Information transmission method, communication node, storage medium and computer program product

Abstract

The invention provides an information transmission method, a communication node, a storage medium and a computer program product. The method is applied to first communication equipment and comprises the following steps: sending a discovery request to second communication equipment; receiving first communication equipment information fed back by the second communication equipment, wherein the first communication equipment information is determined according to model life cycle capability information of the communication equipment; and determining a third communication device according to the first communication device information, sending a data processing request to the third communication device, and receiving processed data fed back by the third communication device, thereby solving the problem that a data processing task cannot be normally executed due to the fact that the first communication device cannot perform data processing. The third communication device assists the first communication device in data processing, so that the data processing task is prevented from being forced to be terminated or delayed, and the data processing efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and in particular to information transmission methods, communication nodes, storage media, and computer program products. Background Technology

[0002] Different communication devices have varying computing power. Taking network elements with Network Data Analysis Functions (NWDAF) as an example, some NWDAFs have more computing power than others. Therefore, different communication devices undertake different types of tasks. Those with higher computing power can handle computationally expensive tasks like training, while those with lower computing power can handle tasks such as model testing and dataset preprocessing. In the current TS23.388 protocol, the model lifecycle definition supported by communication devices is not comprehensive. During the model lifecycle, only the communication device can perform data processing tasks; if the communication device cannot perform the data processing task, the task must be terminated or wait, and cannot be executed normally. Summary of the Invention

[0003] This application provides an information transmission method, a communication node, a storage medium, and a computer program product to solve the problem that data processing tasks cannot be executed normally when communication equipment is unable to process data.

[0004] To achieve the above objectives, embodiments of this application provide an information transmission method applied to a first communication device, comprising:

[0005] Send a discovery request to the second communication device;

[0006] Receive first communication device information fed back by the second communication device, wherein the first communication device information is determined based on the model lifecycle capability information of the communication device;

[0007] The third communication device is determined based on the information of the first communication device, a data processing request is sent to the third communication device, and the processed data is received from the third communication device.

[0008] To achieve the above objectives, embodiments of this application provide another information transmission method applied to a third communication device, including:

[0009] Receive data processing requests sent by the first communication device;

[0010] The processed data is fed back to the first communication device.

[0011] To achieve the above objectives, embodiments of this application provide another information transmission method, applied to a first communication device, including:

[0012] Receive the second model management request sent by the sixth communication device;

[0013] The model is then processed.

[0014] To achieve the above objectives, embodiments of this application provide another information transmission method applied to a second communication device, including:

[0015] Receive model lifecycle capability information sent by a seventh communication device, wherein the model lifecycle capability information is used to indicate the model lifecycle capability of the seventh communication device;

[0016] Store the model's lifecycle capability information.

[0017] To achieve the above objectives, embodiments of this application provide another information transmission method applied to a second communication device, including:

[0018] Receive a machine learning task sent by an eighth communication device, the machine learning task including the task's estimated power;

[0019] The second communication device information is determined based on the estimated capabilities in the machine learning task and the pre-stored model lifecycle capability information of the communication device.

[0020] The information from the second communication device is fed back to the eighth communication device.

[0021] To achieve the above objectives, embodiments of this application provide a communication node, including: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for implementing communication between the processor and the memory. When the program is executed by the processor, it implements the steps of the information transmission method as described in any one of the embodiments of this application.

[0022] To achieve the above objectives, embodiments of this application provide a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the information transmission method described in any one of the embodiments of this application.

[0023] To achieve the above objectives, embodiments of this application provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the information transmission method described in any one of the embodiments of this application.

[0024] The information transmission method, communication node, storage medium, and computer program product provided in this application embodiment involve a first communication device sending a discovery request to a second communication device. Upon receiving the discovery request, the second communication device determines the information of the first communication device based on its model lifecycle capability information and feeds back the first communication device information. The first communication device then determines a third communication device based on the first communication device information, sends a data processing request to the third communication device, instructs the third communication device to perform data processing, and receives the processed data fed back by the third communication device. By instructing the third communication device to perform data processing, the first communication device solves the problem of data processing tasks failing to execute normally due to the first communication device's inability to perform data processing. By having the third communication device assist the first communication device in performing data processing, the data processing task is prevented from being forcibly terminated or delayed, thereby improving data processing efficiency.

[0025] Further details regarding the above embodiments and other aspects of this application, as well as their implementations, are provided in the accompanying drawings, detailed description, and claims. Attached Figure Description

[0026] Figure 1a An architecture diagram of a core network is provided for one embodiment;

[0027] Figure 1b An architecture diagram of data collection for a network data analysis function is provided as an embodiment;

[0028] Figure 1c A network data analysis exposure architecture is provided as one embodiment;

[0029] Figure 1d An architecture diagram of data collection coordination and functionality is provided for one embodiment;

[0030] Figure 1e Another network data analysis exposure architecture is provided as an embodiment;

[0031] Figure 2 A flowchart illustrating the model lifecycle is provided for one embodiment.

[0032] Figure 3 A flowchart illustrating an information transmission method provided in one embodiment;

[0033] Figure 4 A flowchart illustrating another information transmission method provided in one embodiment;

[0034] Figure 5 A flowchart illustrating another information transmission method provided in one embodiment;

[0035] Figure 6A flowchart illustrating another information transmission method provided in one embodiment;

[0036] Figure 7 A flowchart illustrating another information transmission method provided in one embodiment;

[0037] Figure 8 A flowchart illustrating another information transmission method provided in one embodiment;

[0038] Figure 9 A flowchart illustrating another information transmission method provided in one embodiment;

[0039] Figure 10 An example diagram illustrating the interaction of registration model lifecycle capability information is provided in one embodiment.

[0040] Figure 11 An example diagram illustrating the interaction for updating model lifecycle capability information is provided in one embodiment.

[0041] Figure 12 An interactive example diagram of data preprocessing is provided as an embodiment;

[0042] Figure 13 An example diagram illustrating the interaction between model training and validation is provided for one embodiment.

[0043] Figure 14 An example diagram illustrating the interaction of a communication device matching method is provided in one embodiment.

[0044] Figure 15 A schematic diagram of the structure of an information transmission device provided in one embodiment;

[0045] Figure 16 A schematic diagram of another information transmission device provided in one embodiment;

[0046] Figure 17 A schematic diagram of another information transmission device provided in one embodiment;

[0047] Figure 18 A schematic diagram of another information transmission device provided in one embodiment;

[0048] Figure 19 A schematic diagram of another information transmission device provided in one embodiment;

[0049] Figure 20 This is a schematic diagram of the structure of a communication node provided in one embodiment. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be arbitrarily combined with each other.

[0051] Figure 1a This document provides an architecture diagram of a core network as an embodiment, specifically a 5G core network architecture. This architecture has the following functions:

[0052] 1)UE, User Equipment.

[0053] 2) RAN, Radio Access Network. The RAN manages radio resources, transmits user data received through N3 to the UE, and transmits user data from the UE through the N3 interface. The RAN maps QoS traffic between DRB (Dedicated Radio Bearer) and PDU sessions.

[0054] 3) AMF (Access and Mobility Management Function). This function includes registration management, connection management, reachability management, and mobility management. It also performs access authentication and authorization. The AMF is a NAS (Security Attached Function) endpoint that forwards SM NAS information between the UE and the SMF.

[0055] 4) SMF, Session Management Function, includes the following functions: session establishment, modification, and release; UE IP address allocation and management (including optional authorization functions); UP function selection and control; downlink data notification, etc. SMF controls UPF through N4 association. SMF provides UPF with PDR (Packet Inspection Rules), instructing how to inspect user data traffic; and provides FAR (Forwarding Operation Rules), QER (QoS Enforcement Rules), and URR (Usage Reporting Rules), instructing UPF how to perform user data traffic forwarding, QoS processing, and usage reporting on user data traffic detected using PDR.

[0056] 5) UPF, User Plane Functions. This function includes the following: acting as an anchor point for movement within / between Radio Access Types (RATs), packet routing and forwarding, traffic usage reporting, user plane QoS processing, downlink packet buffering, and downlink data notification triggering. GTP-U tunnels are used for the N3 interface between the RAN and UPF. GTP-U tunnels are implemented per PDU session. For downlink traffic, the UPF binds downlink traffic to QoS traffic within the PDU session's GTP-U tunnel using the FAR received from the SMF. For uplink traffic, the RAN transmits user plane traffic to a QoS stream recognized by the UE.

[0057] 6) PCF, Policy Control Function. The PCF provides QoS policy rules to control plane functions for enforcement. The PCF translates AF requests into PCC rules applicable to PDU sessions.

[0058] 7) UDM, Unified Data Management. UDM performs tasks such as generating 3GPP AKA authentication credentials, granting access based on subscription data, managing UE service NF registration (e.g., storing AMF for UE storage services, and SMF for UE PDU session storage services), and subscription management. UDM accesses UDR to retrieve UE subscription data and stores the UE context in the UDR. UDM and UDR can be deployed together.

[0059] NWDAF (Network Data Analysis Function) is a 5GC NF located in the control plane, performing statistical data and machine learning related tasks within 5GS. NWDAF can interact with different entities for various purposes:

[0060] Data is collected based on event subscriptions provided by AMF, SMF, UPF, PCF, UDM, NSACF, AF (directly or via NEF), and OAM;

[0061] In some embodiments, analysis and data collection are performed using DCCF (Data Collection Coordination Function);

[0062] Retrieve information from data repositories (e.g., retrieve UDRs related to users via UDM or retrieve PFD information via NEF (PFDF));

[0063] Collect location information data from the LCS system;

[0064] In some embodiments, information is stored and retrieved from ADRF (Analysis Data Storage Function);

[0065] In some embodiments, data is analyzed and collected from MFAF (Message Transfer Frame Adapter Function);

[0066] Retrieve information about NF (e.g., retrieve NF-related information from NRF);

[0067] Provide analytics to consumers on demand.

[0068] Provides batch data related to the analysis ID.

[0069] Provides information on the accuracy of the analysis ID.

[0070] Provides information about the accuracy of the ML model or an indication of ML model accuracy degradation.

[0071] A single instance or multiple instances of NWDAF can be deployed in a PLMN. NWDAF can contain the following logical functions:

[0072] Analysis Logic Function (AnLF): A logic function in NWDAF used to perform inference, derive analytical information (i.e., derive statistical data and / or predictions based on analytical consumer requests) and expose analytical services (i.e., Nnwdaf_AnalyticsSubscription or Nnwdaf_AnalyticsInfo).

[0073] Model Training Logic Function (MTLF): A logic function in NWDAF used to train machine learning (ML) models and expose new training services (such as providing pre-trained ML models).

[0074] An NWDAF can contain an MTLF or an AnLF, or both of these logical functions.

[0075] The Data Collection Coordination and Function (DCCF) is also an NF on the 5GC control plane. DCCF is responsible for coordinating the collection and distribution of data requested by NF consumers. It prevents data sources from processing multiple subscriptions to the same data and prevents multiple notifications containing the same information from being sent due to miscoordination of data consumer requests.

[0076] DCCF is applicable to:

[0077] NWDAF requests data from a data source (such as for computational analysis).

[0078] NF consumers are analyzed from the NWDAF data source.

