Communication method and communication apparatus

Abstract

Provided in the present application are a communication method and a communication apparatus. The method comprises: sending a first model, which is used for model training; sending a second model, which is used for model training; and sending first indication information, which is used for determining whether a model parameter and / or dataset corresponding to the first model and a model parameter and / or dataset corresponding to the second model can be used for training the same model. On the basis of the present application, if a first apparatus sends a plurality of models to a second apparatus, the first apparatus may send indication information to the second apparatus, such that the second apparatus can determine, on the basis of the indication information, whether model parameters and / or datasets corresponding to at least two of the plurality of models can be used for training the same model, thereby facilitating a reduction in overheads for transmitting the model parameters and / or datasets, and also causing the second apparatus to perform different training tasks on the basis of the plurality of received models, thus ensuring the performance of trained or updated models.

Description

Communication methods and communication devices

[0001] This application claims priority to Chinese Patent Application No. 202411994102.X, filed with the China National Intellectual Property Administration on December 30, 2024, entitled "Communication Method and Communication Device", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of wireless communication, and more particularly to a communication method and a communication device. Background Technology

[0003] Currently, artificial intelligence (AI) has been introduced into wireless communication networks and has been widely applied in many application scenarios of air interface technology, such as AI-based channel state information (CSI) prediction, AI-based beam management, and AI-based CSI feedback, playing an increasingly important role.

[0004] In the practical application of AI models, it is necessary to provide models and / or datasets to the devices that train and / or deploy models for training. Therefore, how to transmit models and / or datasets to reduce air interface overhead has become a pressing technical problem to be solved. Summary of the Invention

[0005] This application provides a communication method and a communication device to reduce the overhead of transmitting models and / or datasets over an air interface.

[0006] In a first aspect, a communication method is provided, the method comprising: sending a first model, the first model being used for model training; sending a second model, the second model being used for model training; and sending first indication information, the first indication information being used to determine one or more of the following: the first model parameters corresponding to the first model and the second model parameters corresponding to the second model can be used for training the same model, or the first model parameters and the second model parameters cannot be used for training the same model; the first dataset corresponding to the first model and the second dataset corresponding to the second model can be used for training the same model, or the first dataset corresponding to the first model and the second dataset corresponding to the second model cannot be used for training the same model.

[0007] The method can be executed by a first device, which can be replaced by a device on the terminal device side, a device on the network device side, or a device on the core network element side.

[0008] The equipment on the terminal device side can include the terminal device itself, the communication module within the terminal device, or the circuits or chips within the terminal device responsible for communication functions (such as a modem chip, also known as a baseband chip, or a system-on-chip (SoC) chip containing a modem core, or a system-in-package (SIP) chip, etc.). Alternatively, the equipment on the terminal device side can include AI entities on the terminal device side. AI entities on the terminal device side can be the terminal device itself, or AI entities serving the terminal device, such as servers, such as over-the-top (OTT) servers or cloud servers.

[0009] Network-side devices can include the network device itself, communication modules within the network device, or circuits or chips responsible for communication functions within the network device (such as modem chips, also known as baseband chips, or system-on-a-chip (SoC) chips or SIP chips containing modem cores, etc.). Alternatively, network-side devices can include AI entities on the network device side. These AI entities can be the network device itself or AI entities serving the network device, such as radio access network (RAN) intelligent controllers (RICs), operation administration and maintenance (OAM) systems, or servers, such as OTT servers or cloud servers.

[0010] The equipment on the core network element side can include the core network element itself, functional modules within the core network element, or circuits or chips within the core network element. Alternatively, the equipment on the core network element side can include AI entities on the core network element side. These AI entities can be the core network element itself or AI entities serving the core network element, such as servers, like OTT servers or cloud servers.

[0011] Based on the above technical solution, if the first device sends multiple models to the second device, the first device can send instruction information to the second device so that the second device can determine whether the model parameters and / or datasets corresponding to at least two of the multiple models can be used for training the same model. This is beneficial to save the transmission overhead of model parameters and / or datasets, and enable the second device to perform different training tasks based on the received multiple models, thereby ensuring the performance of the trained or updated models.

