Capture additional conditions

Abstract

The present disclosure relates to capturing additional conditions. The method comprises: at a first apparatus, receiving, from a second apparatus, model feature information, the model feature information comprising a plurality of first training configurations identified by respective first configuration identifiers, wherein the plurality of first training configurations are associated with a communication condition; determining, for each first training configuration, a plurality of second training configurations identified by respective second configuration identifiers; obtaining, for each second training configuration, a plurality of data sets associated with the communication condition based on the plurality of first training configurations, wherein one data set is obtained based on one first training configuration from the plurality of first training configurations; and transmitting, to the second apparatus, data collection information, the data collection information comprising the plurality of second training configurations identified by respective second configuration identifiers.

Description

[0001] field This application claims priority and benefit to GB application number 2402173.5, filed on February 16, 2024, the contents of which are incorporated herein by reference in their entirety. Technical Field

[0002] Various exemplary embodiments of this disclosure generally relate to the telecommunications field, and more particularly to methods, apparatuses, devices, and computer-readable storage media for capturing additional conditions. Background Technology

[0003] With the integration of artificial intelligence (AI) and machine learning (ML) into the air interfaces of 5G and emerging 6G New Radio (NR), new frontiers in network adaptability and efficiency are being explored. 3GPP Release-18 research emphasizes the importance of model generalization across various network scenarios to address the need for AI / ML models to maintain robust performance under diverse conditions. This includes strategically incorporating additional conditions to refine model training, ensuring that the model adapts well to the requirements of both the network and user equipment sides. Therefore, capturing these additional conditions warrants in-depth research, as understanding and utilizing these factors is crucial for AI / ML applications in 5G and its subsequent evolution, and can significantly impact network performance and user experience. Summary of the Invention

[0004] In a first aspect of this disclosure, a first apparatus is provided. The first apparatus includes: at least one processor; and at least one memory storing instructions, which, when executed by the at least one processor, cause the first apparatus to: receive model feature information from a second apparatus, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; for each first training configuration, determine a plurality of second training configurations identified by corresponding second configuration identifiers; for each second training configuration, acquire a plurality of datasets associated with the communication conditions based on the plurality of first training configurations, wherein one dataset is acquired based on a first training configuration from the plurality of first training configurations; and transmit data collection information to the second apparatus, the data collection information including the plurality of second training configurations identified by corresponding second configuration identifiers.

[0005] In a second aspect of this disclosure, a second apparatus is provided. The second apparatus includes: at least one processor; and at least one memory storing instructions, which, when executed by the at least one processor, cause the second apparatus to: transmit model feature information to a first apparatus, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; for each first training configuration, receive data collection information from the first apparatus, the data collection information including a plurality of second training configurations identified by corresponding second configuration identifiers; and for each second training configuration, acquire a plurality of datasets associated with the communication conditions based on the plurality of second training configurations, wherein one dataset is acquired based on a second training configuration from the plurality of second training configurations.

[0006] In a third aspect of this disclosure, a method is provided. The method includes: receiving model feature information from a second device, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; for each first training configuration, determining a plurality of second training configurations identified by corresponding second configuration identifiers; for each second training configuration, acquiring a plurality of datasets associated with the communication conditions based on the plurality of first training configurations, wherein one dataset is acquired based on a first training configuration from the plurality of first training configurations; and transmitting data collection information to the second device, the data collection information including the plurality of second training configurations identified by corresponding second configuration identifiers.

[0007] In a fourth aspect of this disclosure, a method is provided. The method includes: transmitting model feature information to a first device, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; receiving data collection information from the first device for each first training configuration, the data collection information including a plurality of second training configurations identified by corresponding second configuration identifiers; and for each second training configuration, acquiring a plurality of datasets associated with the communication conditions based on the plurality of second training configurations, wherein one dataset is acquired based on a second training configuration from the plurality of second training configurations.

[0008] In a fifth aspect of this disclosure, a first apparatus is provided. The first apparatus includes: means for receiving model feature information from a second apparatus, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; means for determining, for each first training configuration, a plurality of second training configurations identified by corresponding second configuration identifiers; means for acquiring, for each second training configuration, a plurality of datasets associated with the communication conditions based on the plurality of first training configurations, wherein one dataset is acquired based on a first training configuration from the plurality of first training configurations; and means for transmitting data collection information to the second apparatus, the data collection information including the plurality of second training configurations identified by corresponding second configuration identifiers.

[0009] In a sixth aspect of this disclosure, a second apparatus is provided. The second apparatus includes: means for transmitting model feature information to a first apparatus, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; means for receiving data collection information from the first apparatus for each first training configuration, the data collection information including a plurality of second training configurations identified by corresponding second configuration identifiers; and means for acquiring, for each second training configuration, a plurality of datasets associated with the communication conditions based on the plurality of second training configurations, wherein one dataset is acquired based on a second training configuration from the plurality of second training configurations.

[0010] In a seventh aspect of this disclosure, a computer-readable medium is provided. The computer-readable medium includes instructions stored thereon for causing a device to perform at least the method according to a third aspect.

[0011] In an eighth aspect of this disclosure, a computer-readable medium is provided. The computer-readable medium includes instructions stored thereon for causing a device to perform at least the method according to the fourth aspect.

[0012] It should be understood that the summary portion is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0013] Some exemplary embodiments will now be described with reference to the accompanying drawings, in which: Figure 1 An example communication environment in which example embodiments of this disclosure may be implemented is shown; Figure 2A and Figure 2B The signaling flow of additional conditional groupings according to some example embodiments of this disclosure is shown; Figure 3 The signaling flow of user equipment side and additional condition packets according to some example embodiments of this disclosure is shown; Figure 4 The signaling flow shown illustrates model training and identification using information from network and user equipment-side additional conditional packets according to some example embodiments of this disclosure; Figure 5 A flowchart is shown illustrating a method implemented at a first device according to some example embodiments of the present disclosure; Figure 6 A flowchart illustrating a method implemented at a second device according to some example embodiments of the present disclosure is shown; Figure 7 A simplified block diagram of a device suitable for implementing example embodiments of the present disclosure is shown; and Throughout all the accompanying figures, the same or similar reference numerals denote the same or similar elements. Detailed Implementation

[0014] The principles of this disclosure will now be described with reference to some exemplary embodiments. It should be understood that these embodiments are described for illustrative purposes only and to assist those skilled in the art in understanding and implementing this disclosure, and do not imply any limitation on the scope of this disclosure. The embodiments described herein can be implemented in various ways other than those described below.

[0015] In the following description and claims, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

[0016] References to "an embodiment," "embodiment," "example embodiment," etc., in this disclosure indicate that the described embodiments may include specific features, structures, or characteristics, but it is not necessary for every embodiment to include that specific feature, structure, or characteristic. Furthermore, these phrases do not necessarily refer to the same embodiment. Further, when a specific feature, structure, or characteristic is described in connection with an embodiment, it is to be noted that those skilled in the art will recognize, whether explicitly described or not, that such features, structures, or characteristics apply in conjunction with other embodiments.

[0017] It should be understood that although terms such as "first," "second," etc., preceding nouns in this document may be used to describe various elements, these elements should not be limited to these terms. These terms are only used to distinguish one element from another and they do not restrict the order of nouns. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. As used herein, the term "and / or" includes any or all combinations of one or more of the listed terms.

[0018] As used herein, “at least one of the following: ” and “at least one of ” and similar expressions, wherein the list of two or more elements is connected by “and” or “or”, means at least any one of these elements, or at least any two or more of these elements, or at least all of these elements.

[0019] As used herein, unless explicitly stated otherwise, the “responding to A” action does not indicate that the action is performed immediately after “A” occurs, and may include one or more intermediate steps.

[0020] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments. As used herein, the singular forms “a,” “an,” and “the” are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising,” “including,” “having,” “possessing,” “containing,” and / or “covering,” as used herein, specify the presence of the stated features, elements, and / or components, but do not exclude the presence or addition of one or more other features, elements, components, and / or combinations thereof.

[0021] As used in this application, the term "circuit" may refer to one or more or all of the following: (a) Hardware circuit implementation only (such as implementation in analog and / or digital circuits only) and (b) A combination of hardware circuitry and software, such as (if applicable): (i) A combination of analog and / or digital hardware circuitry with software / firmware, and (ii) Any part of a hardware processor (including (multiple) digital signal processors), software, and memory that work together to enable a device (such as a mobile phone or server) to perform various functions; and (c) (Multiple) hardware circuits and / or (multiple) processors, such as (multiple) microprocessors or portions of (multiple) microprocessors, which require software (e.g., firmware) to operate, but may not exist when the software is not required to operate.