[0079] An NF consumer that requests data from an ADRF data source.

[0080] ADRF that receives data from NF data sources.

[0081] Figure 1b This is an architecture diagram of data collection for a network data analysis function provided in one embodiment. The network data analysis function NWDAF collects data from any network element NF through the network function Nnf, so that NWDAF can perform data analysis and model training. Figure 1c As one embodiment, a network data analysis exposure architecture is provided, in which NWDAF provides analysis, statistics, and prediction to other NFs for use by any NF consumer that subscribes to or requests analysis.

[0082] Figure 1dThis diagram illustrates an architecture for data collection coordination and functionality in one embodiment. The Data Collection Coordination and Function (DCCF) coordinates data collection and distribution between arbitrary network elements (NFs) and the Network Data Analysis Function (NWDAF) through the Ndccf and Nnf services. The DCCF registers with the Network Resource Management Function (NRF), Unified Data Management (UDM), or Binding Support Function (BSF). The DCCF interacts with the Message Framework Adapter Function (MFAF) through the Data Management Configuration Service (Nmfaf) to achieve data collection and distribution. The MFAF interacts with any NF through the Nnf service to collect data, and with the NWDAF through the Nmfaf service to collect data. By collecting data from any NF, the NWDAF can perform data analysis and model training. Figure 1e As one embodiment, another network data analytics exposure architecture is provided, in which DCCF helps send analytics and prediction results provided by NWDAF. DCCF interacts with NWDAF through the Nnwdaf service, which is available to any NF consumer that subscribes to or requests analytics.

[0083] Figure 2 This is a flowchart illustrating a model lifecycle as an example. The model lifecycle typically includes: data preprocessing, training / validation, testing, model deployment / sending to other entities / model inference. Model testing is usually repeated until the test results meet deployment requirements; data preprocessing operations may be repeated during model testing.

[0084] In machine learning, dataset preprocessing is a crucial step that directly impacts model training effectiveness and final performance. Data preprocessing encompasses various techniques and methods designed to transform raw data into a format more suitable for machine learning models.

[0085] Here are some of the main functions of data preprocessing:

[0086] Improving data quality: Raw data often contains noise, errors, missing values, and inconsistencies. Data cleaning can correct errors, fill in missing values, and standardize data formats from different sources, thereby improving the overall quality and usability of the data.

[0087] Standardized Format: Machine learning models typically require input data to have a specific format (e.g., numerical data). The data preprocessing stage needs to transform categorical data (such as text labels) into a format that the model can process, such as through one-hot encoding or label encoding.

[0088] Data normalization / standardization: Different features often have different dimensions and distributions, which can affect the model's learning efficiency and final performance. By normalizing (scaling the data to between 0 and 1) or standardizing (transforming the data to have a distribution with a mean of 0 and a standard deviation of 1), we can ensure that the model treats all features equally and speeds up the learning process.

[0089] Reducing dimensionality: Raw data may contain a large number of features, some of which may be redundant or irrelevant to the prediction task. Feature selection and extraction (such as Principal Component Analysis, PCA) can reduce the dimensionality of the data, improving the training efficiency and generalization ability of the model.

[0090] Handling imbalanced data: In some machine learning tasks, the data classes may be extremely imbalanced, which may cause the model to favor the majority class and ignore the minority class. By oversampling the minority class or undersampling the majority class, the dataset can be made more balanced, improving the model's ability to identify the minority class.

[0091] Feature engineering: By creating new features (based on patterns and associations in existing data), the expressive power of the data can be increased, helping the model learn and make predictions better. Feature engineering is one of the key steps in improving the performance of machine learning models.

[0092] Good data preprocessing can not only improve the model's performance on the training set, but also help the model generalize better on unseen test data.

[0093] Data preprocessing typically refers to the preprocessing of training and test sets. Common techniques include:

[0094] 1. Data Cleaning

[0095] Handling missing values: Missing data can be handled by deleting rows / columns containing missing values, filling missing values ​​(e.g., using the mean, median, mode, or other algorithms).

[0096] Remove duplicate records: Delete duplicate rows in the dataset to avoid biasing the model.

[0097] 2. Data Conversion:

[0098] Standardization / Normalization: Normalizing data to a specific range or distribution. Common methods include Z-score standardization and Min-Max normalization.

[0099] Categorical coding: Converting non-numerical categorical variables into numerical data. Common methods include one-hot encoding and label encoding.

[0100] Ordered coding: Applying meaningful numerical coding to ordered categorical data.

[0101] 3. Data dimensionality reduction:

[0102] Feature selection: Select the most relevant features to train the model, reducing computational complexity and the risk of overfitting.

[0103] Feature extraction: Using methods such as principal component analysis (PCA) and linear discriminant analysis (LDA), high-dimensional data is transformed into a low-dimensional space.

[0104] 4. Outlier handling:

[0105] Identify and handle outliers that may adversely affect model training. Handling methods include deletion, replacement, or identification and processing using statistical methods (such as IQR, Z-score).

[0106] 5. Data Augmentation:

[0107] This is particularly common in image and text data, where methods such as rotating, flipping, scaling images, or replacing text with synonyms are used to increase sample diversity.

[0108] 6. Data bucketing:

[0109] Classifying numerical features into different buckets or intervals helps in handling continuous variables or reducing the impact of noise.

[0110] 7. Feature construction (feature engineering):

[0111] Build new features from existing data / features.

[0112] 8. Handling sample imbalance:

[0113] Employ resampling techniques, such as random oversampling (increasing minority class samples), random undersampling (reducing majority class samples), or use synthetic sample generation techniques such as SMOTE.

[0114] Model training:

[0115] Model training is a core part of the machine learning process, involving using datasets to train algorithms so that they can learn the patterns and structures of the data, thereby making accurate predictions or decisions on new and unseen data.

[0116] Commonly used model training methods include: supervised learning, unsupervised learning, semi-supervised learning, self-supervised learning, reinforcement learning, transfer learning, ensemble learning, etc.

[0117] In machine learning, the types of models trained can be divided into two categories: linear and nonlinear.

[0118] Commonly used linear models include: linear regression, logistic regression, and linear discriminant analysis.

[0119] Commonly used nonlinear models include: decision trees, support vector machines, random forests, gradient boosting machines, K-nearest neighbors, feedforward neural networks, convolutional neural networks, recurrent neural networks, long short-term memory networks, and generative adversarial networks.

[0120] Model validation:

[0121] Model validation is a crucial step in machine learning, aiming to confirm the model's effectiveness and generalization ability. Effective model validation not only helps avoid overfitting but also provides reliable estimates of model performance. It can also assist in tuning model hyperparameters.

[0122] Commonly used model validation methods include: hold-out validation, k-fold cross-validation, leave-one-out validation, time series cross-validation, and hierarchical cross-validation.

[0123] Commonly used metrics for model validation:

[0124] Model accuracy, precision, recall, sensitivity, F1 score, area under the ROC curve, confusion matrix, mean squared error, root mean square error, mean absolute error, R^2, coefficient of determination, silhouette coefficient, Davis-Bourdin index.

[0125] Model deployment / sending to other entities / model inference:

[0126] After the model completes training and passes testing, it is sent to other subjects for deployment and inference. Alternatively, it can be deployed and inference performed on its training subject.

[0127] Figure 3 A flowchart of an information transmission method provided in one embodiment, such as Figure 3 As shown, the information transmission method described in this application embodiment is applied to a first communication device, and the method includes S110-S130:

[0128] S110, Send a discovery request to the second communication device.

[0129] The discovery request is used to instruct the second communication device to find a communication device to perform data processing tasks. Data processing can refer to preprocessing data during the model's lifecycle, such as preprocessing the training dataset, preprocessing the test dataset, etc. The discovery request may carry information indicating the type and capability requirements of the communication device needed by the first communication device. Alternatively, the discovery request may not carry any information, and the second communication device may find a matching communication device based on pre-negotiated rules or information.

[0130] The first communication device can generate a discovery request while performing a data processing task. For example, the first communication device may generate a discovery request when it detects that certain requirements are met during data processing; or, the first communication device may generate a discovery request every time it performs data processing; or, the first communication device may generate a discovery request after receiving instruction information from other communication devices before performing data processing, and so on. After generating the discovery request, the first communication device sends the discovery request to the second communication device.

[0131] S120: Receive the first communication device information fed back by the second communication device. The first communication device information is determined based on the model lifecycle capability information of the communication device.

[0132] In this embodiment, the first communication device information can be understood as information describing the communication device, such as the communication device's ID, name, etc., which can be information used to uniquely identify the communication device; the model lifecycle capability information can be understood as information describing the communication device's ability to perform the model lifecycle, such as whether it can support model testing, the time period during which it supports model testing, etc.

[0133] The second communication device pre-stores model lifecycle capability information for different communication devices. Upon receiving a discovery request, the second communication device filters suitable communication devices from the pre-stored model lifecycle capability information. The number of communication devices selected can be one, more, or zero. Based on the filtering results, the second communication device determines the information of the first communication device and feeds it back to the second communication device. The first communication device receives the first communication device information fed back by the second communication device.

[0134] S130. Determine the third communication device based on the information of the first communication device, send a data processing request to the third communication device, and receive the processed data fed back by the third communication device.

[0135] Among them, a data processing request can be understood as a communication request used to instruct a third communication device to perform a data processing task.

[0136] The first communication device determines which communication devices can perform data processing tasks based on its own information, and selects one of these devices as the third communication device. The first communication device generates a data processing request and sends it to the third communication device. This request may include information instructing the third communication device on how to perform the data processing, such as the execution time, processing method, data to be processed, and storage address of the processed data; alternatively, it may not include any information. The data processing request can be generated according to a pre-defined format. Upon receiving the data processing request, the third communication device executes the data processing task. During processing, it can determine the data to be processed based on the data processing request or a pre-defined plan. After completing the data processing, it sends the processed data back to the first communication device. The first communication device receives the processed data from the third communication device. The first communication device can instruct the third communication device which data to process based on the data processing request, other communication requests, or pre-negotiated results.

[0137] When processing data, the first communication device can also perform data processing tasks itself without going through the third communication device.

[0138] The information transmission method provided in this application embodiment involves a first communication device sending a discovery request to a second communication device. Upon receiving the discovery request, the second communication device determines the information of the first communication device based on its model lifecycle capability information and feeds back the information to the first communication device. The first communication device then determines a third communication device based on the information, sends a data processing request to the third communication device, instructs the third communication device to perform data processing, and receives the processed data from the third communication device. By instructing the third communication device to perform data processing, the first communication device solves the problem of data processing tasks failing to execute normally due to the inability of the first communication device to perform data processing. By having the third communication device assist the first communication device in performing data processing, the data processing task is prevented from being forced to terminate or delayed, thereby improving data processing efficiency.

[0139] In some embodiments, the method further includes, before sending a discovery request to the second communication device:

[0140] Receive the first model management request sent by the fourth communication device.

[0141] The first model management request can be understood as an instruction to the fourth communication device to perform operations related to the model lifecycle; for example, performing preprocessing of datasets such as model training and testing, model analysis, subscription, etc.