[0012] For example, if the first dataset sent by the first device is used for training the first model, and the second dataset sent by the first device is used for training the second model, then if the second device determines, based on the received first instruction information, that the first dataset and the second dataset can be used to train the same model, it can train the first model based on the first dataset and the second dataset. This is equivalent to ensuring the performance of the trained third model while increasing the dataset used to train the first model.

[0013] For another example, if the second device can train the same model based on at least two previously received datasets after cell handover, thereby meeting the model requirements after cell handover, then the first device does not need to send a new model to the second device after the second device performs cell handover, thus saving data transmission requirements.

[0014] For example, the model parameters corresponding to the model mentioned in this application may include all the model parameters, or may include some of the model parameters. If the model parameters corresponding to the model include some of the model parameters, then the model parameters corresponding to the model may be predefined, or may be indicated by the first device to the second device. For example, if the model parameters corresponding to the model include some of the model parameters, the model parameters corresponding to the model may include weights.

[0015] For example, the dataset corresponding to the model mentioned in this application may include the entire dataset corresponding to the model, or it may include a portion of the dataset corresponding to the model. If the dataset corresponding to the model includes a portion of the dataset corresponding to the model, then the dataset corresponding to the model may be predefined, or it may be indicated by the first device to the second device. For example, if the dataset corresponding to the model includes some of the parameters corresponding to the model, the dataset corresponding to the model may include the training dataset corresponding to the model.

[0016] For example, the model structure of the first model is the same as that of the second model, and the first indication information is used to determine one or more of the following: the parameters of the first model and the parameters of the second model can be used for training the same model, or the parameters of the first model and the parameters of the second model cannot be used for training the same model; the first dataset and the second dataset can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model.

[0017] For example, the model structure of the first model is different from that of the second model. The first indication information is used to determine whether the first dataset and the second dataset can be used for training the same model, or to determine whether the first dataset and the second dataset can be used for training the same model.

[0018] In conjunction with the first aspect, in certain implementations of the first aspect, sending the first model includes: sending the first model and its identifier; sending the second model includes: sending the second model and its identifier; the first indication information includes identifiers of multiple models, and the model parameters and / or datasets corresponding to the multiple models can be used for training the same model; or, the first indication information is used to indicate that the model parameters and / or datasets corresponding to models with different identifiers can be used for training the same model; or, the first indication information is used to indicate that the model parameters and / or datasets corresponding to models with different identifiers cannot be used for training the same model; or, the first indication information is used to indicate the conditions that the identifiers of at least two models whose corresponding model parameters and / or datasets can be used for training the same model must satisfy.

[0019] Based on the above technical solution, when the first device sends the model identifier to the second device, the first instruction information sent by the first device to the second device can indicate rule #1. Rule #1 is related to the model identifier, so that the second device can determine whether the model parameters and / or datasets corresponding to at least two models can be used or not used for training the same model according to rule #1.

[0020] In conjunction with the first aspect, in certain implementations of the first aspect, sending a first model includes: sending a first model and a first identifier, the first identifier being used to identify a first value corresponding to a first feature of the first model; the first feature belongs to a first feature set; sending a second model includes: sending a second model and a second identifier, the second identifier being used to identify a second value corresponding to a first feature of the second model; the first indication information includes multiple first feature identifiers, the multiple first feature identifiers being used to identify multiple third values ​​corresponding to the first feature, or the first indication information includes one or more first feature identifiers, each first feature identifier being used to identify a set of first values ​​corresponding to the first feature; wherein, the values ​​corresponding to the first features of at least two models are contained in multiple third values ​​or at least one set of first values, and the model parameters corresponding to at least two models can be used for training the same model; or, the first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the first features can be used for training the same model; or, the first indication information is used to indicate that the values ​​corresponding to the first features of at least two models whose corresponding model parameters can be used for training the same model must satisfy certain conditions.

[0021] Based on the above technical solution, when the first device sends an identifier to the second device to identify the value corresponding to the first feature of the model, the first indication information sent by the first device to the second device can indicate rule #2 or parameters related to rule #2. The parameters related to rule #2 are related to the values ​​corresponding to the features in the first feature set of the model, so that the second device can determine whether the model parameters corresponding to at least two models can be used or cannot be used for training the same model based on the parameters related to rule #2 or rule #2.