[0022] This definition of "circuit" applies to all uses of the term in this application (including in any claim). As another example, as used herein, the term "circuit" also encompasses only hardware circuitry or a processor (or multiple processors) or portions thereof, and their accompanying software and / or firmware implementations. The term "circuit" also encompasses, for example and if applicable to a particular claim element, baseband integrated circuits or processor integrated circuits for mobile devices, or similar integrated circuits in servers, cellular network devices, or other computing or network devices.

[0023] As used herein, the term "communication network" refers to a network that conforms to any suitable communication standard, such as New Radio (NR), LTE, LTE-A Advanced, WCDMA, HSPA, NB-IoT, WLAN, WiFi, etc. Furthermore, communication between terminal devices and network devices in a communication network can be performed according to any suitable generation of communication protocol, including but not limited to first-generation (1G), second-generation (2G), 2.5G, 2.75G, third-generation (3G), fourth-generation (4G), 4.5G, fifth-generation (5G), sixth-generation (6G) communication protocols, and / or any other currently known or to be developed in the future. Embodiments of this disclosure can be applied to various communication systems. Given the rapid development of communications, there will naturally be future types of communication technologies and systems that can implement this disclosure. The scope of this disclosure should not be considered limited to the systems described above.

[0024] As used herein, the term "network device" refers to a node in a communication network through which terminal devices access the network and receive services. Depending on the terminology and technology applied, a network device can refer to a base station (BS) or access point (AP), such as a Node B (or NB), an evolved Node B (eNodeB or eNB), an NR NB (also known as a gNB), a Remote Radio Unit (RRU), a Radio Head (RH), a Remote Radio Head (RRH), an Operation and Maintenance Entity (OAM), a Positioning Reference Unit (PRU), a Location Management Function (LMF), a relay, an Integrated Access and Backhaul (IAB) node, a low-power node (such as a femto or pico), a non-terrestrial network (NTN) or non-terrestrial network device (such as satellite network devices, low Earth orbit (LEO) satellites, and geostationary Earth orbit (GEO) satellites), an aircraft network device, etc. In some example embodiments, the Radio Access Network (RAN) split architecture includes a central unit (CU) and a distributed unit (DU) at the IAB host node. An IAB node includes a mobile terminal (IAB-MT) portion similar to the UE facing the parent node, and the DU portion of the IAB node is similar to the base station facing the next-hop IAB node.

[0025] The term "terminal device" refers to any end device with wireless communication capabilities. As an example and not a limitation, a terminal device can refer to communication equipment, user equipment (UE), subscriber station (SS), portable subscriber station, mobile station (MS), or access terminal (AT). Terminal devices can include, but are not limited to, mobile phones, cellular phones, smartphones, VoIP phones, wireless local loop phones, tablets, wearable terminal devices, personal digital assistants (PDAs), portable computers, desktop computers, image acquisition terminal devices (such as digital cameras), gaming terminal devices, music storage and playback devices, in-vehicle wireless terminal devices, wireless endpoints, mobile stations, laptop embedded devices (LEE), laptop mounted devices (LME), USB dongles, smart devices, wireless client devices (CPE), Internet of Things (IoT) devices, watches or other wearable devices, head-mounted displays (HMDs), vehicles, drones, medical devices and applications (e.g., remote surgery), industrial devices and applications (e.g., robots and / or other wireless devices operating in the context of industrial and / or automated processing chains), consumer electronics devices, devices operating on commercial and / or industrial wireless networks, etc. The terminal device may also correspond to the mobile terminal MT portion of an IAB node (e.g., a relay node). In the following description, the terms "terminal device," "communication device," "terminal," "user equipment," and "UE" are used interchangeably.

[0026] As used herein, the terms “resource,” “transmission resource,” “resource block,” “physical resource block” (PRB), “uplink resource,” or “downlink resource” can refer to any resource used to perform communication, such as communication between a terminal device and a network device, including resources in the time domain, frequency domain, spatial domain, code domain, or any other combination of time-domain, frequency-domain, spatial-domain, and / or code-domain resources that enable communication. In the following, unless explicitly stated otherwise, resources in the frequency and time domains will be used as examples of transmission resources used to describe some exemplary embodiments of this disclosure. Note that the exemplary embodiments of this disclosure are equally applicable to other resources in other domains.

[0027] As used herein, the term "model feature information" can refer to specific characteristics or properties in a computational model, particularly in the fields of machine learning and statistical analysis. These features represent key factors in model analysis for making predictions, performing classifications, or performing other data-driven tasks. Note that the exemplary embodiments of this disclosure are equally applicable to model feature information in other fields.

[0028] As used herein, the term "communication conditions" can refer to the current state or quality of a communication link or network, including various factors such as location, signal strength, noise level, bandwidth availability, and interference. Note that the exemplary embodiments of this disclosure are equally applicable to communication conditions in other domains.

[0029] As used herein, the term "training configuration" can refer to the system setup and parameters defined for collecting and preparing data specifically for developing and refining machine learning algorithms within a communications sector. A training configuration may also include machine learning (ML)-specific training parameters, such as ML model hyperparameters, ML input data preprocessing, or ML output data postprocessing configurations. ML-specific training parameters are crucial for accurate predictions. For example, when optimizing network performance through predictive analytics, a data collection configuration might involve the strategic capture of network load patterns, error rates, and quality of service indicators at different times and locations to train a model that can predict and mitigate potential network problems. Note that the exemplary embodiments of this disclosure are equally applicable to data collection configurations in other domains.

[0030] As used herein, the term "configuration identifier" can refer to a unique label or code that distinguishes a specific set of parameters or settings used during data collection, preprocessing, and model training processes. This identifier enables the tracking, replication, and comparison of different machine learning experiments or training sessions. For example, when training a model to predict network congestion, a configuration identifier can be used to represent a specific combination of features extracted from network traffic data. Note that the exemplary embodiments of this disclosure are equally applicable to configuration identifiers in other domains.

[0031] As used herein, the term "buffer" can refer to a temporary storage area, typically in memory, used to hold data as it is transferred from one place to another, thereby ensuring smooth and efficient data processing. Buffers can be used to manage data packet flows across a network, adapting to changes in data transmission rates and preventing packet loss during high-traffic conditions. Note that the exemplary embodiments of this disclosure are equally applicable to buffering in other domains.

[0032] As used herein, the term "feature difference" refers to variations or differences in the properties, characteristics, or behaviors of data configurations or datasets. For example, in the field of machine learning, identifying feature differences between data points is essential for training algorithms to accurately distinguish categories or predict outcomes. Note that the exemplary embodiments of this disclosure are equally applicable to identifying feature differences in other domains.

[0033] As used herein, the term "core network equipment" refers to the essential hardware components within a telecommunications network that provide critical functions for data routing, management, and connectivity across the network. These devices facilitate centralized network operation and support data transmission between different parts of the network and to external networks. Note that the exemplary embodiments of this disclosure are equally applicable to core network equipment in other domains.

[0034] As used herein, the term "radio access network equipment" refers to components and devices within a telecommunications system that connect mobile devices to the core network and facilitate wireless communication. These devices are integral to establishing and maintaining radio links between users' mobile devices and the network, handling tasks such as signal transmission, reception, and modulation. Note that the exemplary embodiments of this disclosure are equally applicable to radio access network equipment in other domains.

[0035] As used herein, a machine learning (ML) entity can contain an ML model and related metadata. An ML entity can be managed as a single composite entity. In some example implementations, an ML entity can be implemented as an ML application (MLApp).

[0036] To facilitate understanding of the terminology, the RAN1 convention for the list of terms used in AI / ML is provided below.

[0037] AI / ML Models: Data-driven algorithms that apply AI / ML techniques to generate a set of outputs based on a set of inputs.

[0038] AI / ML Model Delivery: The general term refers to delivering an AI / ML model from one entity to another in any way. Note: Entities can refer to network nodes / functions (e.g., gNB, LMF, etc.), UEs, dedicated servers, etc. AI / ML model inference: The process of using a trained AI / ML model to produce a set of outputs based on a set of inputs.

[0039] AI / ML Model Testing: A sub-process of training used to evaluate the performance of the final AI / ML model using a different dataset than that used for model training and validation. Unlike AI / ML model validation, testing does not assume subsequent adjustments to the model.

[0040] AI / ML model training: The process of training an AI / ML model in a data-driven manner [by learning input / output relationships] and obtaining the trained AI / ML model for inference.

[0041] AI / ML model transmission: AI / ML models are delivered over the air interface in a manner opaque to 3GPP signaling. These models may contain parameters of a model structure known to the receiving end, or they may be new models with parameters. The transmission may include a complete model or a partial model.