[0142] After receiving the first model management request from the fourth communication device, the first communication device determines that the fourth communication device needs the model or needs to perform related operations through the model, such as inference. After receiving the first model management request from the fourth communication device, the first communication device sends a discovery request to the second communication device.

[0143] In some embodiments, the first model management request includes one or more of the following information:

[0144] Dataset size requirements;

[0145] Dataset preprocessing instructions;

[0146] Data preprocessing instructions;

[0147] Dataset preprocessing methods;

[0148] Time information.

[0149] The dataset size requirement indicates the size of the dataset to be processed. For example, the dataset size cannot be greater than N, or the dataset size cannot be less than M, or the dataset size is between M and N, etc., where M and N are positive integers. The dataset preprocessing instruction instructs the first communication device to perform dataset processing, specifically preprocessing the data in the dataset. The data preprocessing instruction instructs the first communication device to perform data processing. The dataset preprocessing method indicates the method used to process the dataset, such as data cleaning, data transformation, data dimensionality reduction, outlier handling, data augmentation, data binning, feature construction, and imbalanced sample handling. The dataset preprocessing method may include one or more of the above processing methods, and when multiple processing methods are included, the execution order of these methods may also be included. Time information can be understood as indicating the time of data processing. The fourth communication device can instruct the first communication device to perform data processing according to this requirement by including one or more of the above information in the first model management request.

[0150] In some embodiments, the first model management request may further include: a data preprocessing method; the dataset preprocessing method is used to indicate how the data is processed, such as data cleaning, data transformation, etc. The processing method can refer to the dataset preprocessing method described above, and will not be repeated here.

[0151] In some embodiments, the dataset is at least one of the following:

[0152] Training dataset;

[0153] Test dataset.

[0154] The dataset can be one or more of the training dataset and the test dataset, and the data processing can be performed on at least one of the training dataset and the test dataset.

[0155] Accordingly, the dataset size requirement includes at least one of the training dataset and the test dataset;

[0156] Dataset preprocessing instructions include at least one of training dataset preprocessing instructions and test dataset preprocessing instructions;

[0157] Data preprocessing instructions include at least one of training data preprocessing instructions and test data preprocessing instructions;

[0158] Dataset preprocessing methods include at least one of training dataset preprocessing methods and test dataset preprocessing methods;

[0159] Data preprocessing methods include at least one of training data preprocessing methods and test data preprocessing methods.

[0160] In some embodiments, the first model management request includes at least one of the following:

[0161] Model request;

[0162] Model subscription;

[0163] Analysis request;

[0164] Analytics subscription.

[0165] Among them, the model request can be used to instruct the fourth communication device to request the first communication device to return a model; the model subscription can be used to instruct the fourth communication device to request to subscribe to a model. After subscribing to a model, the fourth communication device will usually receive multiple models. For example, after the first communication device updates the model, it will send the updated model to the first communication device; the analysis request can be used to instruct the fourth communication device to request the first communication device to perform analysis through the model and return the analysis results; the analysis subscription can be used to instruct the fourth communication device to subscribe to the analysis results of the model. After subscribing, the fourth communication device will usually receive multiple analysis results. For example, after the first communication device performs inference analysis through the model, it will send the analysis results to the first communication device.

[0166] In some embodiments, before sending a discovery request to a second communication device, the method further includes:

[0167] The first communication device is determined to satisfy at least one of the following conditions:

[0168] The computing power of the first communication device is insufficient;

[0169] The first communication device does not support the specified data preprocessing method;

[0170] The first communication device lacks data preprocessing capabilities.

[0171] If the computing power of the first communication device is insufficient, it cannot perform the data processing task. The fourth communication device can specify a data preprocessing method. For example, the fourth communication device can instruct the first communication device to use a specified data preprocessing method through a first model management request (i.e., the fourth communication device can specify the data preprocessing method by carrying a dataset preprocessing method or a data preprocessing method in the first model management request), or the fourth communication device can specify the data preprocessing method through prior negotiation with the first communication device, and so on. If the first communication device does not support the specified data preprocessing method, it cannot perform the data processing task. If the first communication device lacks data preprocessing capabilities, it cannot perform the data processing task.

[0172] Before sending a discovery request to the second communication device, the first communication device determines whether it meets at least one of the above conditions. If it does, it determines that other communication devices need to perform data processing tasks, and the first communication device generates a discovery request and sends it to the second communication device.

[0173] In some embodiments, the discovery request includes one or more of the following information:

[0174] Types of communication equipment;

[0175] Data preprocessing instructions;

[0176] Dataset preprocessing instructions;

[0177] Data preprocessing methods;

[0178] Dataset preprocessing methods;

[0179] Time information;

[0180] Information on the computational power required for data preprocessing.

[0181] The time information is used to indicate the execution time of the data processing task.

[0182] The first communication device can instruct the second communication device to find a communication device of a specified type by means of the communication device type; instruct the second communication device to find a communication device that can perform data processing tasks by means of data preprocessing instructions; instruct the second communication device to find a communication device that can perform data processing tasks by means of dataset preprocessing instructions; instruct the second communication device to find a communication device that can perform data processing using this data preprocessing method by means of dataset preprocessing instructions; select a communication device that can perform data processing tasks within this time period by means of time information; or find a communication device with the required computing power by means of the computing power estimated for data preprocessing.

[0183] In some embodiments, a data processing request includes one or more of the following information:

[0184] Data set to be processed;

[0185] Data preprocessing method instructions;

[0186] Time information.

[0187] The dataset to be processed can be used to instruct a third communication device to process the dataset; a data preprocessing method can be used to instruct the third communication device to process the data according to a specific data preprocessing method, that is, to instruct the third communication device to use a particular data preprocessing method for data processing; time information can be used to instruct the third communication device to perform data processing tasks based on the time information, for example, to instruct the third communication device to perform data processing within the time range corresponding to the time information or at the start time.

[0188] In some embodiments, the method further includes:

[0189] Receive dataset labels and storage address information from a third communication device.

[0190] Among them, the dataset label is used to identify the set of processed data. The dataset label can uniquely identify the dataset and distinguish different datasets; the storage address information is used to indicate the address where the processed data is stored.

[0191] After receiving a data processing request, the third communication device processes the data and stores the processed data in its own or other communication devices' storage space, obtaining dataset tags and storage address information, which it then sends back to the first communication device. The first communication device receives the dataset tags and storage address information from the third communication device and can then retrieve the processed data from the corresponding storage space based on these information.

[0192] In some embodiments, model lifecycle capability information includes one or more of the following:

[0193] Dataset preprocessing capabilities;

[0194] Supported dataset preprocessing operations;

[0195] Model training capability;

[0196] Model validation capability;

[0197] Supported model validation methods;

[0198] The time period information corresponding to the capability;

[0199] Model application instruction information, used to indicate the application of the model;

[0200] Available computing power information;

[0201] The time period corresponding to available computing power information.

[0202] The data set preprocessing capability indicates whether the communication device supports data set preprocessing. Supported preprocessing operations include data cleaning, data transformation, dimensionality reduction, outlier handling, data augmentation, data binning, feature construction, and imbalanced sample handling. The model training capability indicates whether the communication device supports model training. The model validation capability indicates whether the communication device supports model validation. Supported model validation methods include at least one of the following: hold-out validation, k-fold cross-validation, leave-one-out validation, time-series cross-validation, and hierarchical cross-validation. The time period information corresponding to the capability indicates the time period during which the communication device supports different capabilities, i.e., which capability the communication device supports in which time period. For example, it supports data set preprocessing in time T1-T2 and model validation in time T3-T4, etc. The model application indication information indicates the application of the model, such as whether the model is used to predict communication congestion or communication quality. The model application indication information can be analysis ID information or other information. Available computing power information is used to indicate the computing power that the communication device can use; the time period corresponding to the available computing power information is used to indicate the available computing power of the communication device in different time periods. For example, the available computing power of the communication device in the time period T5-T6 is X1, and the available computing power of the communication device in the time period T7-T8 is X2.

[0203] In some embodiments, the first communication device is an NWDAF;

[0204] In some embodiments, the third communication device is NWDAF.

[0205] In some embodiments, the second communication device is a network element with network warehousing functionality.

[0206] In some embodiments, the fourth communication device is a core network element, a network management device, or a base station.

[0207] The information transmission method provided in this application embodiment involves a first communication device sending a discovery request to a second communication device. This discovery request can carry relevant information about the communication device. Upon receiving the discovery request, the second communication device determines the information of the first communication device based on its model lifecycle capability information and feeds this information back to the first communication device. The first communication device then determines a third communication device based on this information and sends a data processing request to the third communication device. This data processing request can carry data processing-related information, instructing the third communication device to perform data processing, and receiving the processed data from the third communication device. By instructing the third communication device to perform data processing, the first communication device solves the problem of data processing tasks failing to execute normally due to the first communication device's inability to perform data processing. By having the third communication device assist the first communication device in data processing, the data processing task is prevented from being forcibly terminated or delayed, thus improving data processing efficiency.

[0208] Figure 4 A flowchart of another information transmission method provided in one embodiment, such as Figure 4 As shown, the information transmission method described in this application embodiment is applied to a third communication device, and the method includes S210-S220:

[0209] S210, Receive the data processing request sent by the first communication device.

[0210] S220: Feed back the processed data to the first communication device.

[0211] The third communication device receives a data processing request sent by the first communication device. The data processing request may contain information instructing the third communication device on how to perform data processing, such as the execution time of the data processing task and the dataset to be processed; alternatively, it may not contain any information. Upon receiving the data processing request, the third communication device executes the data processing task according to the instructions in the request. The first communication device may instruct the third communication device, through the data processing request, other communication requests, or pre-negotiated results, to specify which data to process. The third communication device processes the data according to the instructions of the first communication device or the pre-negotiated results. After completing the data processing, the third communication device sends the processed data back to the first communication device.

[0212] The information transmission method provided in this application embodiment involves a third communication device receiving a data processing request sent by a first communication device, assisting the first communication device in performing data processing tasks, and feeding back the processed data to the first communication device. This solves the problem that the data processing task cannot be executed normally because the first communication device is unable to perform data processing. In cases where the first communication device is unable to perform data processing, it can send a data processing request to the third communication device to request the third communication device to assist in data processing, thereby avoiding the forced termination or delay of the data processing task and improving data processing efficiency.

[0213] In some embodiments, a data processing request includes one or more of the following information:

[0214] Data set to be processed;

[0215] Data preprocessing method instructions;

[0216] Time information.

[0217] In some embodiments, before feeding back the processed data to the first communication device, the method further includes:

[0218] Identify the data to be processed;

[0219] Data preprocessing is performed on the data to be processed to obtain the processed data.

[0220] Here, the data to be processed can be understood as data that requires processing. After receiving a data processing request, the third communication device determines the data to be processed based on the information carried in the data processing request or pre-agreed information, and processes the data to be processed according to the information carried in the data processing request or the pre-agreed data preprocessing method to obtain the processed data.

[0221] In some embodiments, determining the data to be processed includes:

[0222] If the data processing request does not include the dataset to be processed or the included dataset does not meet the requirements, data is collected to form the data to be processed.