[0022] For example, the first feature set includes one or more of the following: learning rate, retention probability to prevent overfitting, total number of training epochs, batch size, size of training dataset, neural network dimension, and number of neural network layers.

[0023] In conjunction with the first aspect, in certain implementations of the first aspect, sending the first model includes: sending the first model and a third identifier, the third identifier being used to identify a fourth value corresponding to a second feature of the first model; the second feature belongs to a second feature set; sending the second model includes: sending the second model and a fourth identifier, the fourth identifier being used to identify a fifth value corresponding to a second feature of the second model; the first indication information includes multiple second feature identifiers, the multiple second feature identifiers being used to identify multiple sixth values ​​corresponding to the second feature, or the first indication information includes one or more second feature identifiers, each second feature identifier being used to identify a set of second values ​​corresponding to the second feature; wherein, the values ​​corresponding to the second features of at least two models are contained in multiple sixth values ​​or at least one set of second values, and the model parameters corresponding to at least two models can be used for training the same model; or, the first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the second features can be used for training the same model; or, the first indication information is used to indicate that the values ​​corresponding to the second features of at least two models whose corresponding model parameters can be used for training the same model must satisfy certain conditions.

[0024] Based on the above technical solution, when the first device sends an identifier to the second device to identify the value corresponding to the second feature of the model, the first indication information sent by the first device to the second device can indicate rule #3 or parameters related to rule #3. Rule #3 or parameters related to rule #3 are related to the value corresponding to the feature in the second feature set of the model, so that the second device can determine whether the model parameters corresponding to at least two models can be used or cannot be used for training the same model based on rule #3 or parameters related to rule #3.

[0025] For example, the second feature set includes one or more of the following: the configuration of the historical time window length corresponding to the data in the training dataset corresponding to the model, the configuration of the prediction time window length corresponding to the data in the training dataset corresponding to the model, and the function for compressing channel state information (CSI) and / or predicting CSI.

[0026] In conjunction with the first aspect, in certain implementations of the first aspect, sending the first model includes: sending the first model, the first dataset, and the identifier of the first dataset; sending the second model includes: sending the second model, the second dataset, and the identifier of the second dataset; the first indication information includes the identifiers of multiple datasets, which can be used to train the same model; or, the first indication information is used to indicate that datasets with different identifiers can be used to train the same model; or, the first indication information is used to indicate that datasets with different identifiers cannot be used to train the same model; or, the first indication information is used to indicate the conditions that the identifiers of at least two datasets that can be used to train the same model must satisfy.

[0027] Based on the above technical solution, when the first device sends the identifier of the dataset to the second device, the first indication information sent by the first device to the second device can indicate rule #4. Rule #4 is related to the identifier of the dataset, so that the second device can determine whether at least two datasets can be used or not used for training the same model according to rule #4.

[0028] In conjunction with the first aspect, in certain implementations of the first aspect, sending the first model includes: sending the first model, the first dataset, and the fifth identifier, wherein the fifth identifier is used to identify the seventh value corresponding to the third feature of the first dataset, and the third feature belongs to the third feature set; sending the second model includes: sending the second model, the second dataset, and the sixth identifier, wherein the sixth identifier is used to identify the eighth value corresponding to the third feature of the second dataset; the first indication information includes multiple third feature identifiers, which are used to identify multiple ninth values ​​corresponding to the third feature, or the first indication information includes one or more third feature identifiers, each of which is used to identify a set of third values ​​corresponding to the third feature; wherein the values ​​corresponding to the third feature of at least two datasets are included in multiple ninth values ​​or at least one set of third values, and at least two datasets can be used for training the same model; or the first indication information is used to indicate that datasets with different values ​​corresponding to the third feature can be used for training the same model; or the first indication information is used to indicate that datasets with different values ​​corresponding to the third feature cannot be used for training the same model; or the first indication information is used to indicate the conditions that the values ​​corresponding to the third feature of at least two datasets that can be used for training the same model must satisfy.