[0042] AI / ML Model Validation: A sub-process of training that evaluates the quality of an AI / ML model using a dataset different from the dataset used for model training, and helps in selecting model parameters that generalize to datasets other than those used for model training.

[0043] Data collection: The process by which network nodes, management entities, or UEs collect data for use in AI / ML model training, data analysis, and inference.

[0044] Federated learning / federated training: A machine learning technique that trains AI / ML models across multiple distributed edge nodes (e.g., UE, gNB), with each node performing local model training using local data samples. This technique requires multiple interactions between models but does not exchange local data samples.

[0045] Function Identification: Identifies the process / method for mutual understanding between the NW and UE regarding AI / ML functions. Note: Information about AI / ML functions can be shared during function identification. The residency of AI / ML functions depends on specific use cases and sub-use cases.

[0046] Model Activation: Enables AI / ML models for specific functions.

[0047] Model deactivation: Disables AI / ML models for specific functions.

[0048] Model download: The transfer of the model from the network to the UE.

[0049] Model Identification: The process / method for identifying the AI / ML model used for mutual understanding between the NW and UE. Note: The process / method for model identification may or may not be applicable. Note: Information about the AI / ML model may be shared during model identification.

[0050] Model monitoring: The process of monitoring the inference performance of AI / ML models.

[0051] Model parameter update: The process of updating the model parameters.

[0052] Model selection: The process of selecting the AI / ML model to be activated from among multiple models that share the same AI / ML-supporting features. Note: Model selection may or may not be performed concurrently with model activation.

[0053] Model switching: Deactivate the currently active AI / ML model and activate different AI / ML models for specific functions.

[0054] Model update: The process of updating the model parameters and / or model structure.

[0055] Model upload: The transfer of the model from the UE to the network.

[0056] Network-side (AI / ML) model: An AI / ML model in which inference is performed entirely at the network.

[0057] Offline field data: Data collected from the field and used for offline training of AI / ML models.

[0058] Offline training: An AI / ML training process in which a model is trained based on a collected dataset, and the trained model is later used or delivered for inference. Note: This definition is for guidance only. There may be cases that do not perfectly fit this definition but can still be classified as offline training through generally accepted conventions.

[0059] Online field data: Data collected from the field and used for online training of AI / ML models.

[0060] Online training: The AI / ML training process in which the model used for inference is trained (usually continuously) (nearly) real-time as new training samples arrive. Note: The concepts of (near) real-time and non-real-time are context-dependent and relative to the inference timescale. Note: This definition is for guidance only. There may be cases that do not perfectly fit this definition but can still be classified as online training by generally accepted conventions. Note: Fine-tuning / retraining can be done via online or offline training. (This note can be removed when we define the term fine-tuning.) Reinforcement learning (RL): The process of training an AI / ML model based on inputs (also called states) and feedback signals (also called rewards) generated by the model's outputs (also called actions) in an environment where the model interacts with itself.

[0061] Semi-supervised learning: The process of training a model using a mixture of labeled and unlabeled data.

[0062] Supervised learning: The process of training a model based on inputs and their corresponding labels.

[0063] Two-sided (AI / ML) model: A pair of AI / ML models that perform joint inference, where joint inference includes AI / ML inference jointly performed across the UE and the network, i.e., the first part of the inference is first performed by the UE, and then the remaining part is performed by the gNB, and vice versa.

[0064] UE-side (AI / ML) model: An AI / ML model where inference is performed entirely at the UE.

[0065] Unsupervised learning: The process of training a model without labeled data.

[0066] Proprietary format models: From a 3GPP perspective, these are ML models using vendor / device-specific proprietary formats. Different vendors cannot recognize these models, and sharing them hides model design information from other vendors. Note: An example is a device-specific binary executable format.

[0067] Open format models: From a 3GPP perspective, these are specified format ML models that are mutually identifiable across suppliers and allow interoperability. They are mutually identifiable between suppliers and do not hide model design information from other suppliers when shared.

[0068] In the 3GPP Release-18 study on Artificial Intelligence (AI) / Machine Learning (ML) for the NR air interface, several use cases were considered, and evaluation studies were conducted to examine the generalization performance of the ML models used, particularly for Channel State Information (CSI) feedback enhancements with ML models. For model generalization performance, various scenarios / configurations and combinations thereof were used during the training and validation phases to ensure the model could perform across diverse scenarios / configurations. Some of these parameters are specific to the simulated scenario (e.g., UE distribution), some are known based on system information (e.g., carrier frequency), and some are specific to the product implementation (e.g., antenna panel structure and orientation), and may even be specific to RF aspects, such as the power used for a given transmission bandwidth.

[0069] The set of scenarios considered focuses on one or more of the following aspects: - Various deployment scenarios (e.g., City Micro (UMi), Universal Mobile Access (UMa), Indoor Hotspot (InH)); - Various outdoor / indoor UE distributions for UMA / UMi (e.g., 10:0, 8:2, 5:5, 2:8, 0:10); - Various carrier frequencies (e.g., 2GHz, 3.5GHz); - Other aspects of the scenario cannot be excluded, such as various antenna spacings, various antenna virtualizations (TxRU mapping), various ISDs, various UE speeds, etc. - A method for mixing cross-scenario / configuration datasets.

[0070] For various configurations (e.g., configurations that may result in different dimensions of model input / output), the set of configurations considered focuses on one or more of the following aspects: - Various bandwidths (e.g., 10MHz, 20MHz) and / or frequency granularity (e.g., subband size); -CSI feedback provides various payload sizes; - Various antenna port layouts, such as (N1 (number of antenna ports) / N2 / P) and / or antenna port numbers (e.g., 32 ports, 16 ports); -CSI predicts various UE speeds for sub-use cases (e.g., 10km / h, 30km / h, 60km / h, 120km / h, etc.); - Other configuration aspects cannot be excluded, such as various parameter sets, various ranks / layers, etc. In further discussion, some of the aspects mentioned above are referred to as "additional conditions." For AI / ML-supporting features / FG, additional conditions can refer to any aspect of the model's training assumptions, but are not part of the UE capabilities of the AI / ML-supporting features / FG. This does not mean that additional conditions must be specified. Additional conditions can be divided into two categories: NW-side additional conditions and UE-side additional conditions.

[0071] These additional conditions can be used in discussions of model identification and model ID-LCM, where: • For AI / ML model identification and model ID-based lifecycle management (LCM) of the UE portion of the UE-side model and / or dual-side model, model ID-based LCM operates based on the identified model, where the model can be associated with a specific configuration / condition that is associated with the UE capability of the AI / ML enabled feature / FG and additional conditions (e.g., scene, site, and dataset) determined / identified between the UE side and NW side.

[0072] • From RAN1's ​​perspective, the AI / ML model identified by the model ID can be logical, and how it maps to the physical AI / ML model can depend on the implementation. When a distinction is necessary for discussion purposes, companies may use the term "logical AI / ML model" to refer to the model that is identified and assigned a model ID, and "physical AI / ML model" to refer to the actual implementation of such a model.

[0073] The model identifier for RAN1 has been further clarified in Technical Report (TR) 38.843 as follows.

[0074]

[0075]

[0076] The term "network (NW) side additional conditions" used in this paper refers to network-side conditions that may potentially affect different dimensions of model input / output. Based on the definition of model identification above, the model identification process clarifies these "additional conditions," which can include, for example, training dataset categories, site-related information, timestamps, implicit identification information (such as labels for specific gNB / UE implementation details), statistical information (such as latency spread, angular spread, LOS / NLOS data, etc.), and other factors. These conditions can also include model hyperparameters that control how the neural network learns from the data. These hyperparameters affect the underlying model's neural network architecture, optimization, and regularization (controlling overfitting or underfitting). For convolutional neural networks (CNNs), hyperparameters include kernel size, number of kernels, stride length, and pooling size. These parameters directly affect the CNN's performance and training speed; and as network complexity increases, the number of parameters increases, and a chosen set of hyperparameters influences model training, causing significant differences in model performance based on this choice. However, accurately describing these additional conditions remains a challenging task.

[0077] Figure 1 An example communication environment 100 in which exemplary embodiments of the present disclosure can be implemented is shown. In the communication environment 100, two communication devices, including a first device 110 and a second device 120, can communicate with each other.

[0078] exist Figure 1 In the example, the first device 110 may be a terminal device, such as a UE, and the second device 120 may be a network device, such as a core network device or a radio access network device. The service area of ​​the second device 120 may be referred to as cell 102. The communication link 130 between the first device 110 and the second device 120 is an air link.