[0223] If the data processing request does not include the dataset to be processed, the third communication device can collect data within its service area and use the collected data as the data to be processed. Alternatively, if the data processing request includes the dataset to be processed, but the dataset does not meet the requirements, the third communication device can collect data within its service area, supplement the data with the collected data, and use the collected data as the data to be processed. Or, it can use both the collected data and the dataset to be processed in the data processing request as the data to be processed. Pre-set judgment conditions can be used to determine whether the dataset to be processed in the data processing request meets the requirements. For example, if the amount of data in the dataset to be processed is less than a certain threshold, it is determined that the requirements are not met; or if the collection area of ​​the dataset to be processed is not within a preset area, it is determined that the requirements are not met, etc. Multiple conditions can also be set simultaneously for comprehensive judgment.

[0224] In some embodiments, the method further includes:

[0225] Send a dataset storage request to the fifth communication device. The dataset storage request is used to instruct the storage of the processed data.

[0226] The dataset storage request can be understood as a communication request instructing the fifth communication device to store data. The dataset storage request instructs the fifth communication device to store the processed data. After completing data processing and obtaining the processed data, the third communication device can also send a dataset storage request to the fifth communication device, instructing the fifth communication device to store the processed data. The processed data can be sent to the fifth communication device via the dataset storage request, or via other means.

[0227] In some embodiments, the method further includes:

[0228] Receive the dataset labels fed back by the fifth communication device.

[0229] After storing the processed data, the fifth communication device generates a dataset tag to uniquely identify the data and sends the dataset tag back to the third communication device. The third communication device receives the dataset tag from the fifth communication device.

[0230] In some embodiments, the method further includes:

[0231] The dataset label and storage address information are fed back to the first communication device.

[0232] After the third communication device sends the processed data to the fifth communication device for storage, it can determine the storage address information accordingly. For example, the storage address information is the identifier ID of the fifth communication device. The third communication device then feeds back the dataset label and storage address information to the first communication device.

[0233] In some embodiments, the first communication device is an NWDAF;

[0234] In some embodiments, the third communication device is NWDAF.

[0235] In some embodiments, the fifth communication device is a network element with ADRF (Analysis Data Storage Function).

[0236] The information transmission method provided in this application embodiment involves a third communication device receiving a data processing request from a first communication device and assisting the first communication device in performing data processing tasks. During data processing, the third communication device can independently collect data to construct or expand the dataset to be processed and then feed back the processed data to the first communication device. This solves the problem of data processing tasks failing to execute normally due to the first communication device's inability to perform data processing. In cases where the first communication device is unable to perform data processing, it can send a data processing request to the third communication device to request assistance in data processing, thus avoiding forced termination or delay of the data processing task and improving data processing efficiency. Furthermore, the third communication device can send the processed data to a fifth communication device for storage, saving the third communication device's data storage space. Data stored on the fifth communication device can be freely accessed by other communication devices, improving data availability.

[0237] Figure 5 A flowchart of another information transmission method provided in one embodiment, such as Figure 5 As shown, the information transmission method described in this application embodiment is applied to a second communication device, and the method includes S310-S320:

[0238] S310, Receive a discovery request sent by the first communication device;

[0239] S320. Send first communication device information to the first communication device. The first communication device information is determined based on the model lifecycle capability information of the communication device.

[0240] The second communication device pre-stores model lifecycle capability information of different communication devices. After receiving a discovery request, the second communication device filters out suitable communication devices from the pre-stored model lifecycle capability information. At this time, the number of communication devices obtained by filtering can be one or more, or zero. Based on the filtering results, the information of the first communication device is determined and the information of the first communication device is fed back to the first communication device.

[0241] The information transmission method provided in this application embodiment involves a second communication device receiving a discovery request sent by a first communication device. The first communication device instructs the second communication device to determine a communication device capable of performing data processing tasks through the discovery request. After receiving the discovery request, the second communication device determines the information of the first communication device based on the model lifecycle capability information of the communication device and feeds back the information of the first communication device to the first communication device so that the first communication device can determine a third communication device based on the information of the first communication device. This solves the problem that the data processing task cannot be executed normally because the first communication device cannot perform data processing. Under the instruction of the first communication device, the second communication device determines the information of the first communication device capable of performing data processing tasks and feeds it back to the first communication device, assisting the first communication device in performing data processing, avoiding the forced termination or delay of data processing tasks, and thus improving data processing efficiency.

[0242] In some embodiments, the discovery request includes one or more of the following information:

[0243] Types of communication equipment;

[0244] Data preprocessing instructions;

[0245] Dataset preprocessing instructions;

[0246] Data preprocessing methods;

[0247] Dataset preprocessing methods;

[0248] Time information;

[0249] Information on the computational power required for data preprocessing.

[0250] In some embodiments, determining the first communication device information includes:

[0251] Based on the discovery request, the first matching communication device information is filtered from the stored model lifecycle capability information of the communication devices.

[0252] Upon receiving a discovery request, the second communication device determines the information carried by the request and matches it with the stored model lifecycle capability information of each communication device to identify matching communication devices. Based on the information of the matching communication devices, it forms the first communication device information. For example, it may filter communication devices based on computing power, communication device type, data preprocessing method, or dataset preprocessing method, or communication devices capable of performing data processing tasks within a given timeframe, or a combination of these filtering methods to select matching communication devices, and so on.

[0253] In some embodiments, model lifecycle capability information includes one or more of the following:

[0254] Dataset preprocessing capabilities;

[0255] Supported dataset preprocessing operations;

[0256] Model training capability;

[0257] Model validation capability;

[0258] Supported model validation methods;

[0259] The time period information corresponding to the capability;

[0260] Model application instruction information, used to indicate the application of the model;

[0261] Available computing power information;

[0262] The time period corresponding to available computing power information.

[0263] In some embodiments, the first communication device is an NWDAF.

[0264] In some embodiments, the second communication device is a network element with network warehousing functionality.

[0265] The information transmission method provided in this application embodiment involves a first communication device sending a discovery request to a second communication device. The discovery request instructs the second communication device to identify a communication device capable of performing data processing tasks. The discovery request can also carry relevant information about the communication device. Upon receiving the discovery request, the second communication device determines the information of the first communication device based on its model lifecycle capability information and feeds this information back to the first communication device, assisting it in data processing. This solves the problem of data processing tasks failing to execute normally due to the first communication device's inability to perform data processing. Under the instruction of the first communication device, the second communication device identifies and feeds back the information of the first communication device capable of performing data processing tasks, assisting it in data processing and preventing forced termination or delays in data processing tasks, thereby improving data processing efficiency.

[0266] Figure 6 A flowchart of another information transmission method provided in one embodiment, such as Figure 6 As shown, the information transmission method described in this application embodiment is applied to a first communication device, and the method includes S410-S420:

[0267] S410: Receive the second model management request sent by the sixth communication device.

[0268] The second model management request can be understood as an instruction to the first communication device to perform operations related to the model lifecycle; for example, performing any one or more operations such as model training, model validation, model analysis, and model subscription.

[0269] After receiving the second model management request sent by the sixth communication device, the first communication device determines that the sixth communication device needs the model or needs to perform related operations through the model, such as reasoning through the model.

[0270] S420. Process the model.

[0271] The model to be processed is identified and relevant processing is performed. The model to be processed can be indicated through a second model management request, or through a pre-agreed method or other communication request. By analyzing the second model management request, the processing method for the model is determined. The second model management request may carry instructions on how the first communication device should process the model, or the processing method for the model may be determined based on the second model management request. For example, if the second model management request is for model analysis, the model can be trained and validated, and then model analysis can be performed on the validated model to obtain the analysis results, and so on.

[0272] The information transmission method provided in this application embodiment allows a sixth communication device to instruct a first communication device to perform model processing via a second model management request. The sixth communication device instructs the first communication device to perform model processing according to its needs, so that the first communication device processes the model as required and obtains a result that meets the requirements of the sixth communication device, thereby improving model processing efficiency, avoiding situations where the first communication device cannot meet the requirements of the sixth communication device during model processing, and saving resources.

[0273] In some embodiments, processing the model includes at least one of the following:

[0274] Train the model;

[0275] Validate the model.

[0276] In some embodiments, the second model management request includes one or more of the following information:

[0277] Model training methods;

[0278] Model type;

[0279] Model validation methods;

[0280] Model validation metrics;

[0281] Time information;

[0282] Training dataset size requirements;

[0283] Verify the required size of the dataset;

[0284] Requirements for partitioning the training dataset;

[0285] Verify the requirements for splitting the dataset.

[0286] The model training method indicates the training method used, such as supervised learning, unsupervised learning, semi-supervised learning, self-supervised learning, reinforcement learning, transfer learning, and ensemble learning. The model type can be linear or nonlinear. Linear models include linear regression, logistic regression, and linear discriminant analysis, while nonlinear models include decision trees, support vector machines, random forests, gradient boosting machines, K-nearest neighbors, feedforward neural networks, convolutional neural networks, recurrent neural networks, long short-term memory networks, and generative adversarial networks. The model validation method can be one or more of hold-out validation, k-fold cross-validation, leave-one-out validation, time-series cross-validation, and hierarchical cross-validation. The model validation metric can be at least one of the following: accuracy, precision, recall, sensitivity, F1 score, area under the ROC curve, confusion matrix, mean squared error, root mean square error, mean absolute error, R², coefficient of determination, silhouette coefficient, and Davis-Burding index. The time information indicates the time spent on model training and validation. The training dataset size requirement indicates the size of the training dataset. For example, the training dataset size cannot be greater than P, or the dataset size cannot be less than Q, or the dataset size is between Q and P, etc., where P and Q are positive integers. The validation dataset size requirement indicates the size of the validation dataset. The training dataset partitioning requirement indicates the requirements for the training dataset when partitioning it. For example, the proportion of the training dataset in the total dataset is not less than k1, the ratio of the training dataset to the validation dataset is within a certain range, etc. The validation dataset partitioning requirement indicates the requirements for the validation dataset when partitioning it. For example, the proportion of the validation dataset in the total dataset is not less than k2, the ratio of the training dataset to the validation dataset is within a certain range, etc.

[0287] In some embodiments, the second model management request includes at least one of the following:

[0288] Model request;

[0289] Model subscription;

[0290] Analysis request;

[0291] Analytics subscription.

[0292] In some embodiments, the method further includes:

[0293] Send at least one of the following data to the sixth communication device:

[0294] Model;

[0295] Analysis results;

[0296] Verification method;

[0297] Verification results.

[0298] After processing the model, the first communication device can also send back at least one of the following data to the sixth communication device: model, analysis results, verification method, and verification results. For example, if the sixth communication device sends a second model management request to request the model, the first communication device can send back the model; if the sixth communication device sends a second model management request to request the analysis results, the first communication device will obtain the analysis results after model inference analysis and send them back to the sixth communication device. The verification results can be model verification indicators and their corresponding numerical values.

[0299] In some embodiments, the sixth communication device may be the fourth communication device; that is, the fourth communication device may send a second model management request to the first communication device as the sixth communication device.

[0300] In some embodiments, the sixth communication device is a core network element, a network management device, or a base station.

[0301] In some embodiments, the first communication device is an NWDAF.