[0029] Based on the above technical solution, when the first device sends an identifier to the second device to identify the value corresponding to the third feature of the dataset, the first indication information sent by the first device to the second device can indicate rule #5 or a parameter related to rule #5. Rule #5 or a parameter related to rule #5 is related to the value corresponding to the feature in the third feature set of the dataset, thereby enabling the second device to determine whether at least two datasets can be used or not used for training the same model based on rule #5 or a parameter related to rule #5.

[0030] For example, the third feature set includes one or more of the following: whether the data was acquired during training with a quantizer, the range of transmit power distribution corresponding to the data, the range of receive power distribution corresponding to the data, the range of receive noise power distribution corresponding to the data, the range of time delay spread distribution corresponding to the data, the range of angle spread distribution corresponding to the data, the range of interference level distribution corresponding to the data, and the range of signal to interference plus noise ratio (SINR) distribution corresponding to the data.

[0031] In conjunction with the first aspect, in certain implementations of the first aspect, sending the first model includes: sending the first model, the first dataset, and the seventh identifier, wherein the seventh identifier is used to identify the tenth value corresponding to the fourth feature of the first dataset, and the fourth feature belongs to the fourth feature set; sending the second model includes: sending the second model, the second dataset, and the eighth identifier, wherein the eighth identifier is used to identify the eleventh value corresponding to the fourth feature of the second dataset; the first indication information includes multiple fourth feature identifiers, which are used to identify multiple twelfth values ​​corresponding to the fourth feature, or the first indication information includes one or more fourth feature identifiers, each fourth feature identifier being used to identify a set of fourth values ​​corresponding to the fourth feature; wherein the values ​​corresponding to the fourth feature of at least two datasets are included in multiple twelfth values ​​or at least one set of fourth values, and at least two datasets can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature cannot be used for training the same model; or, the first indication information is used to indicate the conditions that the values ​​corresponding to the fourth feature of at least two datasets that can be used for training the same model must satisfy.

[0032] Based on the above technical solution, when the first device sends an identifier to the second device to identify the value corresponding to the fourth feature of the dataset, the first indication information sent by the first device to the second device can indicate rule #6 or a parameter related to rule #6. The parameter related to rule #6 is related to the value corresponding to the feature in the fourth feature set of the dataset, so that the second device can determine whether at least two datasets can be used or not used for training the same model based on the parameter related to rule #6 or rule #6.

[0033] For example, the fourth feature set includes one or more of the following: the number of transmission ports corresponding to the data in the dataset, the carrier frequency corresponding to the data in the dataset, the transmission bandwidth corresponding to the data in the dataset, the number of transmission subbands corresponding to the data in the dataset, the granularity of the transmission subbands corresponding to the data in the dataset, and the number of bits of CSI feedback overhead corresponding to the data in the dataset.

[0034] In conjunction with the first aspect, in certain implementations of the first aspect, sending the first model includes: sending the first model, the first dataset, and the ninth identifier, wherein the ninth identifier is used to identify the thirteenth value corresponding to the fifth feature of the first dataset, and the fifth feature belongs to the fifth feature set; sending the second model includes: sending the second model, the second dataset, and the tenth identifier, wherein the tenth identifier is used to identify the fourteenth value corresponding to the fifth feature of the second dataset; the first indication information includes multiple fifth feature identifiers, which are used to identify multiple fifteenth values ​​corresponding to the fifth feature, or the first indication information includes one or more fifth feature identifiers, each fifth feature identifier being used to identify a set of fifth values ​​corresponding to the fifth feature; wherein the values ​​corresponding to the fifth feature of at least two datasets are included in multiple fifteenth values ​​or at least one set of fifth values, and at least two datasets can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature cannot be used for training the same model; or, the first indication information is used to indicate the conditions that the values ​​corresponding to the fifth feature of at least two datasets that can be used for training the same model must satisfy.

[0035] Based on the above technical solution, when the first device sends an identifier to the second device to identify the value corresponding to the fifth feature of the dataset, the first indication information sent by the first device to the second device can indicate rule #7 or a parameter related to rule #7. The parameter related to rule #7 is related to the value corresponding to the feature in the fifth feature set of the dataset, so that the second device can determine whether at least two datasets can be used or not used for training the same model based on the parameter related to rule #7 or rule #7.

