 to 4C, and other details are omitted here for conciseness.
[0278] FIG. 10 illustrates a flowchart illustrating another example method 1000 of communication implemented at a network device in accordance with some embodiments of the present disclosure. For example, the method 1000 may be performed at the network device 120 as shown in FIG. 1. For the purpose of discussion, in the following, the method 1000 will be described with reference to FIG. 1. It is to be understood that the method 1000 may include additional blocks not shown and / or may omit some blocks as shown, and the scope of the present disclosure is not limited in this regard.
[0279] At block 1010, the network device 120 may transmit, to the terminal device 110, a first configuration indicating a transmission of first information related to a functionality. The first information is associated with a first inference related to the functionality using historical information.
[0280] At block 1020, the network device 120 may determine that a third configuration indicating an activation of the first inference is transmitted. In some embodiments, the network device 120 may receive, from the terminal device 110, information indicating whether the terminal device supports the activation of the first reference.
[0281] In some embodiments, the third configuration may comprise a dataset for the first inference, or a burst of resources for data collection. In some embodiments, the burst of resources may comprise a set of persistent or semi-persistent resources, or a set of aperiodic resources.
[0282] At block 1030, the network device 120 may receive, from the terminal device 110, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time. The first number of time units is based on at least time required for collecting the historical information for the first inference.
[0283] In some embodiments, the functionality may be beam management, and the first information may comprise a set of predicted results for a first set of beams in a set of future time instances.
[0284] In some embodiments, the functionality may be CSI prediction, and the first information may comprise a set of predicted CSI for a set of future time instances.
[0285] In some embodiments, the functionality may be CSI compression, and the first information may comprise first compressed CSI with historical CSI information.
[0286] With the method 1000, a process to get the first valid report for a model inference may be accelerated. It is to be understood that operations of the method 1000 correspond to that described in connection with FIGs. 5 and 6, and other details are omitted here for conciseness.
[0287] EXAMPLE IMPLEMENTATION OF DEVICES
[0288] FIG. 11 is a simplified block diagram of a device 1100 that is suitable for implementing embodiments of the present disclosure. The device 1100 can be considered as a further example implementation of the terminal device 110 or the network device 120 as shown in FIG. 1. Accordingly, the device 1100 can be implemented at or as at least a part of the terminal device 110 or the network device 120.
[0289] As shown, the device 1100 includes a processor 1110, a memory 1120 coupled to the processor 1110, a suitable transceiver 1140 coupled to the processor 1110, and a communication interface coupled to the transceiver 1140. The memory 1110 stores at least a part of a program 1130. The transceiver 1140 may be for bidirectional communications or a unidirectional communication based on requirements. The transceiver 1140 may include at least one of a transmitter 1142 or a receiver 1144. The transmitter 1142 and the receiver 1144 may be functional modules or physical entities. The transceiver 1140 has at least one antenna to facilitate communication, though in practice an Access Node mentioned in this application may have several ones. The communication interface may represent any interface that is necessary for communication with other network elements, such as X2 / Xn interface for bidirectional communications between eNBs / gNBs, S1 / NG interface for communication between a mobility management entity (MME) / access and mobility management function (AMF) / SGW / UPF and the eNB / gNB, Un interface for communication between the eNB / gNB and a relay node (RN) , or Uu interface for communication between the eNB / gNB and a terminal device.
[0290] The program 1130 is assumed to include program instructions that, when executed by the associated processor 1110, enable the device 1100 to operate in accordance with the embodiments of the present disclosure, as discussed herein with reference to FIGs. 1 to 10. The embodiments herein may be implemented by computer software executable by the processor 1110 of the device 1100, or by hardware, or by a combination of software and hardware. The processor 1110 may be configured to implement various embodiments of the present disclosure. Furthermore, a combination of the processor 1110 and memory 1120 may form processing means 1150 adapted to implement various embodiments of the present disclosure.
[0291] The memory 1120 may be of any type suitable to the local technical network and may be implemented using any suitable data storage technology, such as a non-transitory computer readable storage medium, semiconductor based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory, as non-limiting examples. While only one memory 1120 is shown in the device 1100, there may be several physically distinct memory modules in the device 1100. The processor 1110 may be of any type suitable to the local technical network, and may include one or more of general purpose computers, special purpose computers, microprocessors, digital signal processors (DSPs) and processors based on multicore processor architecture, as non-limiting examples. The device 1100 may have multiple processors, such as an application specific integrated circuit chip that is slaved in time to a clock which synchronizes the main processor.