[0079] It should be understood that Figure 1 The number of devices and their connections shown are for illustrative purposes only and do not imply any limitation. Communication environment 100 may include any suitable number of devices configured to implement the exemplary embodiments of this disclosure. Although not shown, it should be understood that one or more additional devices may be located in cell 102, and one or more additional cells may be deployed in communication environment 100. In some exemplary embodiments, both first device 110 and second device 120 may be UEs. Alternatively, both first device 110 and second device 120 may be radio network devices. In some other exemplary embodiments, both first device 110 and second device 120 may be network devices.

[0080] In the following description, for illustrative purposes, some exemplary embodiments are depicted in which the first device 110 operates as a terminal device and the second device 120 operates as a network device. However, in some exemplary embodiments, the operations described in connection with the terminal device can be implemented at the network device or other devices, and the operations described in connection with the network device can be implemented at the terminal device or other devices.

[0081] In some example embodiments, if the first device 110 is a terminal device and the second device 120 is a network device, the link from the second device 120 to the first device 110 is referred to as a downlink (DL), and the link from the first device 110 to the second device 120 is referred to as an uplink (UL). In the DL, the second device 120 is a transmission (TX) device (or transmitter), and the first device 110 is a reception (RX) device (or receiver). In the UL, the first device 110 is a TX device (or transmitter), and the second device 120 is an RX device (or receiver).

[0082] Communication in communication environment 100 can be implemented according to any suitable communication protocol, including but not limited to cellular communication protocols such as first-generation (1G), second-generation (2G), third-generation (3G), fourth-generation (4G), fifth-generation (5G), and sixth-generation (6G), wireless local network communication protocols such as IEEE 802.11, and / or any other currently known or future-developed protocols. Furthermore, communication can utilize any suitable wireless communication technology, including but not limited to: Code Division Multiple Access (CDMA), Frequency Division Multiple Access (FDMA), Time Division Multiple Access (TDMA), Frequency Division Duplex (FDD), Time Division Duplex (TDD), Multiple-Input Multiple-Output (MIMO), Orthogonal Frequency Division Multiple Access (OFDM), Discrete Fourier Transform Extended OFDM (DFT-s-OFDM), and / or any other currently known or future-developed technologies.

[0083] According to some example embodiments of this disclosure, a solution is provided to address the complex challenges of managing additional conditions on both the network side and the user equipment side (UE side) within the context of training and deploying artificial intelligence (AI) and machine learning (ML) models. Traditional methods suffer from difficulties due to a lack of transparency in the encoding of these conditions, complicating tasks for data processing parties such as UEs or over-the-top (OTT) servers and making it difficult to discern compatibility between different conditions. This disclosure introduces standardized protocols for the identification, encoding, and communication of these conditions. The system ensures that AI / ML models and data samples are precisely matched to the corresponding network and UE environments, aiming to significantly improve the accuracy and efficiency of AI / ML applications within the network and enhance the user experience by adapting services to the different characteristics of each UE.

[0084] The exemplary embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0085] refer to Figure 2A and Figure 2B This illustrates a signaling flow including two portions 200-1 and 200-2, comprising additional conditional groupings, according to some example embodiments of this disclosure. For discussion purposes, reference will be made to... Figure 1 For example, signaling flow can be discussed using the first device 110 and the second device 120. (See reference) Figure 2A and Figure 2B The described signaling flow can be applied to different scenarios, such as beam management, channel state information (CSI) estimation, or location.

[0086] like Figure 2A As shown, the second device 120 transmits (2005) model feature information to the first device 110. In other words, the first device 110 receives (2005) model feature information from the second device 120. The model feature information includes a plurality of first training configurations identified by corresponding first configuration identifiers, and the plurality of first training configurations are associated with communication conditions. For example, the first training configurations may include one or more of the following: a scenario, a reference signal configuration, a data measurement configuration, or a data log configuration. The first configuration identifier may be any suitable kind of information that can identify the first training configuration. In some example embodiments, the model feature information may include any characteristic or property of the computational model, particularly in the fields of machine learning and statistical analysis. Note that the model feature information may include any other suitable model-related information.

[0087] The first device 110 determines (2010) a plurality of second training configurations for each first training configuration. Each second training configuration is identified by a corresponding second configuration identifier. For example, a second training configuration may include one or more of the following: a scenario, a reference signal configuration, a data measurement configuration, or a data log configuration.

[0088] The first device 110 acquires (2015) multiple datasets associated with communication conditions for each second training configuration. The communication conditions are based on multiple first training configurations, and a dataset is acquired based on one of the multiple first training configurations. In some embodiments, the communication conditions may include reference signal transmission. Alternatively or additionally, the communication conditions may include beam management configurations. In some other embodiments, the communication conditions may include positioning configurations.

[0089] The first device 110 transmits (2020) data collection information to the second device 120. In other words, the second device 120 receives (2020) data collection information from the first device 110. The data collection information includes multiple second training configurations identified by corresponding second configuration identifiers.

[0090] The second device 120 acquires (2025) multiple datasets associated with communication conditions for each second training configuration. The communication conditions are based on multiple second training configurations, and a dataset is acquired based on one of the multiple second training configurations.

[0091] In some example embodiments, the first device 110 stores (2030) multiple datasets in a buffer. The buffer can store data according to the training configuration. In this case, each dataset is stored together with a corresponding first configuration identifier and a corresponding second configuration identifier. In some example embodiments, the second device 120 stores (2032) multiple datasets in a buffer, and each dataset is stored together with a corresponding first configuration identifier and a corresponding second configuration identifier. Note that... Figure 2A The storage order of 2030 and 2032 in the example is just an example.

[0092] In some example embodiments, the second device 120 transmits (2035) a first indication to the first device 110. In other words, the first device 110 receives (2035) a first indication from the second device 120. The first indication may indicate a set of first configuration identifiers, and the characteristic differences between a set of first training configurations identified by the corresponding set of first configuration identifiers are less than a threshold. For example, if the characteristic differences are less than the threshold, it means that a set of first training configurations is compatible.

[0093] In some example embodiments, the first device 110 determines (2040) a set of second configuration identifiers. In this case, a set of second training configurations identified by the corresponding set of second configuration identifiers can be compatible with each other. In some other example embodiments, the first device 110 determines (2045) a set of datasets from multiple datasets, and the datasets are based on a set of first configuration identifiers and a set of second configuration identifiers. In some example embodiments, the first device 110 stores (2050) a set of datasets together with a set of first configuration identifiers and a set of second configuration identifiers in a buffer. In some other example embodiments, the first device 110 determines (2055) data identifiers for a set of datasets. In some example embodiments, the first device 110 trains a first model based on a set of datasets.

[0094] In some other example embodiments, the first device 110 transmits (2060) a second instruction to the second device 120. In other words, the second device 120 receives (2060) a second instruction from the first device 110. The second instruction may indicate a set of second configuration identifiers.

[0095] In some example embodiments, the second device 120 determines (2061) a set of datasets from multiple datasets based on a set of first configuration identifiers and a set of second configuration identifiers. In some other example embodiments, the second device 120 stores (2062) a set of datasets together with a set of first configuration identifiers and a set of second configuration identifiers in a buffer that stores data according to a training configuration. In some example embodiments, the second device 120 determines (2063) a dataset identifier for a set of datasets. In some example embodiments, the second device 120 trains (2064) a first model based on a set of datasets.

[0096] like Figure 2B As shown, in some example embodiments, the first device 110 receives a model transfer request (2065) from the second device 120. In other words, the second device 120 transmits the model transfer request (2065) to the first device 110. The model transfer request may include a set of second configuration identifiers.

[0097] In some example embodiments, the first device 110 determines (2070) a subgroup of second configuration identifiers from a set of second configuration identifiers. In this case, the second training configuration identified by the corresponding subgroup of second configuration identifiers is compatible.

[0098] In some example embodiments, the first device 110 transmits a (2075) mode transmission response to the second device 120. In other words, the second device 120 receives a (2075) mode transmission response from the first device 110. The mode transmission response may indicate a second configuration identifier subgroup.

[0099] In some example embodiments, the first device 110 receives (2080) the second model from the second device 120. In other words, the second device 120 transmits (2080) the second model to the first device 110. The second model can be trained at the second device 120 based on a training configuration having a second configuration identifier subgroup.

[0100] In some example embodiments, the first device 110 receives a model transfer request (2085) from the OTT server 210. In other words, the OTT server 210 transmits the model transfer request (2085) to the first device 110. The model transfer request may include a set of second configuration identifiers.

[0101] In some example embodiments, the first device 110 determines (2090) a subgroup of second configuration identifiers from a set of second configuration identifiers. In this case, the second training configuration identified by the corresponding subgroup of second configuration identifiers is compatible.

[0102] In some example embodiments, the first device 110 transmits a (2095) mode transfer response to the OTT server 210. In other words, the OTT server 210 receives a (2095) mode transfer response from the first device 110. The mode transfer response may indicate a second configuration identifier subgroup.