[0036] For example, the fifth feature set includes one or more of the following: the length of the historical time window corresponding to the data in the dataset, the length of the prediction time window corresponding to the data in the dataset, and the moving speed of the device corresponding to the data in the dataset.

[0037] Secondly, a communication method is provided, the method comprising: receiving a first model, the first model being used for model training; receiving a second model, the second model being used for model training; receiving first indication information, the first indication information being used to determine one or more of the following: the first model parameters corresponding to the first model and the second model parameters corresponding to the second model can be used for training the same model, or the first model parameters and the second model parameters cannot be used for training the same model; the first dataset corresponding to the first model and the second dataset corresponding to the second model can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model.

[0038] This method can be executed by a second device, which can be replaced by a device on the terminal device side, a device on the network device side, or a device on the core network element side. Further description of the device on the terminal device side, the device on the network device side, or the device on the core network element side can be found in the first aspect above.

[0039] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: the model structure of the first model is the same as the model structure of the second model, and one or more of the following are determined according to the first indication information: the parameters of the first model and the parameters of the second model can be used for training the same model, or the parameters of the first model and the parameters of the second model cannot be used for training the same model; the first dataset and the second dataset can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model; or, the model structure of the first model and the model structure of the second model are different, and one or more of the following are determined according to the first indication information: the first dataset and the second dataset can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model.

[0040] In conjunction with the second aspect, in some implementations of the second aspect, receiving a first model includes: receiving a first model and its identifier; receiving a second model includes: receiving a second model and its identifier; the first indication information includes identifiers of multiple models, and the model parameters and / or datasets corresponding to the multiple models can be used for training the same model; or, the first indication information is used to indicate that the model parameters and / or datasets corresponding to models with different identifiers can be used for training the same model; or, the first indication information is used to indicate that the model parameters and / or datasets corresponding to models with different identifiers cannot be used for training the same model; or, the first indication information is used to indicate the conditions that the identifiers of at least two models whose corresponding model parameters and / or datasets can be used for training the same model must satisfy.

[0041] In conjunction with the second aspect, in some implementations of the second aspect, receiving a first model includes: receiving a first model and a first identifier, the first identifier being used to identify a first value corresponding to a first feature of the first model; the first feature belongs to a first feature set; receiving a second model includes: receiving a second model and a second identifier, the second identifier being used to identify a second value corresponding to a first feature of the second model; the first indication information includes multiple first feature identifiers, the multiple first feature identifiers being used to identify multiple third values ​​corresponding to the first feature, or the first indication information includes one or more first feature identifiers, each first feature identifier being used to identify a set of first values ​​corresponding to the first feature; wherein, the values ​​corresponding to the first features of at least two models are contained in multiple third values ​​or at least one set of first values, and the model parameters corresponding to at least two models can be used for training the same model; or, the first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the first features can be used for training the same model; or, the first indication information is used to indicate that the values ​​corresponding to the first features of at least two models whose corresponding model parameters can be used for training the same model must satisfy certain conditions.

[0042] For example, the first feature set includes one or more of the following: learning rate, retention probability to prevent overfitting, total number of training epochs, batch size, size of training dataset, neural network dimension, and number of neural network layers.

[0043] In conjunction with the second aspect, in certain implementations of the second aspect, receiving the first model includes: receiving the first model and a third identifier, the third identifier being used to identify a fourth value corresponding to a second feature of the first model; the second feature belongs to a second feature set; receiving the second model includes: receiving the second model and a fourth identifier, the fourth identifier being used to identify a fifth value corresponding to a second feature of the second model; the first indication information includes multiple second feature identifiers, the multiple second feature identifiers being used to identify multiple sixth values ​​corresponding to the second feature, or the first indication information includes one or more second feature identifiers, each second feature identifier being used to identify a set of second values ​​corresponding to the second feature; wherein, the values ​​corresponding to the second features of at least two models are contained in multiple sixth values ​​or at least one set of second values, and the model parameters corresponding to at least two models can be used for training the same model; or, the first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the second features can be used for training the same model; or, the first indication information is used to indicate that the values ​​corresponding to the second features of at least two models whose corresponding model parameters can be used for training the same model must satisfy certain conditions.