[0292] In some embodiments, a terminal device comprises a circuitry configured to: receive, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; transmit the first information to the network device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; and perform, before the first occasion, a first operation comprising one of the following: transmitting second information related to the functionality to the network device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, or skipping the transmission of the first information.
[0293] In some embodiments, a terminal device comprises a circuitry configured to: receive, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; and in accordance with a determination that a third configuration indicating an activation of the first inference is received, transmit, to the network device, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.
[0294] In some embodiments, a network device comprises a circuitry configured to: transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; receive the first information from the terminal device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; and perform, before the first occasion, a second operation comprising one of the following: receiving second information related to the functionality from the terminal device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, or skipping reception of the first information.
[0295] In some embodiments, a network device comprises a circuitry configured to: transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; and in accordance with a determination that a third configuration indicating an activation of the first inference is transmitted, receive the first information from the terminal device based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.
[0296] The term ‘circuitry’ used herein may refer to hardware circuits and / or combinations of hardware circuits and software. For example, the circuitry may be a combination of analog and / or digital hardware circuits with software / firmware. As a further example, the circuitry may be any portions of hardware processors with software including digital signal processor (s) , software, and memory (ies) that work together to cause an apparatus, such as a terminal device or a network device, to perform various functions. In a still further example, the circuitry may be hardware circuits and or processors, such as a microprocessor or a portion of a microprocessor, that requires software / firmware for operation, but the software may not be present when it is not needed for operation. As used herein, the term circuitry also covers an implementation of merely a hardware circuit or processor (s) or a portion of a hardware circuit or processor (s) and its (or their) accompanying software and / or firmware.
[0297] Generally, various embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, logic or any combination thereof. Some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software which may be executed by a controller, microprocessor or other computing device. While various aspects of embodiments of the present disclosure are illustrated and described as block diagrams, flowcharts, or using some other pictorial representation, it will be appreciated that the blocks, apparatus, systems, techniques or methods described herein may be implemented in, as non-limiting examples, hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof.
[0298] The present disclosure also provides at least one computer program product tangibly stored on a non-transitory computer readable storage medium. The computer program product includes computer-executable instructions, such as those included in program modules, being executed in a device on a target real or virtual processor, to carry out the process or method as described above with reference to FIGs. 1 to 10. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, or the like that perform particular tasks or implement particular abstract data types. The functionality of the program modules may be combined or split between program modules as desired in various embodiments. Machine-executable instructions for program modules may be executed within a local or distributed device. In a distributed device, program modules may be located in both local and remote storage media.
[0299] Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may execute entirely on a machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
[0300] The above program code may be embodied on a machine readable medium, which may be any tangible medium that may contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine readable medium may be a machine readable signal medium or a machine readable storage medium. A machine readable medium may include but not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine readable storage medium would include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM) , a read-only memory (ROM) , an erasable programmable read-only memory (EPROM or Flash memory) , an optical fiber, a portable compact disc read-only memory (CD-ROM) , an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0301] Further, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are contained in the above discussions, these should not be construed as limitations on the scope of the present disclosure, but rather as descriptions of features that may be specific to particular embodiments. Certain features that are described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment may also be implemented in multiple embodiments separately or in any suitable sub-combination.
[0302] Although the present disclosure has been described in language specific to structural features and / or methodological acts, it is to be understood that the present disclosure defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.