[0302] The information transmission method provided in this application embodiment allows a sixth communication device to instruct a first communication device to perform model processing via a second model management request. The sixth communication device instructs the first communication device to perform model processing by carrying one or more types of information in the second model management request, so that the first communication device processes the model as required and obtains a result that meets the requirements of the sixth communication device, thereby improving model processing efficiency, avoiding situations where the first communication device cannot meet the requirements of the sixth communication device during model processing, and saving resources.

[0303] Figure 7 A flowchart of another information transmission method provided in one embodiment, such as Figure 7 As shown, the information transmission method described in this application embodiment is applied to a second communication device, and the method includes S510-S520:

[0304] S510. Receive model lifecycle capability information sent by the seventh communication device, wherein the model lifecycle capability information is used to indicate the model lifecycle capability of the seventh communication device;

[0305] S520, storage model lifecycle capability information.

[0306] The second communication device receives the model generation cycle capability information sent by the seventh communication device and stores the received model lifecycle capability information.

[0307] The second communication device can receive and store model generation cycle capability information sent by one or more seventh communication devices, so that during data processing, it can find a matching communication device to perform data processing and other tasks based on the model generation cycle capability information.

[0308] The information transmission method provided in this application embodiment involves a second communication device receiving model lifecycle capability information sent by a seventh communication device. The model lifecycle capability information is used to indicate the model lifecycle capability of the seventh communication device. When the first communication device is unable to perform data processing, the second communication device assists the first communication device in determining communication device information that can perform data processing tasks through the model lifecycle capability information, thereby avoiding the forced termination or delay of the first communication device's data processing tasks and improving data processing efficiency.

[0309] After receiving the model lifecycle capability information, the second communication device stores the model lifecycle capability information. If the saving is successful, a saving success indication is generated; if the saving fails, a saving failure indication is generated.

[0310] In some embodiments, receiving model lifecycle capability information sent by a seventh communication device includes:

[0311] Receive model lifecycle capability information registered or updated by the seventh communication device.

[0312] In some embodiments, a successful saving indication or a failed saving indication is sent back to the seventh communication device;

[0313] The "Save Success" indicator indicates that the registration or update of model lifecycle capability information was successful, while the "Save Failure" indicator indicates that the registration or update of model lifecycle capability information failed.

[0314] In some embodiments, the save success indicator can also be used to indicate that the second communication device has successfully saved the model lifecycle capability information, and the save failure indicator can also be used to indicate that the second communication device has failed to save the model lifecycle capability information.

[0315] In some embodiments, model lifecycle capability information includes one or more of the following:

[0316] Dataset preprocessing capabilities;

[0317] Supported dataset preprocessing operations;

[0318] Model training capability;

[0319] Model validation capability;

[0320] Supported model validation methods;

[0321] The time period information corresponding to the capability;

[0322] Model application instruction information, used to indicate the application of the model;

[0323] Available computing power information;

[0324] The time period corresponding to available computing power information.

[0325] In some embodiments, the seventh communication device is either the first communication device or the third communication device.

[0326] In some embodiments, the second communication device is a network element with network warehousing functionality.

[0327] In some embodiments, the seventh communication device is NWDAF.

[0328] The information transmission method provided in this application embodiment involves a seventh communication device sending model lifecycle capability information to a second communication device. The model lifecycle capability information indicates the model lifecycle capability of the seventh communication device, so that when the first communication device is unable to perform data processing, the second communication device can assist the first communication device in determining communication device information that can perform data processing tasks, thereby avoiding the forced termination or delay of the first communication device's data processing tasks and improving data processing efficiency.

[0329] Figure 8 A flowchart of another information transmission method provided in one embodiment, such as Figure 8 As shown, the information transmission method described in this application embodiment is applied to a seventh communication device, and the method includes S610-S620:

[0330] S610. Send model lifecycle capability information to the second communication device. The model lifecycle capability information is used to indicate the model lifecycle capability of the seventh communication device.

[0331] The seventh communication device determines its own model lifecycle capability information and sends the model lifecycle capability information to the second communication device. The model lifecycle capability information indicates the model lifecycle capability of the seventh communication device. The second communication device can save the model lifecycle capability information of different communication devices so as to match the appropriate communication device for the first communication device when performing data processing.

[0332] The information transmission method provided in this application embodiment involves a seventh communication device sending model lifecycle capability information to a second communication device. The model lifecycle capability information indicates the model lifecycle capability of the seventh communication device, so that when the first communication device is unable to perform data processing, the second communication device can assist the first communication device in determining communication device information that can perform data processing and other tasks, thereby avoiding the forced termination or delay of the first communication device's data processing tasks and improving data processing efficiency.

[0333] Send model lifecycle capability information to the second communication device, including:

[0334] Register or update model lifecycle capability information with the second communication device.

[0335] The seventh communication device can register or update model lifecycle capability information with the second communication device. That is, the seventh communication device registers the model lifecycle capability information when it sends the model lifecycle capability information to the second communication device for the first time; and updates the model lifecycle capability information when it sends the model lifecycle capability information again in subsequent communications.

[0336] In some embodiments, the method further includes:

[0337] Receive a save success indication or save failure indication from the second communication device;

[0338] The "Save Success" indicator indicates that the registration or update of model lifecycle capability information was successful, while the "Save Failure" indicator indicates that the registration or update of model lifecycle capability information failed.

[0339] The seventh communication device can resend the model lifecycle capability information in the event of a saving failure, thus preventing the failure to save the model lifecycle capability information from affecting the normal execution of subsequent tasks.

[0340] In some embodiments, the save success indicator can also be used to indicate that the second communication device has successfully saved the model lifecycle capability information, and the save failure indicator can also be used to indicate that the second communication device has failed to save the model lifecycle capability information.

[0341] In some embodiments, model lifecycle capability information includes one or more of the following:

[0342] Dataset preprocessing capabilities;

[0343] Supported dataset preprocessing operations;

[0344] Model training capability;

[0345] Model validation capability;

[0346] Supported model validation methods;

[0347] The time period information corresponding to the capability;

[0348] Model application instruction information, used to indicate the application of the model;

[0349] Available computing power information;

[0350] The time period corresponding to available computing power information.

[0351] In some embodiments, the seventh communication device is either the first communication device or the third communication device.

[0352] The seventh communication device can be either the first or the third communication device. That is, both the first and third communication devices can act as the seventh communication device to send model lifecycle capability information to the second communication device, thereby enabling the registration or updating of model lifecycle capability information.

[0353] In some embodiments, the seventh communication device is NWDAF.

[0354] Any NWDAF can act as a seventh communication device to send its own model lifecycle capability information to the second communication device, and register or update its model lifecycle capability information with the second communication device.

[0355] In some embodiments, the second communication device is a network element with network warehousing functionality.

[0356] In some embodiments, the first communication device sends model lifecycle capability information to the second communication device, the model lifecycle capability information being used to indicate the model lifecycle capability of the first communication device.

[0357] In some embodiments, the first communication device sends model lifecycle capability information to the second communication device, including registering or updating the model lifecycle capability information with the second communication device.

[0358] In some embodiments, the first communication device receives a save success indication or a save failure indication fed back by the second communication device;

[0359] The "Save Success" indicator indicates that the registration or update of model lifecycle capability information was successful, while the "Save Failure" indicator indicates that the registration or update of model lifecycle capability information failed.

[0360] In some embodiments, model lifecycle capability information includes one or more of the following:

[0361] Dataset preprocessing capabilities;

[0362] Supported dataset preprocessing operations;

[0363] Model training capability;

[0364] Model validation capability;

[0365] Supported model validation methods;

[0366] The time period information corresponding to the capability;

[0367] Model application instruction information, used to indicate the application of the model;

[0368] Available computing power information;

[0369] The time period corresponding to available computing power information.

[0370] In some embodiments, the third communication device sends model lifecycle capability information to the second communication device, the model lifecycle capability information being used to indicate the model lifecycle capability of the third communication device.

[0371] In some embodiments, the third communication device sends model lifecycle capability information to the second communication device, including registering or updating the model lifecycle capability information with the second communication device.

[0372] In some embodiments, the third communication device receives a save success indication or a save failure indication fed back by the second communication device;

[0373] The "Save Success" indicator indicates that the registration or update of model lifecycle capability information was successful, while the "Save Failure" indicator indicates that the registration or update of model lifecycle capability information failed.

[0374] In some embodiments, model lifecycle capability information includes one or more of the following:

[0375] Dataset preprocessing capabilities;

[0376] Supported dataset preprocessing operations;

[0377] Model training capability;

[0378] Model validation capability;

[0379] Supported model validation methods;

[0380] The time period information corresponding to the capability;

[0381] Model application instruction information, used to indicate the application of the model;

[0382] Available computing power information;

[0383] The time period corresponding to available computing power information.

[0384] The information transmission method provided in this application embodiment involves a seventh communication device sending model lifecycle capability information to a second communication device. This model lifecycle capability information indicates the seventh communication device's model lifecycle capabilities, enabling the second communication device to assist the first communication device in determining available communication devices to perform data processing tasks when the first communication device is unable to do so. This avoids forced termination or delay of the first communication device's data processing tasks, thereby improving data processing efficiency. Simultaneously, the seventh communication device determines whether the model lifecycle capability information has been successfully saved based on a received save success indication or save failure indication.

[0385] Figure 9 A flowchart of another information transmission method provided in one embodiment, such as Figure 9 As shown, the information transmission method described in this application embodiment is applied to a second communication device, and the method includes S710-S730:

[0386] S710 receives a machine learning task sent by an eighth communication device, the machine learning task including the task's prediction capability.

[0387] Here, machine learning tasks can be understood as AI tasks; estimated computing power can be understood as the computing power required to perform this machine learning task, which can be obtained by analyzing and estimating the task.

[0388] The eighth communication device can generate machine learning tasks under conditions such as needing to acquire models, perform inference analysis through models, or meet other conditions. The machine learning task includes a preliminary estimate of the computing power required for the task, which is used to indicate the computing power required for this machine learning task.

[0389] S720. Determine the second communication device information based on the estimated capabilities in the machine learning task and the pre-stored model lifecycle capability information of the communication device.

[0390] The second communication device information can be understood as information describing the communication device, such as the communication device's ID, name, etc., which can be information used to uniquely identify the communication device.

[0391] After receiving a machine learning task, the second communication device determines the predicted estimation capability corresponding to the task, matches the predicted estimation capability with the model lifecycle capability information of one or more pre-stored communication devices, and identifies the communication devices that can execute the machine learning task, thus forming second communication device information. The number of communication devices executing the machine learning task can be one or more, meaning the second communication device information includes one or more communication devices.

[0392] S730: Feed back the information from the second communication device to the eighth communication device.

[0393] The second communication device information is fed back to the eighth communication device, instructing the eighth communication device to perform corresponding operations through the communication device in the second communication device information, such as model training, model testing, data preprocessing, etc.

[0394] During the process of matching the estimated capability with the pre-stored model lifecycle capability information of the communication device, the second communication device may fail to find a matching communication device. In this case, the information of the second communication device can be empty, indicating that there is no matching communication device that can perform the machine learning task.