[0103] In some example embodiments, the first device 110 receives (2100) a second model from the OTT server 210. In other words, the OTT server 210 transmits (2100) the second model to the first device 110. The second model can be trained at the OTT server based on a dataset associated with a second configuration identifier subgroup.

[0104] In some example embodiments, the first device 110 transmits (2105) a model transfer request to the second device 120. In other words, the second device 120 receives (2105) a model transfer request from the first device 110. The model transfer request may include a set of first configuration identifiers.

[0105] In some example embodiments, the second device 120 determines (2107) a subgroup of first configuration identifiers from a set of first configuration identifiers. In this case, the first training configuration identified by the corresponding subgroup of first configuration identifiers may be compatible.

[0106] In some example embodiments, the first device 110 receives (2110) a model transfer response from the second device 120. In other words, the second device 120 transfers (2110) a model transfer response to the first device 110. The model transfer response may indicate a first configuration identifier subgroup. In some other example embodiments, the first device 110 transfers (2115) a third model to the second device 120. In other words, the second device 120 may receive (2115) a third model from the first device 110. The third model may be trained based on a training configuration having the first configuration identifier subgroup.

[0107] According to the reference Figure 2A and Figure 2B The described example implementation demonstrates how appropriately selecting data samples can improve the accuracy of AI / ML models. Note that... Figure 2A and Figure 2B The order of steps shown is merely an example and not a limitation.

[0108] refer to Figure 3Example embodiments of additional condition grouping are described in detail, and references are made to... Figure 4 A detailed description of an example implementation of model training is provided.

[0109] refer to Figure 3 This illustrates the signaling flow of user equipment-side and additional conditions packets according to some example embodiments of this disclosure. For discussion purposes, signaling flow 300 can... Figure 1 This is implemented in the communication environment 100 shown. For example, Figure 1 The first device 110 can act as a UE. In some embodiments, the second device 120 can act as a network device.

[0110] like Figure 3 As shown, the second device 120 can define (301) a scenario and select one of M "network additional conditions". The second device 120 can transmit (302) configuration data collection (NW Additional Condition Code ID (A)) to the first device 110. m ), (E){Configuration(Scenario, RS Configuration, Data Measurement Configuration (DataMeasConfig), Data Log Configuration (DataLogConfig))}). In these steps, the second device 120 initiates data collection by the first device 110 by selecting a set of initial "additional conditions" on the network side from a sample space of 1 to M entries; each vector in this space has an index m, thus encoding vector A m This represents one of the vectors. Network "additional conditions" may indicate, for example, the gNB antenna panel configuration, which may contain network-specific information such as the number of antenna elements, the grouping of antenna elements to antenna ports, and the baseband version of the products used. Such information cannot be publicly disclosed and therefore needs to be encoded. Data collection configurations related to the data collection scenario, reference signal configuration, measurement configuration, and logging are encoded in vector E.

[0111] The first device 110 can set (303) a scenario and select N “UE additional conditions”. In this step, the first device 110 sets the reference scenario as indicated in step 302. A novel aspect is that the first device 110 can set some “additional conditions” before initiating data collection. These additional conditions are proprietary information that cannot be publicly disclosed, such as the number of UE antenna panels and their relative positions, the antenna ports given by the manufacturer, the antenna panel orientation, the number of UE antenna elements, the antenna panel switching duty cycle, the number of CPU cores and physical memory allocated for data collection, the physical sampling rate, and the hardware (HW) and software (SW) baseband versions of the chipset used. These additional conditions consist of 1 to N entries, whose encoding forms vector C. n .

[0112] The first device 110 and the second device 120 can perform (304) a data collection initialization handshake. Then, the first device 110 and the second device 120 can transmit (305) the required reference signals to each other according to their configuration. After transmitting the reference signals, the first device 110 and the second device 120 can perform (306) a data collection completion handshake. In these steps, for each vector C... n Data collection proceeds to the initial handshake at step 304, the necessary reference signals are transmitted at step 305 and alignment is configured on both sides of the second device 120 and the first device 110, the handshake is completed at step 306, and the novel aspect at step 307 is that the first device 110 indicates the encoding vector C to the second device 120. n .

[0113] The first device 110 can store (308) UE-side data buffer (data samples at NW, A) m C n The second device 120 can store (309) NW-side data buffer (data samples at the UE, A) m C n The second device 120 and the first device 110 may have separate data buffers; for each data buffer, due to A m and C n The results of the iterative data collection are stored in the second device 120 and the first device 110. Steps 301 to 309 do not preclude data collection and storage only on one side (i.e., the UE or the network side). These steps are described to illustrate how data collection is organized on both sides and to clarify the aspects involved on both sides.

[0114] In some example embodiments, the second device 120 transmits (310) NW Add-Cond Grouping (NWAdd-CondGrouping). In some other example embodiments, the first device 110 buffers the UE-side data (each group {A m C n The data samples of {A} and the UE AddlCondGrouping and NW AddlCondGrouping are finally determined (311) as the UE dataset ID (DataSetID_UE). In some example embodiments, the first device 110 transmits (312) the UE AddlCondGrouping. In some other example embodiments, the second device 120 buffers the NW side data (each group {A}) and determines (311) the UE dataset ID (DataSetID_UE). m C nThe data samples of E) and UEAddlCondGrouping and NWAddlCondGrouping were finally determined to be the network dataset ID (DataSetID_NW).

[0115] In steps 310 to 313, the second device 120 can transmit NWAddlCondGrouping via signaling to indicate to the first device 110 which coded vectors are compatible with each other from a network perspective, for example, having the same antenna panel orientation and number of elements, the same reference signal configuration, and the same enabled beam grid. The reason for indicating a set of compatible vectors is as follows: • The data collected for each compatible vector share similar characteristics (e.g., sample mean and variance, signal strength values ​​conforming to a given range, etc.); • Model training can incorporate data from datasets with compatible vectors to improve robustness during model training; • Model training can mix data from datasets with incompatible vectors to improve model generalization performance; • The trained model should reflect which mixture of compatible and incompatible vectors was used during model training—remember that (multiple) neighboring gNBs should still be able to use the features of the ML model when trained on a dataset based on a mixture of compatible vectors, or not use the features when the vectors are incompatible for a given gNB.

[0116] • Hyperparameters used for model training; • Input data preprocessing; • Post-processing of output data.

[0117] The first device 110 can transmit UEAddlCondGrouping to the second device 120 via signaling to indicate to the second device 120 which encoded vectors are compatible with each other from a network perspective, according to the same standard described above for NWAddlCondGrouping. Both the second device 120 and the first device 110 mark the final data buffer, which includes data samples across “additional condition” combinations, and is marked by UEAddlCondGrouping and NWAddlCondGrouping. Data batches and their complete information are pointed to by the identifiers DataSetID_UE and DataSetID_NW. Execution Figure 3 After the steps in the process, it is expected that the data collected at the first device 110 and the second device 120 will have the encoding of “additional conditions” on the UE side, and will also have an understanding of how these “additional conditions” are related, i.e. which IDs are compatible or incompatible with each other.

[0118] In some example embodiments, A m and Cn The encoding can represent the combination of NW-side additional conditions (antenna element assumption 1, beamcodebook assumption 1, etc.) as a vector with X entries (the number of entries depends on the number of different additional conditions. It also usually depends on the use case, and in this example, beam management is considered). Each entry in the vector depends on the possible number of the corresponding additional condition. For example, the beam management (BM) codebook can have 12 possibilities, while the antenna setup can have only 4 possibilities, and another additional condition can have 8 possibilities. Therefore, the number of NW-side additional condition combinations can be the total number of vectors that the second device 120 can have. The space formed by all combinations is M = 12 × 4 × 8 = 384 combinations. In the case of log2(384), the second device 120 can share only the index, i.e., 9 bits are sufficient. The second device 120 can also add more bits, without directly sharing the index, but instead perform some kind of encryption and make it 16 bits (the encryption algorithm and the number of added bits can be specified by the standard or determined by the vendor). In this case, 16 bits is A m The encoded ID.

[0119] In some other example embodiments, the UE-side "additional conditions" for beam management use cases may consider receiver antenna panel element assumption 1, the number of configured antenna panels assumption 2, antenna panel switching algorithm assumption 3, physical memory configuration assumption 4, baseband receiver architecture assumption 5, etc., and may be represented as a vector with 5 entries. The number of UE-side "additional conditions" formed based on this can be 14 bits, and the first device 110 may add encryption to make it 24 bits (the encryption algorithm and the number of added bits may be specified by standards or vendor-specific). In this case, 24 bits is C. n The encoded ID, such as Figure 4 As shown.