Claims
1.A terminal device, comprising:a processor configured to cause the terminal device to:receive, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information;transmit the first information to the network device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; andperform, before the first occasion, a first operation comprising one of the following:transmitting second information related to the functionality to the network device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, orskipping the transmission of the first information.2.The terminal device of claim 1, wherein the first number of time units are larger than or equal to a sum of one or more of the following:a time length of an observation window configured for the collection of the historical information;time used for the first inference;time used for validation or monitoring of the first inference;time used for preparing an uplink channel to carry the transmission of the first information;time used for processing a signaling carrying the first configuration;time used for a model transfer; ortime used for reporting one or more applicable functionalities.3.The terminal device of claim 2, wherein the observation window is associated with the following:number of historical time instances within the observation window, anda time interval between two neighbor historical time instances.4.The terminal device of claim 2, wherein an ending time point of the observation window is the same as a starting time point of a prediction window configured for the first reference; orwherein the ending time point of the observation window is different from the starting time point of the prediction window, and a time interval between the ending time point of the observation window and the starting time point of the prediction window depends on at least time used for the first inference.5.The terminal device of claim 4, wherein the prediction window is associated with the following:number of future time instances within the prediction window, anda time interval between two neighbor future time instances.6.The terminal device of claim 4, wherein the ending time point of the observation window or the starting time point of the prediction window is based on one of the following:a configured time of the transmission of the first information associated with the prediction window or the observation window, ora configured time of a latest reference signal before the configured time of the transmission of the first information associated with the prediction window or the observation window.7.The terminal device of claim 1, wherein the first reference time is based on one of the following:a time in which the first configuration is received;a time in which an indication of activating the first configuration is received;a time in which a response to the reception of the first configuration is transmitted; ora time in which an acknowledgement to the reception of the indication is transmitted.8.The terminal device of claim 1, wherein the terminal device is further caused to:in accordance with a determination that the first configuration is received, consider that the functionality is activated; orin accordance with a determination that the first information is transmitted, consider that the activation of the functionality is successful.9.The terminal device of claim 1, wherein the terminal device is further caused to:receive, from the network device, a second configuration indicating the transmission of the second information related to the functionality.10.The terminal device of claim 9, wherein the first configuration is invalid and the second configuration is valid before the first occasion, and the first configuration is valid and the second configuration is invalid after the first occasion.11.The terminal device of claim 1, wherein the functionality is beam management, andwherein the first information comprises a set of predicted results for a first set of beams in a set of future time instances, and the second information comprises a set of measured results for a second set of beams.12.The terminal device of claim 1, wherein the functionality is channel status information (CSI) prediction, andwherein the first information comprises a set of predicted CSI for a set of future time instances, and the second information comprises a set of measured CSI.13.The terminal device of claim 1, wherein the functionality is channel status information (CSI) compression,wherein the first information comprises first compressed CSI with historical CSI information, and the second information comprises second compressed CSI without the historical CSI information, andwherein the first compressed CSI is determined based on a first encoder or quantization, and the second compressed CSI is determined based on a second encoder or quantization and an initial setting of a vector.14.A terminal device, comprising:a processor configured to cause the terminal device to:receive, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; andin accordance with a determination that a third configuration indicating an activation of the first inference is received, transmit, to the network device, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.15.The terminal device of claim 14, wherein the third configuration comprises a dataset for the first inference, or a burst of resources for data collection.16.The terminal device of claim 15, wherein the burst of resources comprises a set of persistent or semi-persistent resources, or a set of aperiodic resources.17.The terminal device of claim 14, wherein the terminal device is further caused to:transmit, to the network device, information indicating whether the terminal device supports the activation of the first reference.18.The terminal device of claim 14, wherein the functionality is beam management, and the first information comprises a set of predicted results for a first set of beams in a set of future time instances; orwherein the functionality is channel status information (CSI) prediction, and the first information comprises a set of predicted CSI for a set of future time instances; orwherein the functionality is CSI compression, and the first information comprises first compressed CSI with historical CSI information.19.A network device, comprising:a processor configured to cause the network device to:transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information;receive the first information from the terminal device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; andperform, before the first occasion, a second operation comprising one of the following:receiving second information related to the functionality from the terminal device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, orskipping reception of the first information.20.The network device of claim 19, wherein the first number of time units are larger than or equal to a sum of one or more of the following:a time length of an observation window configured for the collection of the historical information;time used for the first inference;time used for validation or monitoring of the first inference;time used for preparing an uplink channel to carry the transmission of the first information;time used for processing a signaling carrying the first configuration;time used for a model transfer; ortime used for reporting one or more applicable functionalities.21.The network device of claim 20, wherein the observation window is associated with the following:number of historical time instances within the observation window, anda time interval between two neighbor historical time instances.22.The network device of claim 20, wherein an ending time point of the observation window is the same as a starting