[0395] The information transmission method provided in this application embodiment allows an eighth communication device to send a machine learning task to a second communication device. The second communication device then searches for a suitable communication device to execute the machine learning task based on the predicted capability in the machine learning task and the model lifecycle capability information, and provides feedback to the eighth communication device. This assists the eighth communication device in selecting a suitable communication device to execute the machine learning task, thus avoiding situations where the eighth communication device selects an unsuitable communication device, resulting in the machine learning task failing to execute normally or wasting communication device resources.

[0396] In some embodiments, the machine learning task further includes at least one of the following information:

[0397] Analyze the request;

[0398] Model training request;

[0399] Model validation request;

[0400] Model test request;

[0401] Dataset preprocessing request;

[0402] Time information.

[0403] Among them, the analysis acquisition request is used to instruct the communication device to generate corresponding analysis results through model training and inference; the model training request is used to request the communication device to perform model training; the model verification request is used to request the communication device to perform model verification; the model testing request is used to request the communication device to perform model testing; the dataset preprocessing request is used to request the communication device to preprocess the dataset; and the time information is used to indicate the time for the communication device to execute each task or request, for example, performing model testing at time t1-t2, performing model verification at time t3-t4, and so on.

[0404] When determining the information of the second communication device, the second communication device can comprehensively consider the above information and, together with the model lifecycle capability information of the communication device, select suitable communication devices to form the information of the second communication device.

[0405] In some embodiments, the second communication device is a network element with network warehousing functionality.

[0406] In some embodiments, the eighth communication device is the fourth communication device; that is, the fourth communication device can send machine learning tasks to the second communication device as the eighth communication device, and receive the second communication device information fed back by the second communication device; the eighth communication device selects the first communication device according to the second communication device information, and sends a first model management request to the first communication device; after receiving the first model management request, the first communication device can process the data according to the method provided in the embodiments of this application, etc.

[0407] In some embodiments, the eighth communication device is a core network element, a network management device, or a base station.

[0408] The information transmission method provided in this application embodiment allows an eighth communication device to send a machine learning task to a second communication device. The second communication device can filter and match model lifecycle capability information based on one or more types of information contained in the machine learning task, find a suitable communication device to execute the machine learning task, and provide feedback to the eighth communication device. This assists the eighth communication device in selecting a suitable communication device to execute the machine learning task, avoiding situations where the eighth communication device selects an unsuitable communication device, resulting in the machine learning task failing to execute normally or wasting communication device resources.

[0409] The information transmission process is illustrated through the following examples:

[0410] Example 1

[0411] The information transmission process is illustrated by the example of a seventh communication device registering model lifecycle capability information with a second communication device; wherein the seventh communication device can be an NWDAF and the second communication device can be an NRF. The first and third communication devices can act as the seventh communication device to register model lifecycle capability information with the second communication device; other communication devices can also act as the seventh communication device to register model lifecycle capability information with the second communication device.

[0412] Figure 10 An interactive example diagram of registering model lifecycle capability information is provided, including the following steps:

[0413] 1. The seventh communication device registers its model lifecycle capability information with the second communication device. The model lifecycle capability information includes AI model lifecycle-related capabilities, which are used to indicate the model lifecycle capabilities of the seventh communication device.

[0414] Model lifecycle capability information includes at least one of the following:

[0415] Dataset preprocessing capabilities;

[0416] Supported dataset preprocessing operations;

[0417] Model training capability;

[0418] Model validation capability;

[0419] Supported model validation methods;

[0420] The time period information corresponding to the capability;

[0421] Model application instruction information, used to indicate the application of the model;

[0422] Available computing power information;

[0423] The time period corresponding to available computing power information.

[0424] All of the above information can be included in the NF profile of the network element.

[0425] 2. The second communication device stores model lifecycle capability information.

[0426] 3. The second communication device returns a save success indication or a save failure indication.

[0427] The "Save Success" indicator indicates that the registration of the model lifecycle capability information was successful, while the "Save Failure" indicator indicates that the registration of the model lifecycle capability information failed.

[0428] Example 2

[0429] The information transmission process is illustrated by the example of a seventh communication device updating model lifecycle capability information to a second communication device; wherein the seventh communication device can be an NWDAF and the second communication device can be an NRF. The first and third communication devices can act as the seventh communication device to update model lifecycle capability information to the second communication device; other communication devices can also act as the seventh communication device to update model lifecycle capability information to the second communication device.

[0430] Figure 11 An interactive example diagram for updating model lifecycle capability information is provided, including the following steps:

[0431] 1. The seventh communication device updates its model lifecycle capabilities to the second communication device, wherein the model lifecycle capability information includes AI model lifecycle-related capabilities, which are used to indicate the model lifecycle capabilities of the fifth communication device.

[0432] Model lifecycle capability information includes at least one of the following:

[0433] Dataset preprocessing capabilities;

[0434] Supported dataset preprocessing operations;

[0435] Model training capability;

[0436] Model validation capability;

[0437] Supported model validation methods;

[0438] The time period information corresponding to the capability;

[0439] Model application instruction information, used to indicate the application of the model;

[0440] Available computing power information;

[0441] The time period corresponding to available computing power information.

[0442] All of the above information is contained in the NF profile of the network element.

[0443] 2. The second communication device stores model lifecycle capability information.

[0444] 3. The second communication device returns a save success indication or a save failure indication.

[0445] The "Save Success" indicator indicates that the update of the model lifecycle capability information was successful, while the "Save Failure" indicator indicates that the update of the model lifecycle capability information failed.

[0446] Example 3

[0447] Taking the interaction of the first, second, third, fourth, and fifth communication devices as an example, the information transmission process is explained; wherein the first and third communication devices can be NWDAF, the second communication device can be NRF, the fourth communication device is the core network element 5GC, the network management device OAM or the base station, and the fifth communication device is ADRF.

[0448] Figure 12 An interactive example diagram of data preprocessing is provided, including the following steps:

[0449] 1. The fourth communication device initiates a first model management request to the first communication device. The first model management request may include at least one of the following: training dataset size requirement, test dataset size requirement, dataset preprocessing instruction, data preprocessing instruction, dataset preprocessing method, and time information.

[0450] 2. The first communication device performs dataset preprocessing itself, or the first communication device decides that it needs to find another communication device to perform dataset preprocessing.

[0451] The first communication device decides that other communication devices are needed for data processing if at least one of the following conditions is met: the first communication device has insufficient computing power, the first communication device does not support the specified data preprocessing method, or the first communication device does not have data preprocessing capabilities.

[0452] 3. The first communication device sends a discovery request to the second communication device.

[0453] The discovery request contains at least one of the following information: communication device type, data preprocessing instruction, dataset preprocessing instruction, data preprocessing method, dataset preprocessing method, time information, and estimated computing power required for data preprocessing.

[0454] 4. The second communication device matches the information registered / updated by the communication device and returns the information of the first communication device; the information of the first communication device includes a set of NWDAF or other subject information that supports data preprocessing.

[0455] 5. The first communication device determines the third communication device based on the returned information and sends a data processing request to the third communication device. This data processing request may carry at least one of the following information: the dataset to be processed, the data preprocessing method instruction, and time information.

[0456] 6. If the first communication device does not carry the dataset to be processed or the dataset to be processed does not meet the requirements, the third communication device can collect data within its own service area to supplement or construct the dataset and form the data to be processed.

[0457] 7. The third communication device performs dataset preprocessing according to the instructions of the first communication device.

[0458] 8a. The third communication device returns the processed dataset to the first communication device.

[0459] 8b. The third communication device initiates a request to the fifth communication device to store the processed dataset.

[0460] 8c. The fifth communication device returns the dataset labels.

[0461] 8d. The third communication device returns the dataset label and storage address information to the first communication device.

[0462] The storage address information can be ADRFID.

[0463] 9. The first communication device performs subsequent tasks such as model training, model inference, model validation, and model testing based on the preprocessed dataset.

[0464] If preprocessing of the test dataset is required during model testing, steps 2-8 can be repeated.

[0465] 10. The first communication device sends at least one of the model inference results and the trained model to the fourth communication device.

[0466] It is important to emphasize that the dataset in this process can be either a training dataset or a test dataset.

[0467] Example 4

[0468] Taking the interaction between the first communication device and the sixth communication device as an example, the information transmission process is explained; wherein, the first communication device can be NWDAF, the sixth communication device is the core network element 5GC, the network management device OAM or the base station, and the fourth communication device can act as the sixth communication device to interact with the first communication device to achieve the following functions.

[0469] Figure 13 An interactive example diagram for model training / validation is provided, including the following steps:

[0470] 1. The sixth communication device sends a second model management request to the first communication device. The second model management request may contain at least one of the following information: model training method, model type, model validation method, model validation metric, time information, size requirement of training dataset, size requirement of validation dataset, partitioning requirement of training dataset, and partitioning requirement of validation dataset.

[0471] 2. The first communication device can process the dataset itself, or it can choose another communication device to process the data.

[0472] For example, the first communication device can select a third communication device to process the dataset through the second communication device.

[0473] 3. The first communication device performs model training / validation according to the model type and model training method required by the sixth communication device.

[0474] 4. The first communication device performs model testing; the model testing can be performed by the first communication device or another communication device can be selected for model testing.

[0475] For example, the first communication device can select a third communication device through the second communication device to perform model testing.

[0476] 5. If the sixth communication device subscribes to / requests a model, skip this step. If the sixth communication device subscribes to / requests an analysis result, the first communication device performs model inference to obtain the analysis result.

[0477] 6. The first communication device returns at least one of the following data to the sixth communication device: model, analysis results, verification method, and verification results; wherein, the verification results are the model verification indicators and the corresponding values ​​of the indicators in step 1.

[0478] Example 5

[0479] Taking the interaction between the first communication device, the second communication device, and the eighth communication device as an example, the information transmission process is explained; wherein, the first communication device can be an NWDAF, the second communication device is an NRF, the eighth communication device is a core network element 5GC, a network management device OAM, or a base station, and the fourth communication device can act as the eighth communication device to interact with the first and second communication devices to achieve the following functions.

[0480] Figure 14 An interactive example diagram for matching communication devices is provided, including the following steps:

[0481] 1. The eighth communication device sends a machine learning task to the second communication device.

[0482] Machine learning tasks include at least one of the following: analysis acquisition requests (e.g., Analytics ID), model training requests, model validation requests, model testing requests, dataset preprocessing requests, time information, and task prediction capabilities.

[0483] 2. The second communication device matches one or more suitable communication devices according to the machine learning task, forms the second communication device information, and returns it to the eighth communication device.

[0484] For example, the second communication device matches one or more suitable NWDAFs based on the estimated capabilities of the task.

[0485] 3. The eighth communication device sends a request to one or more communication devices based on the information returned by one or more communication devices. For example, a request is sent to the first communication device.

[0486] This request can be at least one of the following: model training request, model validation request, model testing request, dataset preprocessing request, analysis request, analysis subscription, etc.

[0487] 4. The first communication device executes the corresponding task according to the request.

[0488] 5. The first communication device returns the execution result to the eighth communication device.

[0489] The result of the execution can be at least one of the following: the trained model, the analysis results, the validation results, the test results, or the processed dataset.

[0490] Figure 15This is a schematic diagram of the structure of an information transmission device provided in one embodiment. The device is applied to a first communication device, such as... Figure 15 As shown, the device includes: a discovery request sending module 810, a first communication device information receiving module 820, and a data processing request sending module 830.