[0120] refer to Figure 4 This illustrates signaling flows trained and identified using information from NW and UE-side additional conditional groups according to some example embodiments of this disclosure. For discussion purposes, signaling flow 400 may be... Figure 1 This is implemented in the communication environment 100 shown. For example, Figure 1 The first device 110 can act as a UE. In some embodiments, the second device 120 can act as a network, OAM, LMF, etc.

[0121] Step 401 involves training a model for the UE at the network, OAM, LMF, etc., based on a mixture of NWAddlCondGrouping and UEAddlCondGrouping data. The second device 120 can transmit (402) an ML model request (UEAddlCondGrouping) to the first device 110. After step 402, the first device 110 can check (403) the compatibility vectors in the UEAddlCondGrouping. Then, the first device 110 can transmit (404) an ML model response (compatibility vectors) to the second device 120. The first device 110 can continue (405) transmitting the ML model for the compatibility vectors. These steps describe the scenario of training an ML model on the network side (e.g., gNB, OAM, LMF, CN, etc.). As described above, different types of ML models can now be implemented using UEAddlCondGrouping and NWAddlCondGrouping information: a generalized ML model considering a mixture of data samples across incompatible vectors, a more localized ML model created by mixing data samples across compatible vectors, or a mixture of both. How data samples are mixed, which samples are selected, and which samples are excluded are proprietary implementations, but the input is explicitly based on UEAddlCondGrouping and NWAddlCondGrouping. When the network trains this model for a UE and needs to download the ML model to a new UE, it can request the first device 110 to perform a compatibility check to ensure that the first device 110 supports at least some compatible vectors in UEAddlCondGrouping.

[0122] The following steps illustrate model training at the first device 110 or OTT server 420. Step 406 is model training for the UE at the OTT, based on a mix of NWAddlCondGrouping and UEAddlCondGrouping data. The OTT server 420 may transmit (407) an ML model request (UEAddlCondGrouping) to the first device 110. After step 407, the first device 110 and the OTT server 420 may check (408) the compatibility vector in the UEAddlCondGrouping. Then, the first device 110 may transmit (409) an ML model response (compatibility vector) to the OTT server 420. The first device 110 and the OTT server 420 may continue (410) with ML model transmission for the compatibility vector. Subsequently, the first device 110, the OTT server 420, and the second device 120 may continue (411) with the ML model identifier shown in step 413 and subsequent steps. These steps describe the steps for training an ML model at the OTT server 420, which are similar in aspects to steps 401 to 405. Finally, the first device 110 can initiate step 413 and subsequent steps in step 411 to perform ML model identification.

[0123] The following steps illustrate training an ML model at a first device 110. Step 412 involves training the model for the UE at the OTT, based on a mix of NWAddlCondGrouping and UEAddlCondGrouping data. The first device 110 may transmit (413) an ML model identification request (NWAddlCondGrouping) to the second device 120. After transmission, the second device 120 may check (414) the compatibility vector in the NWAddlCondGrouping. The second device 120 may transmit (415) an ML model identification response (compatibility vector) to the first device 110. In step 416, the first device 110, the OTT server 420, and the second device 120 may use only the functionality of the ML model in the compatibility vector. These steps describe training an ML model on the UE side based on UEAddlCondGrouping and NWAddlCondGrouping. The mix instruction may come from the OTT server or be provided by the network. After model training is complete, the UE's model identifier, NWAddlCondGrouping, can be transmitted to the network upon arrival to assess which NW additional conditions the network might have that are compatible. If any compatible vectors exist, the network can continue to use the functionality of these ML models and ignore the rest of the models.

[0124] Figure 5A flowchart of an example method 500 implemented at a first device according to some example embodiments of the present disclosure is shown. For the purposes of discussion, [the following will be discussed]. Figure 1 Method 500 is described by the angle of the first device 110 in the middle.

[0125] At block 510, the first device 110 receives from the second device model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions.

[0126] At box 520, the first device 110 determines a plurality of second training configurations identified by a corresponding second configuration identifier for each first training configuration.

[0127] At box 530, the first device 110 acquires multiple datasets associated with communication conditions for each second training configuration based on multiple first training configurations, wherein a dataset is acquired based on a first training configuration from multiple first training configurations.

[0128] At frame 540, the first device 110 transmits data collection information to the second device, including multiple second training configurations identified by corresponding second configuration identifiers.

[0129] In some example embodiments, method 500 further includes storing multiple datasets in a buffer that stores data according to a corresponding first training configuration and a corresponding second training configuration.

[0130] In some example embodiments, method 500 further includes receiving from a second device a first indication indicating a set of first configuration identifiers, wherein a set of first training configurations identified by a corresponding set of first configuration identifiers are compatible with each other.

[0131] In some example embodiments, method 500 further includes: determining a set of second configuration identifiers, wherein a set of second training configurations identified by a corresponding set of second configuration identifiers are compatible with each other; determining a set of datasets from a plurality of datasets based on a set of first configuration identifiers and a set of second configuration identifiers; storing a set of datasets together with a set of first configuration identifiers and a set of second configuration identifiers in a buffer; and determining a dataset identifier for the set of datasets.

[0132] In some example embodiments, method 500 further includes: training a first model based on a set of datasets.

[0133] In some example embodiments, method 500 further includes transmitting a second indication to the second device that indicates a set of second configuration identifiers.

[0134] In some example embodiments, method 500 further includes: receiving a model transfer request from a second device, wherein the model transfer request includes a set of second configuration identifiers; determining a second configuration identifier subgroup from the set of second configuration identifiers, wherein a second training configuration identified by a corresponding second configuration identifier subgroup is compatible; transmitting a model transfer response indicating the second configuration identifier subgroup to the second device; and receiving a second model trained based on a training configuration having the second configuration identifier subgroup from the second device.

[0135] In some example embodiments, method 500 further includes: transmitting a model transmission request to a second device, wherein the model transmission request includes a set of first configuration identifiers; receiving a model transmission response from the second device indicating a subgroup of the first configuration identifiers; and transmitting a third model to the second device, the third model being trained based on a training configuration utilizing the subgroup of the first configuration identifiers.

[0136] In some example embodiments, the first device includes a terminal device, and the second device is a core network device or a radio access network device.

[0137] Figure 6 A flowchart of an example method 600 implemented at a second device according to some example embodiments of the present disclosure is shown. For discussion purposes, [the following will be discussed]. Figure 1 Method 600 is described by the angle of the second device 120 in the middle.

[0138] At box 610, the second device 120 transmits model feature information to the first device, including multiple first training configurations identified by corresponding first configuration identifiers, wherein the multiple first training configurations are associated with communication conditions.

[0139] At box 620, for each first training configuration, the second device 120 receives data collection information from the first device, the data collection information including multiple second training configurations identified by corresponding second configuration identifiers.

[0140] At box 630, the second device 120 acquires multiple datasets associated with communication conditions for each second training configuration based on multiple second training configurations, wherein one dataset is acquired based on one second training configuration from multiple second training configurations.

[0141] In some example embodiments, method 600 further includes storing multiple datasets in a buffer that stores data according to a corresponding first training configuration and a corresponding second training configuration.

[0142] In some example embodiments, method 600 further includes transmitting to a first device a first indication indicating a set of first configuration identifiers, wherein the characteristic differences between a set of first training configurations identified by the corresponding set of first configuration identifiers are less than a threshold.

[0143] In some example embodiments, method 600 further includes receiving from a first device a second indication indicating a set of second configuration identifiers, wherein a set of second training configurations identified by the respective set of second configuration identifiers are compatible with each other.

[0144] In some example embodiments, method 600 further includes: determining a set of datasets from multiple datasets based on a set of first configuration identifiers and a set of second configuration identifiers; storing the set of datasets together with the set of first configuration identifiers and the set of second configuration identifiers in a buffer; and determining a dataset identifier for the set of datasets.

[0145] In some example embodiments, method 600 further includes: training a first model based on a set of datasets.

[0146] In some example embodiments, method 600 further includes: transmitting a model transmission request to a first device, wherein the model transmission request includes a set of second configuration identifiers; receiving a model transmission response from the first device indicating a subgroup of the second configuration identifiers; and transmitting a second model to the first device, the second model being trained based on a training configuration utilizing the subgroup of the second configuration identifiers.

[0147] In some example embodiments, method 600 further includes: receiving a model transfer request from a first device, wherein the model transfer request includes a set of first configuration identifiers; determining a first configuration identifier subgroup from the set of first configuration identifiers, wherein a first training configuration identified by a corresponding first configuration identifier subgroup is compatible; transmitting a model transfer response indicating a second configuration identifier subgroup to the first device; and receiving a third model from the first device based on a training configuration having the first configuration identifier subgroup.