time point of a prediction window configured for the first reference; orwherein the ending time point of the observation window is different from the starting time point of the prediction window, and a time interval between the ending time point of the observation window and the starting time point of the prediction window depends on at least time used for the first inference.23.The network device of claim 22, wherein the prediction window is associated with the following:number of future time instances within the prediction window, anda time interval between two neighbor future time instances.24.The network device of claim 22, wherein the ending time point of the observation window or the starting time point of the prediction window is based on one of the following:a configured time of the transmission of the first information associated with the prediction window or the observation window, ora configured time of a latest reference signal before the configured time of the transmission of the first information associated with the prediction window or the observation window.25.The network device of claim 19, wherein the first reference time is based on one of the following:a time in which the first configuration is transmitted;a time in which an indication of activating the first configuration is transmitted;a time in which a response to the reception of the first configuration is received;a time in which an acknowledgement to the reception of the indication is received.26.The network device of claim 19, wherein the network device is further caused to:in accordance with a determination that the first configuration is transmitted, consider that the functionality is activated; orin accordance with a determination that the first information is received, consider that the activation of the functionality is successful.27.The network device of claim 19, wherein the network device is further caused to:transmit, to the terminal device, a second configuration indicating the transmission of the second information related to the functionality.28.The network device of claim 27, wherein the first configuration is invalid and the second configuration is valid before the first occasion, and the first configuration is valid and the second configuration is invalid after the first occasion.29.The network device of claim 19, wherein the functionality is beam management, andwherein the first information comprises a set of predicted results for a first set of beams in a set of future time instances, and the second information comprises a set of measured results for a second set of beams.30.The network device of claim 19, wherein the functionality is channel status information (CSI) prediction, andwherein the first information comprises a set of predicted CSI for a set of future time instances, and the second information comprises a set of measured CSI.31.The network device of claim 19, wherein the functionality is channel status information (CSI) compression,wherein the first information comprises first compressed CSI with historical CSI information, and the second information comprises second compressed CSI without the historical CSI information, andwherein the first compressed CSI is determined based on a first encoder or quantization, and the second compressed CSI is determined based on a second encoder or quantization and an initial setting of a vector.32.A network device, comprising:a processor configured to cause the network device to:transmit, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; andin accordance with a determination that a third configuration indicating an activation of the first inference is transmitted, receive the first information from the terminal device based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.33.The network device of claim 32, wherein the third configuration comprises a dataset for the first inference, or a burst of resources for data collection.34.The network device of claim 33, wherein the burst of resources comprises a set of persistent or semi-persistent resources, or a set of aperiodic resources.35.The network device of claim 32, wherein the network device is further caused to:receive, from the terminal device, information indicating whether the terminal device supports the activation of the first reference.36.The network device of claim 32, wherein the functionality is beam management, and the first information comprises a set of predicted results for a first set of beams in a set of future time instances; orwherein the functionality is channel status information (CSI) prediction, and the first information comprises a set of predicted CSI for a set of future time instances; orwherein the functionality is CSI compression, and the first information comprises first compressed CSI with historical CSI information.37.A method of communication at a terminal device, comprising:receiving, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information;transmitting the first information to the network device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; andperforming, before the first occasion, a first operation comprising one of the following:transmitting second information related to the functionality to the network device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, orskipping the transmission of the first information.38.A method of communication at a terminal device, comprising:receiving, from a network device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; andin accordance with a determination that a third configuration indicating an activation of the first inference is received, transmitting the first information to the network device based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.39.A method of communication at a network device, comprising:transmitting, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information;receiving the first information from the terminal device from a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference; andperforming, before the first occasion, a second operation comprising one of the following:receiving second information related to the functionality from the terminal device, the second information being associated with a measurement related to the functionality or a second inference related to the functionality without using the historical information, orskipping reception of the first information.40.A method of communication at a network device, comprising:transmitting, to a terminal device, a first configuration indicating a transmission of first information related to a functionality, the first information being associated with a first inference related to the functionality using historical information; andin accordance with a determination that a third configuration indicating an activation of the first inference is transmitted, receiving, from the terminal device, the first information based on the third configuration before a first occasion that is an earliest occasion after a first number of time units since first reference time, the first number of time units being based on at least time required for collecting the historical information for the first inference.
Citation Information
Patent Citations
Prediction method of operational performance based on historical data modeling in grid
CN101697141A
Network mobility management optimization method, base station, device, system and related equipment
CN117560650A
Equipment energy consumption prediction method and system based on graph neural network
CN117575072A
Intelligent prediction method and device for AI computing power network load, and related medium and equipment
CN118802587A
Communication method, user equipment, network entity, and storage medium
WO2023219402A1