[0491] The discovery request sending module 810 is used to send a discovery request to the second communication device;

[0492] The first communication device information receiving module 820 is used to receive the first communication device information fed back by the second communication device, wherein the first communication device information is determined based on the communication device's model lifecycle capability information.

[0493] The data processing request sending module 830 is used to determine a third communication device based on the first communication device information, send a data processing request to the third communication device, and receive the processed data fed back by the third communication device.

[0494] The information transmission device provided in this application embodiment involves a first communication device sending a discovery request to a second communication device. Upon receiving the discovery request, the second communication device determines the information of the first communication device based on its model lifecycle capability information and feeds back the information to the first communication device. The first communication device then determines a third communication device based on the information, sends a data processing request to the third communication device, instructs the third communication device to perform data processing, and receives the processed data from the third communication device. By instructing the third communication device to perform data processing, the first communication device solves the problem of data processing tasks failing to execute normally due to the first communication device's inability to perform data processing. By having the third communication device assist the first communication device in performing data processing, the data processing task is prevented from being forced to terminate or delayed, thereby improving data processing efficiency.

[0495] In some embodiments, the device further includes:

[0496] The first management request receiving module is used to receive the first model management request sent by the fourth communication device before sending the discovery request to the second communication device.

[0497] In some embodiments, the first model management request includes one or more of the following information:

[0498] Dataset size requirements;

[0499] Dataset preprocessing instructions;

[0500] Data preprocessing instructions;

[0501] Dataset preprocessing methods;

[0502] Time information.

[0503] In some embodiments, the dataset is at least one of the following:

[0504] Training dataset;

[0505] Test dataset.

[0506] In some embodiments, the first model management request includes at least one of the following:

[0507] Model request;

[0508] Model subscription;

[0509] Analysis request;

[0510] Analytics subscription.

[0511] In some embodiments, the device is further configured to: determine, before sending the discovery request to the second communication device, that the first communication device satisfies at least one of the following conditions:

[0512] The computing power of the first communication device is insufficient;

[0513] The first communication device does not support the specified data preprocessing method;

[0514] The first communication device does not have data preprocessing capabilities.

[0515] In some embodiments, the discovery request includes one or more of the following information:

[0516] Types of communication equipment;

[0517] Data preprocessing instructions;

[0518] Dataset preprocessing instructions;

[0519] Data preprocessing methods;

[0520] Dataset preprocessing methods;

[0521] Time information;

[0522] Information on the computational power required for data preprocessing.

[0523] In some embodiments, the data processing request includes one or more of the following information:

[0524] Data set to be processed;

[0525] Data preprocessing method instructions;

[0526] Time information.

[0527] In some embodiments, the device further includes:

[0528] The storage address receiving module is used to receive the dataset tag and storage address information fed back by the third communication device.

[0529] In some embodiments, the model lifecycle capability information includes one or more of the following:

[0530] Dataset preprocessing capabilities;

[0531] Supported dataset preprocessing operations;

[0532] Model training capability;

[0533] Model validation capability;

[0534] Supported model validation methods;

[0535] The time period information corresponding to the capability;

[0536] Model application instruction information, used to indicate the application of the model;

[0537] Available computing power information;

[0538] The time period corresponding to available computing power information.

[0539] The information transmission device proposed in this embodiment belongs to the same inventive concept as the information transmission method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in any of the above embodiments. Furthermore, this embodiment has the same beneficial effects as performing the information transmission method.

[0540] Figure 16 This is a schematic diagram of another information transmission device provided in one embodiment. This device is applied to a third communication device, such as... Figure 16 As shown, the device includes: a data processing request receiving module 910 and a data feedback module 920;

[0541] The data processing request receiving module 910 is used to receive data processing requests sent by the first communication device.

[0542] The data feedback module 920 is used to feed back the processed data to the first communication device.

[0543] The information transmission device provided in this application embodiment allows a third communication device to receive a data processing request sent by a first communication device, assist the first communication device in performing data processing tasks, and provide feedback of processed data to the first communication device. This solves the problem that the data processing task cannot be executed normally because the first communication device is unable to perform data processing. In cases where the first communication device is unable to perform data processing, it can send a data processing request to the third communication device to request the third communication device to assist in data processing, thereby avoiding the forced termination or delay of the data processing task and improving data processing efficiency.

[0544] In some embodiments, the data processing request includes one or more of the following information:

[0545] Data set to be processed;

[0546] Data preprocessing method instructions;

[0547] Time information.

[0548] In some embodiments, the device further includes:

[0549] The data to be processed determination module is used to determine the data to be processed before feeding back the processed data to the first communication device;

[0550] The data preprocessing module is used to preprocess the data to be processed to obtain the processed data.

[0551] In some embodiments, the data to be processed determination module is specifically used to: collect data and form data to be processed when the data processing request does not carry a dataset to be processed or the dataset to be processed does not meet the requirements.

[0552] In some embodiments, the device further includes:

[0553] The storage request sending module is used to send a dataset storage request to the fifth communication device, wherein the dataset storage request is used to instruct the storage of processed data.

[0554] In some embodiments, the device further includes:

[0555] The tag receiving module is used to receive the dataset tags fed back by the fifth communication device.

[0556] In some embodiments, the device further includes:

[0557] The storage address sending module is used to send back the dataset label and storage address information to the first communication device.

[0558] The information transmission device proposed in this embodiment belongs to the same inventive concept as the information transmission method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in any of the above embodiments. Furthermore, this embodiment has the same beneficial effects as performing the information transmission method.

[0559] Figure 17 This is a schematic diagram of another information transmission device provided in one embodiment. This device is applied to a first communication device, such as... Figure 17 As shown, the device includes: a second management request receiving module 1010 and a model processing module 1020;

[0560] The second management request receiving module 1010 is used to receive the second model management request sent by the sixth communication device;

[0561] The model processing module 1020 is used to process the model.

[0562] The information transmission device provided in this application embodiment allows the sixth communication device to instruct the first communication device to perform model processing via a second model management request. The sixth communication device instructs the first communication device to perform model processing according to its needs, so that the first communication device processes the model as required and obtains a result that meets the requirements of the sixth communication device, thereby improving model processing efficiency, avoiding situations where the first communication device cannot meet the requirements of the sixth communication device during model processing, and saving resources.

[0563] In some embodiments, processing the model includes at least one of the following:

[0564] Train the model;

[0565] Validate the model.

[0566] In some embodiments, the second model management request includes one or more of the following information:

[0567] Model training methods;

[0568] Model type;

[0569] Model validation methods;

[0570] Model validation metrics;

[0571] Time information;

[0572] Training dataset size requirements;

[0573] Verify the required size of the dataset;

[0574] Requirements for partitioning the training dataset;

[0575] Verify the requirements for splitting the dataset.

[0576] In some embodiments, the second model management request includes at least one of the following:

[0577] Model request;

[0578] Model subscription;

[0579] Analysis request;

[0580] Analytics subscription.

[0581] In some embodiments, the device is further configured to: send at least one of the following data to the sixth communication device:

[0582] Model;

[0583] Analysis results;

[0584] Verification method;

[0585] Verification results.

[0586] The information transmission device proposed in this embodiment belongs to the same inventive concept as the information transmission method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in any of the above embodiments. Furthermore, this embodiment has the same beneficial effects as performing the information transmission method.

[0587] Figure 18 This is a schematic diagram of another information transmission device provided in one embodiment. This device is applied to a second communication device, such as... Figure 18 As shown, the device includes a model capability information receiving module 1110 and a model capability information storage module 1120.

[0588] The model capability information receiving module 1110 is used to receive model lifecycle capability information sent by the seventh communication device, wherein the model lifecycle capability information is used to indicate the model lifecycle capability of the seventh communication device.

[0589] The model capability information storage module 1120 is used to store the model lifecycle capability information.

[0590] The information transmission device provided in this application embodiment includes a second communication device receiving model lifecycle capability information sent by a seventh communication device. The model lifecycle capability information is used to indicate the model lifecycle capability of the seventh communication device. When the first communication device is unable to perform data processing, the second communication device assists the first communication device in determining communication device information that can perform data processing tasks through the model lifecycle capability information, thereby avoiding the forced termination or delay of the data processing tasks of the first communication device and improving data processing efficiency.

[0591] In some embodiments, the model capability information receiving module 1110 is specifically used to: receive model lifecycle capability information registered or updated by the seventh communication device.

[0592] In some embodiments, the device further includes:

[0593] The save instruction information feedback module is used to send a save success instruction or a save failure instruction to the seventh communication device;

[0594] The "save success" indicator indicates that the registration or update of model lifecycle capability information was successful, while the "save failure" indicator indicates that the registration or update of model lifecycle capability information failed.

[0595] In some embodiments, the model lifecycle capability information includes one or more of the following:

[0596] Dataset preprocessing capabilities;

[0597] Supported dataset preprocessing operations;

[0598] Model training capability;

[0599] Model validation capability;

[0600] Supported model validation methods;

[0601] The time period information corresponding to the capability;

[0602] Model application instruction information, used to indicate the application of the model;

[0603] Available computing power information;

[0604] The time period corresponding to available computing power information.

[0605] In some embodiments, the seventh communication device is either the first communication device or the third communication device.

[0606] The information transmission device proposed in this embodiment belongs to the same inventive concept as the information transmission method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in any of the above embodiments. Furthermore, this embodiment has the same beneficial effects as performing the information transmission method.

[0607] Figure 19 This is a schematic diagram of another information transmission device provided in one embodiment. This device is applied to a second communication device, such as... Figure 19 As shown, the device includes: a task receiving module 1210, a second communication device information determination module 1220, and a second communication device information feedback module 1230.

[0608] The task receiving module 1210 is used to receive a machine learning task sent by the eighth communication device, wherein the machine learning task includes the task's estimated power.

[0609] The second communication device information determination module 1220 is used to determine the second communication device information based on the estimated capability in the machine learning task and the pre-stored model lifecycle capability information of the communication device.

[0610] The second communication device information feedback module 1230 is used to feed back the second communication device information to the eighth communication device.

[0611] The information transmission device provided in this application embodiment allows the eighth communication device to send a machine learning task to the second communication device. The second communication device finds a suitable communication device to execute the machine learning task based on the predicted capability in the machine learning task and the model lifecycle capability information, and feeds back to the eighth communication device. This assists the eighth communication device in selecting a suitable communication device to execute the machine learning task, avoiding situations where the eighth communication device selects an unsuitable communication device, resulting in the machine learning task failing to execute normally or wasting communication device resources.

[0612] In some embodiments, the machine learning task further includes at least one of the following information:

[0613] Analyze the request;

[0614] Model training request;

[0615] Model validation request;

[0616] Model test request;

[0617] Dataset preprocessing request;

[0618] Time information.

[0619] The information transmission device proposed in this embodiment belongs to the same inventive concept as the information transmission method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in any of the above embodiments. Furthermore, this embodiment has the same beneficial effects as performing the information transmission method.

[0620] This application also provides a communication node. Figure 20 A schematic diagram of the structure of a communication node is provided as an embodiment, such as... Figure 20 As shown, the communication node provided in this application includes a processor 1310, a memory 1320, and a computer program stored in the memory and executable on the processor. When the processor 1310 executes the program, it implements the information transmission method described above.