[0148] In some example embodiments, the first device includes a terminal device, and the second device is a core network device or a radio access network device.

[0149] In some example embodiments, a first means capable of performing any of the methods in method 500 (e.g. Figure 1 The first device 110 may include components for performing the corresponding operations of method 500. These components may be implemented in any suitable form. For example, these components may be implemented in a circuit or software module. The first device may be implemented as or included in Figure 1 In the first device 110.

[0150] In some example embodiments, the first device includes: means for receiving from the second device model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; means for determining a plurality of second training configurations identified by corresponding second configuration identifiers for each first training configuration; means for acquiring, for each second training configuration, a plurality of datasets associated with the communication conditions based on the plurality of first training configurations, wherein one dataset is acquired based on a first training configuration from the plurality of first training configurations; and means for transmitting to the second device data collection information including the plurality of second training configurations identified by corresponding second configuration identifiers.

[0151] In some example embodiments, the first device further includes a component for storing multiple datasets in a buffer, the buffer storing data according to a corresponding first training configuration and a corresponding second training configuration.

[0152] In some example embodiments, the first device further includes a component for receiving from the second device a first indication of a set of first configuration identifiers, wherein a set of first training configurations identified by a corresponding set of first configuration identifiers are compatible with each other.

[0153] In some example embodiments, the first apparatus further includes: components for determining a set of second configuration identifiers, wherein a set of second training configurations identified by the respective set of second configuration identifiers are compatible with each other; components for determining a set of datasets from a plurality of datasets based on a set of first configuration identifiers and a set of second configuration identifiers; components for storing a set of datasets together with a set of first configuration identifiers and a set of second configuration identifiers in a buffer; and components for determining dataset identifiers for a set of datasets.

[0154] In some example embodiments, the first device further includes a component for training a first model based on a set of datasets.

[0155] In some example embodiments, the first device further includes a component for transmitting to the second device a second indication indicating a set of second configuration identifiers.

[0156] In some example embodiments, the first device further includes: means for receiving a model transfer request from the second device, wherein the model transfer request includes a set of second configuration identifiers; means for determining a second configuration identifier subgroup from the set of second configuration identifiers, wherein the second training configuration identified by the corresponding second configuration identifier subgroup is compatible; means for transmitting a model transfer response indicating the second configuration identifier subgroup to the second device; and means for receiving a second model trained based on a training configuration having the second configuration identifier subgroup from the second device.

[0157] In some example embodiments, the first device further includes: a component for transmitting a model transmission request to the second device, wherein the model transmission request includes a set of first configuration identifiers; a component for receiving a model transmission response indicating a subgroup of the first configuration identifiers from the second device; and a component for transmitting a third model to the second device, the third model being trained based on a training configuration utilizing the subgroup of the first configuration identifiers.

[0158] In some example embodiments, the first device includes a terminal device, and the second device is a core network device or a radio access network device.

[0159] In some example embodiments, the first device further includes components for performing other operations in some example embodiments of method 500 or the first device 110. In some example embodiments, the components include: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the first device to perform.

[0160] In some example embodiments, a second means capable of performing any of the methods in method 600 (e.g. Figure 1 The second device 120 may include components for performing the corresponding operations of method 600. These components may be implemented in any suitable form. For example, these components may be implemented in a circuit or software module. The second device may be implemented as or included in Figure 1 The second device 120 in the middle.

[0161] In some example embodiments, the second device includes: components for transmitting to the first device model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; components for receiving data collection information from the first device for each first training configuration, the data collection information including a plurality of second training configurations identified by corresponding second configuration identifiers; and components for obtaining, for each second training configuration, a plurality of datasets associated with the communication conditions based on the plurality of second training configurations, wherein one dataset is obtained based on a second training configuration from the plurality of second training configurations.

[0162] In some example embodiments, the second apparatus further includes a component for storing multiple datasets in a buffer, the buffer storing data according to a corresponding first training configuration and a corresponding second training configuration.

[0163] In some example embodiments, the second device further includes: a component for transmitting to the first device a first indication indicating a set of first configuration identifiers, wherein the characteristic differences between a set of first training configurations identified by the corresponding set of first configuration identifiers are less than a threshold.

[0164] In some example embodiments, the second device further includes a component for receiving from the first device a second indication indicating a set of second configuration identifiers, wherein a set of second training configurations identified by the respective set of second configuration identifiers are compatible with each other.

[0165] In some example embodiments, the second apparatus further includes: components for determining a set of datasets from a plurality of datasets based on a set of first configuration identifiers and a set of second configuration identifiers; components for storing the set of datasets together with the set of first configuration identifiers and the set of second configuration identifiers in a buffer; and components for determining a dataset identifier for the set of datasets.

[0166] In some example embodiments, the second device further includes a component for training the first model based on a set of datasets.

[0167] In some example embodiments, the second device further includes: means for transmitting a model transmission request to the first device, wherein the model transmission request includes a set of second configuration identifiers; means for receiving a model transmission response indicating a subgroup of the second configuration identifiers from the first device; and means for transmitting a second model to the first device, the second model being trained based on a training configuration utilizing the subgroup of the second configuration identifiers.

[0168] In some example embodiments, the second apparatus further includes: means for receiving a model transfer request from the first apparatus, wherein the model transfer request includes a set of first configuration identifiers; means for determining a first configuration identifier subgroup from the set of first configuration identifiers, wherein the first training configuration identified by the corresponding first configuration identifier subgroup is compatible; means for transmitting to the first apparatus a model transfer response indicating a second configuration identifier subgroup; and means for receiving from the first apparatus a third model transferred based on a training configuration having the first configuration identifier subgroup.

[0169] In some example embodiments, the first device includes a terminal device, and the second device is a core network device or a radio access network device.

[0170] In some example embodiments, the second device further includes components for performing other operations in some example embodiments of method 600 or the second device 120. In some example embodiments, the components include: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the second device to perform a corresponding operation.

[0171] Figure 7 This is a simplified block diagram of a device 700 suitable for implementing exemplary embodiments of the present disclosure. Device 700 can be provided to implement a communication device, for example, Figure 1The first device 110 or the second device 120 shown. As shown, the device 700 includes one or more processors 710, one or more memories 720 coupled to the processors 710, and one or more communication modules 740 coupled to the processors 710.

[0172] Communication module 740 is used for bidirectional communication. Communication module 740 has one or more communication interfaces to facilitate communication with one or more other modules or devices. The communication interface can represent any interface necessary for communication with other network elements. In some example embodiments, communication module 740 may include at least one antenna.

[0173] Processor 710 can be any type suitable for a local technology network and may include one or more of the following as non-limiting examples: general-purpose computer, special-purpose computer, microprocessor, digital signal processor (DSP), and processor based on a multi-core processor architecture. Device 700 may have multiple processors, such as application integrated circuit chips, which are time-subordinate to a clock synchronized with the main processor.

[0174] Memory 720 may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, read-only memory ROM 724, electrically programmable read-only memory EPROM, flash memory, hard disk, optical disc CD, digital video disc DVD, optical disc, laser disc, and other magnetic and / or optical memories. Examples of volatile memories include, but are not limited to, random access memory RAM 722 and other volatile memories that will not be retained during power loss.

[0175] Computer program 730 includes computer-executable instructions that are executed by an associated processor 710. The instructions of program 730 may include instructions for performing operations / actions of some example embodiments of this disclosure. Program 730 may be stored in memory, such as ROM 724. Processor 710 can perform any suitable actions and processes by loading program 730 into RAM 722.

[0176] Example embodiments of this disclosure can be implemented using the method of procedure 730, such that device 700 can perform as shown in Figures 2 to 30. Figure 6 Any process discussed in this disclosure. Exemplary embodiments of this disclosure may also be implemented by hardware or a combination of software and hardware.

[0177] In some example embodiments, program 730 may be tangibly contained in a computer-readable medium, which may be included in device 700 (such as in memory 720) or other storage devices accessible by device 700. Device 700 may load program 730 from the computer-readable medium into RAM 722 for execution. In some example embodiments, the computer-readable medium may include any type of non-transitory storage medium, such as ROM, EPROM, flash memory, hard disk, CD, DVD, etc. As used herein, the term "non-transitory" is a limitation of the medium itself (i.e., tangible, not tactile), and not a limitation of data storage persistence (e.g., RAM vs. ROM).

[0178] In general, the various embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects can be implemented in hardware, and others can be implemented in firmware or software executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of this disclosure are illustrated and described as block diagrams, flowcharts, or using some other graphical representation, it should be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof, as examples of non-limiting examples.

[0179] Some exemplary embodiments of this disclosure also provide at least one computer program product tangibly stored on a computer-readable medium (such as a non-transitory computer-readable medium). The computer program product includes computer-executable instructions, such as those included in a program module, which are executed in a device targeting a physical or virtual processor to implement any of the methods described above. Typically, a program module includes routines, programs, libraries, objects, classes, components, data structures, etc., that perform a particular task or implement a particular abstract data type. The functionality of the program module can be combined or split as needed among program modules in various embodiments. The machine-executable instructions for the program module can be executed within a local or distributed device. In a distributed device, the program module can reside on both local and remote storage media.

[0180] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. The program code may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus, such that, when executed by the processor or controller, the program code enables the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0181] In the context of this disclosure, computer program code or related data may be carried on any suitable carrier to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carriers include signals, computer-readable media, etc.

[0182] Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable media can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any suitable combination thereof. More specific examples of computer-readable storage media will include electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable optical disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0183] Furthermore, although operations are described in a specific order, this should not be construed as requiring such operations to be performed in the specific order shown or in sequential order, or to perform all shown operations to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the discussion above, these should not be construed as limiting the scope of this disclosure, but rather as descriptions of features that may be specific to particular embodiments. Unless explicitly stated otherwise, certain features described in the context of a single embodiment may also be implemented in combination in a single embodiment. Conversely, unless explicitly stated otherwise, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0184] Although this disclosure has been described in language specific to structural features and / or methodological actions, it should be understood that the disclosure as defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are disclosed as exemplary forms for implementing the claims.

Claims

1. A first device, comprising: At least one processor; as well as At least one memory, the at least one memory storing instructions, the instructions, when executed by the at least one processor, cause the first device to: The second device receives model feature information, which includes a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions. For each first training configuration, determine a plurality of second training configurations identified by a corresponding second configuration identifier; For each second training configuration, multiple datasets associated with communication conditions are obtained based on the plurality of first training configurations, wherein one dataset is obtained based on a first training configuration from the plurality of first training configurations; as well as Data collection information is transmitted to the second device, the data collection information including the plurality of second training configurations identified by corresponding second configuration identifiers.

2. The first device according to claim 1, wherein the first device is configured to: The multiple datasets are stored in a buffer, which stores data according to a corresponding first training configuration and a corresponding second training configuration.

3. The first device according to claim 1 or 2, wherein the first device is configured to: The device receives a first instruction, which indicates a set of first configuration identifiers, wherein the characteristic difference between a set of first training configurations identified by the corresponding set of first configuration identifiers is less than a threshold.

4. The first device according to claim 3, wherein the first device is configured to: A set of second configuration identifiers is determined, wherein a set of second training configurations identified by the corresponding set of second configuration identifiers are compatible with each other; A set of datasets is determined from the plurality of datasets based on the first set of configuration identifiers and the second set of configuration identifiers; The set of datasets, together with the set of first configuration identifiers and the set of second configuration identifiers, are stored in the buffer; as well as Determine the dataset identifier for the set of datasets.

5. The first device according to claim 4, wherein the first device is configured to: The first model is trained based on the first set of training configurations.

6. The first device according to claim 5, wherein the first device is configured to: A second instruction is transmitted to the second device, the second instruction indicating the set of second configuration identifiers.

7. The first device according to any one of claims 1-6, wherein the first device is configured to: Receive a model transfer request from the second device, wherein the model transfer request includes a set of second configuration identifiers; A second configuration identifier subgroup is determined from the set of second configuration identifiers, wherein the second training configuration identified by the corresponding second configuration identifier subgroup is compatible; Transmit a mode transmission response to the second device, the mode transmission response indicating the second configuration identifier subgroup; as well as The second model is received from the second device, and the second model is trained based on a training configuration associated with the second configuration identifier subgroup.

8. The first device according to any one of claims 1-6, wherein the first device is configured to: Transmit a model transmission request to the second device, wherein the model transmission request includes a set of first configuration identifiers; Receive a mode transmission response from the second device, the mode transmission response indicating the first configuration identifier subgroup; and A third model is transmitted to the second device, the third model being trained based on a training configuration associated with the first configuration identifier subgroup.

9. The first device according to any one of claims 1-8, wherein the first device includes a terminal device, and the second device is a core network device or a radio access network device.

10. A second device, comprising: At least one processor; as well as At least one memory, the at least one memory storing instructions, the instructions, when executed by the at least one processor, cause the second device to: Transmit model feature information to a first device, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; For each first training configuration, data collection information is received from the first device, the data collection information including a plurality of second training configurations identified by corresponding second configuration identifiers; as well as For each second training configuration, multiple datasets associated with communication conditions are obtained based on the plurality of second training configurations, wherein one dataset is obtained based on a second training configuration from the plurality of second training configurations.

11. The second device according to claim 10, wherein the second device is configured to: The multiple datasets are stored in a buffer, which stores data according to a corresponding first training configuration and a corresponding second training configuration.

12. The second device according to claim 10 or 11, wherein the second device is configured to: A first instruction is transmitted to the first device, the first instruction indicating a set of first configuration identifiers, wherein the characteristic difference between a set of first training configurations identified by the corresponding set of first configuration identifiers is less than a threshold.

13. The second device according to claim 12, wherein the second device is configured to: The device receives a second instruction, which indicates a set of second configuration identifiers, wherein a set of second training configurations identified by the corresponding set of second configuration identifiers are compatible with each other.

14. The second device according to claim 13, wherein the second device is configured to: A set of datasets is determined from the plurality of datasets based on the first set of configuration identifiers and the second set of configuration identifiers; The set of datasets, together with the set of first configuration identifiers and the set of second configuration identifiers, are stored in the buffer; as well as Determine the dataset identifier for the set of datasets.

15. The second device according to claim 14, wherein the second device is configured to: The first model is trained based on the second set of training configurations.

16. The second device according to any one of claims 10-15, wherein the second device is configured to: Transmit a model transmission request to the first device, wherein the model transmission request includes a set of second configuration identifiers; Receive a model transfer response from the first device, the model transfer response indicating the second configuration identifier subgroup; and A second model is transmitted to the first device, the second model being trained based on a training configuration associated with the second configuration identifier subgroup.

17. The second device according to any one of claims 10-15, wherein the second device is configured to: Receive a model transfer request from the first device, wherein the model transfer request includes a set of first configuration identifiers; A first configuration identifier subgroup is determined from the set of first configuration identifiers, wherein the first training configuration identified by the corresponding first configuration identifier subgroup is compatible; A mode transmission response is transmitted to the first device, the mode transmission response indicating the second configuration identifier subgroup; as well as A third model is received from the first device, the third model being transmitted based on a training configuration associated with the first configuration identifier subgroup.

18. The second apparatus according to any one of claims 10-17, wherein the first apparatus includes a terminal device, and the second apparatus is a core network device or a radio access network device.

19. A method implemented at a first device, comprising: The second device receives model feature information, which includes a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions. For each first training configuration, determine a plurality of second training configurations identified by a corresponding second configuration identifier; For each second training configuration, multiple datasets associated with communication conditions are obtained based on the plurality of first training configurations, wherein one dataset is obtained based on a first training configuration from the plurality of first training configurations; as well as Data collection information is transmitted to the second device, the data collection information including the plurality of second training configurations identified by corresponding second configuration identifiers.

20. A method implemented at a second device, comprising: Transmit model feature information to a first device, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; For each first training configuration, data collection information is received from the first device, the data collection information including a plurality of second training configurations identified by corresponding second configuration identifiers; as well as For each second training configuration, multiple datasets associated with communication conditions are obtained based on the plurality of second training configurations, wherein one dataset is obtained based on a second training configuration from the plurality of second training configurations.

21. A first device, comprising: A component for receiving model feature information from a second device, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; For each first training configuration, a component is used to determine a plurality of second training configurations identified by a corresponding second configuration identifier; A component for acquiring multiple datasets associated with communication conditions based on the plurality of first training configurations for each second training configuration, wherein one dataset is acquired based on a first training configuration from the plurality of first training configurations; as well as Components for transmitting data collection information to the second device, the data collection information including the plurality of second training configurations identified by corresponding second configuration identifiers.

22. A second device, comprising: A component for transmitting model feature information to a first device, the model feature information including a plurality of first training configurations identified by corresponding first configuration identifiers, wherein the plurality of first training configurations are associated with communication conditions; A component for receiving data collection information from the first device for each first training configuration, the data collection information including a plurality of second training configurations identified by corresponding second configuration identifiers; as well as A component for acquiring multiple datasets associated with communication conditions based on the plurality of second training configurations for each second training configuration, wherein one dataset is acquired based on a second training configuration from the plurality of second training configurations.

23. A computer-readable medium comprising instructions stored thereon, the instructions being configured to cause a device to perform at least the method of claim 19 or claim 20.

Images