[0621] The communication node may also include a memory 1320; the processor 1310 in the communication node may be one or more. Figure 20 Taking a processor 1310 as an example; memory 1320 is used to store one or more programs; the one or more programs are executed by the one or more processors 1310, so that the one or more processors 1310 implement the information transmission method as described in the embodiments of this application.

[0622] The communication node also includes: a communication device 1330, an input device 1340, and an output device 1350.

[0623] The processor 1310, memory 1320, communication device 1330, input device 1340, and output device 1350 in the communication node can be connected via a bus or other means. Figure 20 Taking the example of a connection between China and Israel via a bus.

[0624] Input device 1340 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the communication node. Output device 1350 may include display devices such as a display screen.

[0625] The communication device 1330 may include a receiver and a transmitter. The communication device 1330 is configured to perform information transmission and reception communication under the control of the processor 1310.

[0626] The memory 1320, as a computer-readable storage medium, can be configured to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the information transmission method described in the embodiments of this application (e.g., the discovery request sending module 810, the first communication device information receiving module 820, and the data processing request sending module 830 in the information transmission device; or the data processing request receiving module 910 and the data feedback module 920 in the information transmission device; or the second management request receiving module 1010 and the model processing module 1020 in the information transmission device; or the model capability information receiving module 1110 and the model capability information storage module 1120 in the information transmission device; or the task receiving module 1210, the second communication device information determination module 1220, and the second communication device information feedback module 1230 in the information transmission device). The memory 1320 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the communication node, etc. Furthermore, memory 1320 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, memory 1320 may further include memory remotely located relative to processor 1310, which can be connected to a communication node via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0627] This application also provides a storage medium storing a computer program, which, when executed by a processor, implements any of the information transmission methods described in this application.

[0628] Optionally, the information transmission method, applied to a first communication device, includes: sending a discovery request to a second communication device; receiving first communication device information fed back by the second communication device, wherein the first communication device information is determined based on the model lifecycle capability information of the communication device; determining a third communication device based on the first communication device information, sending a data processing request to the third communication device, and receiving processed data fed back by the third communication device.

[0629] Optionally, this information transmission method is applied to a third communication device and includes: receiving a data processing request sent by a first communication device; and feeding back processed data to the first communication device.

[0630] Optionally, the information transmission method is applied to the first communication device and includes: receiving a second model management request sent by a sixth communication device; and processing the model.

[0631] Optionally, this information transmission method, applied to a second communication device, includes: receiving model lifecycle capability information sent by a seventh communication device, the model lifecycle capability information being used to indicate the model lifecycle capability of the seventh communication device; and storing the model lifecycle capability information.

[0632] Optionally, the information transmission method, applied to a second communication device, includes: receiving a machine learning task sent by an eighth communication device, the machine learning task including a predicted estimation capability of the task; determining second communication device information based on the predicted estimation capability in the machine learning task and pre-stored model lifecycle capability information of the communication device; and feeding back the second communication device information to the eighth communication device.

[0633] The computer storage medium in this application embodiment can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable CD-ROM, optical storage device, magnetic storage device, or any suitable combination thereof. The computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0634] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit programs for use by or in connection with an instruction execution system, apparatus, or device.

[0635] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, radio frequency (RF), etc., or any suitable combination thereof.

[0636] This application provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the information transmission method described in any one of the embodiments of this application.

[0637] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smallport, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0638] The above description is merely an exemplary embodiment of this application and is not intended to limit the scope of protection of this application.

[0639] Those skilled in the art will understand that the term user terminal encompasses any suitable type of wireless user equipment, such as mobile phones, portable data processing devices, portable web browsers, or vehicle-mounted mobile stations.

[0640] Generally, the various embodiments of this application can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. For example, some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device, although this application is not limited thereto.

[0641] Embodiments of this application can be implemented by executing computer program instructions through the data processor of a mobile device, for example, in a processor entity, or through hardware, or through a combination of software and hardware. The computer program instructions can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages.

[0642] Any block diagram of logical flow in the accompanying drawings of this application may represent program steps, or may represent interconnected logic circuits, modules, and functions, or may represent a combination of program steps and logic circuits, modules, and functions. The computer program may be stored on memory. Memory may be of any type suitable to the local technical environment and may be implemented using any suitable data storage technology, such as, but not limited to, read-only memory (ROM), random access memory (RAM), optical storage devices and systems (Digital Video Disc (DVD) or Compact Disc (CD), etc.). Computer-readable media may include non-transitory storage media. The data processor may be of any type suitable to the local technical environment, such as, but not limited to, general-purpose computers, special-purpose computers, microprocessors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), and processors based on multi-core processor architectures.

[0643] A detailed description of exemplary embodiments of this application has been provided above through exemplary and non-limiting examples. However, various modifications and adjustments to the above embodiments will be apparent to those skilled in the art when considered in conjunction with the accompanying drawings and claims, without departing from the scope of this application. Therefore, the proper scope of this application will be determined by the claims.

Claims

1. An information transmission method, characterized in that, Applied to a first communication device, including: Send a discovery request to the second communication device; Receive first communication device information fed back by the second communication device, wherein the first communication device information is determined based on the model lifecycle capability information of the communication device; The third communication device is determined based on the information of the first communication device, a data processing request is sent to the third communication device, and the processed data is received from the third communication device.

2. The information transmission method according to claim 1, characterized in that, Before sending the discovery request to the second communication device, the method further includes: Receive the first model management request sent by the fourth communication device.

3. The information transmission method according to claim 2, characterized in that, The first model management request includes one or more of the following information: Dataset size requirements; Dataset preprocessing instructions; Data preprocessing instructions; Dataset preprocessing methods; Time information.

4. The information transmission method according to claim 3, characterized in that, The dataset is at least one of the following: Training dataset; Test dataset.

5. The information transmission method according to claim 2, characterized in that, The first model management request includes at least one of the following: Model request; Model subscription; Analysis request; Analytics subscription.

6. The information transmission method according to claim 1, characterized in that, Before sending the discovery request to the second communication device, the method further includes: The first communication device is determined to satisfy at least one of the following conditions: The computing power of the first communication device is insufficient; The first communication device does not support the specified data preprocessing method; The first communication device does not have data preprocessing capabilities.

7. The information transmission method according to claim 1, characterized in that, The discovery request includes one or more of the following information: Types of communication equipment; Data preprocessing instructions; Dataset preprocessing instructions; Data preprocessing methods; Dataset preprocessing methods; Time information; Information on the computational power required for data preprocessing.

8. The information transmission method according to claim 1, characterized in that, The data processing request includes one or more of the following information: Data set to be processed; Data preprocessing method instructions; Time information.

9. The information transmission method according to claim 1, characterized in that, The method further includes: Receive the dataset tag and storage address information fed back by the third communication device.

10. The information transmission method according to any one of claims 1-9, characterized in that, The model lifecycle capability information includes one or more of the following: Dataset preprocessing capabilities; Supported dataset preprocessing operations; Model training capability; Model validation capability; Supported model validation methods; The time period information corresponding to the capability; Model application instruction information, used to indicate the application of the model; Available computing power information; The time period corresponding to available computing power information.

11. An information transmission method, characterized in that, Applied to third-party communication devices, including: Receive data processing requests sent by the first communication device; The processed data is fed back to the first communication device.

12. The information transmission method according to claim 11, characterized in that, The data processing request includes one or more of the following information: Data set to be processed; Data preprocessing method instructions; Time information.

13. The information transmission method according to claim 11, characterized in that, Before feeding back the processed data to the first communication device, the method further includes: Identify the data to be processed; The data to be processed is preprocessed to obtain the processed data.

14. The information transmission method according to claim 13, characterized in that, The determination of the data to be processed includes: If the data processing request does not include a dataset to be processed or the included dataset does not meet the requirements, data is collected to form data to be processed.

15. The information transmission method according to any one of claims 11-14, characterized in that, Also includes: A dataset storage request is sent to the fifth communication device, the dataset storage request being used to instruct the storage of the processed data.

16. The information transmission method according to any one of claims 15, characterized in that, Also includes: Receive the dataset labels fed back by the fifth communication device.

17. The information transmission method according to any one of claims 16, characterized in that, Also includes: The dataset label and storage address information are fed back to the first communication device.

18. An information transmission method, characterized in that, Applied to the first communication device, including Receive the second model management request sent by the sixth communication device; The model is then processed.

19. The information transmission method according to claim 18, characterized in that, The model is processed in at least one of the following ways: Train the model; Validate the model.

20. The information transmission method according to claim 18, characterized in that, The second model management request includes one or more of the following information: Model training methods; Model type; Model validation methods; Model validation metrics; Time information; Training dataset size requirements; Verify the required size of the dataset; Requirements for partitioning the training dataset; Verify the requirements for splitting the dataset.

21. The information transmission method according to claim 18, characterized in that, The second model management request includes at least one of the following: Model request; Model subscription; Analysis request; Analytics subscription.

22. The information transmission method according to claim 18, characterized in that, Also includes: Send at least one of the following data to the sixth communication device: Model; Analysis results; Verification method; Verification results.

23. An information transmission method, characterized in that, Applied to a second communication device, including: Receive model lifecycle capability information sent by a seventh communication device, wherein the model lifecycle capability information is used to indicate the model lifecycle capability of the seventh communication device; Store the model's lifecycle capability information.

24. The information transmission method according to claim 23, characterized in that, The received model lifecycle capability information from the seventh communication device includes: Receive model lifecycle capability information registered or updated by the seventh communication device.

25. The information transmission method according to claim 24, characterized in that, Also includes: The seventh communication device shall be informed of either a successful save or a failed save indication. The "save success" indicator indicates that the registration or update of model lifecycle capability information was successful, while the "save failure" indicator indicates that the registration or update of model lifecycle capability information failed.

26. The information transmission method according to any one of claims 23-25, characterized in that, The model lifecycle capability information includes one or more of the following: Dataset preprocessing capabilities; Supported dataset preprocessing operations; Model training capability; Model validation capability; Supported model validation methods; The time period information corresponding to the capability; Model application instruction information, used to indicate the application of the model; Available computing power information; The time period corresponding to available computing power information.

27. The information transmission method according to any one of claims 23-25, characterized in that, The seventh communication device is either the first communication device or the third communication device.

28. An information transmission method, characterized in that, Applied to a second communication device, including: Receive a machine learning task sent by an eighth communication device, the machine learning task including the task's estimated power; The second communication device information is determined based on the estimated capabilities in the machine learning task and the pre-stored model lifecycle capability information of the communication device. The information from the second communication device is fed back to the eighth communication device.

29. The information transmission method according to claim 28, characterized in that, The machine learning task also includes at least one of the following information: Analyze the request; Model training request; Model validation request; Model test request; Dataset preprocessing request; Time information.

30. A communication node, characterized in that, include: The program includes a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for establishing communication between the processor and the memory, wherein the program, when executed by the processor, implements the steps of the information transmission method as described in any one of claims 1-29.

31. A storage medium for computer-readable storage, characterized in that, The storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the information transmission method according to any one of claims 1-29.

32. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the information transmission method according to any one of claims 1-29.

Images