[0044] For example, the second feature set includes one or more of the following: the configuration of the historical time window length corresponding to the data in the training dataset corresponding to the model, the configuration of the prediction time window length corresponding to the data in the training dataset corresponding to the model, and the function for compressing CSI and / or predicting CSI.

[0045] In conjunction with the second aspect, in some implementations of the second aspect, the first indication information indicates a first rule and a second rule, with the first rule having a higher priority than the second rule. The first rule and the second rule belong to the following rules used to determine whether model parameters corresponding to at least two models can be used or cannot be used for training the same model: rules related to the model's identifier, rules related to the values ​​corresponding to the features included in the first feature set of the model, and rules related to the values ​​corresponding to the features included in the second feature set of the model. The method further includes: determining, according to the first rule, that the first model parameters and the second model parameters can be used for training the same model, or determining that the first model parameters and the second model parameters cannot be used for training the same model.

[0046] Based on the above technical solution, if the first indication information indicates multiple rules, the second device can prioritize determining whether the first model parameter and the second model parameter can be used for training the same model according to the rule with higher priority. Thus, if the second device determines that the first model parameter and the second model parameter cannot be used for training the same model according to the rule with higher priority, it is not necessary to determine whether the first model parameter and the second model parameter can be used for training the same model according to the rule with lower priority.

[0047] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes determining, according to the first rule, that the first model parameters and the second model parameters can be used for training the same model, or determining that the first model parameters and the second model parameters cannot be used for training the same model.

[0048] In conjunction with the second aspect, in some implementations of the second aspect, receiving a first model includes: receiving a first model, a first dataset, and an identifier of the first dataset; receiving a second model includes: receiving a second model, a second dataset, and an identifier of the second dataset; the first indication information includes identifiers of multiple datasets, which can be used to train the same model; or, the first indication information is used to indicate that datasets with different identifiers can be used to train the same model; or, the first indication information is used to indicate that datasets with different identifiers cannot be used to train the same model; or, the first indication information is used to indicate the conditions that the identifiers of at least two datasets that can be used to train the same model must satisfy.

[0049] In conjunction with the second aspect, in some implementations of the second aspect, receiving a first model includes: receiving a first model, a first dataset, and a fifth identifier, wherein the fifth identifier is used to identify the seventh value corresponding to the third feature of the first dataset, and the third feature belongs to a third feature set; receiving a second model includes: receiving a second model, a second dataset, and a sixth identifier, wherein the sixth identifier is used to identify the eighth value corresponding to the third feature of the second dataset; the first indication information includes multiple third feature identifiers, which are used to identify multiple ninth values ​​corresponding to the third feature, or the first indication information includes one or more third feature identifiers, each of which is used to identify a set of third values ​​corresponding to the third feature; wherein the values ​​corresponding to the third feature of at least two datasets are included in multiple ninth values ​​or at least one set of third values, and at least two datasets can be used for training the same model; or the first indication information is used to indicate that datasets with different values ​​corresponding to the third feature can be used for training the same model; or the first indication information is used to indicate that datasets with different values ​​corresponding to the third feature cannot be used for training the same model; or the first indication information is used to indicate the conditions that the values ​​corresponding to the third feature of at least two datasets that can be used for training the same model must satisfy.

[0050] For example, the third feature set includes one or more of the following: whether the data was acquired during training with a quantizer, the absolute received power distribution range of the data, the received power distribution range of the data, the received noise power distribution range of the data, the time delay spread distribution range of the data, the angle spread distribution range of the data, the interference level distribution range of the data, and the SINR distribution range of the data.

[0051] In conjunction with the second aspect, in some implementations of the second aspect, receiving a first model includes: receiving a first model, a first dataset, and a seventh identifier, wherein the seventh identifier is used to identify the tenth value corresponding to the fourth feature of the first dataset, and the fourth feature belongs to a fourth feature set; receiving a second model includes: receiving a second model, a second dataset, and an eighth identifier, wherein the eighth identifier is used to identify the eleventh value corresponding to the fourth feature of the second dataset; the first indication information includes multiple fourth feature identifiers, which are used to identify multiple twelfth values ​​corresponding to the fourth feature, or the first indication information includes one or more fourth feature identifiers, each fourth feature identifier being used to identify a set of fourth values ​​corresponding to the fourth feature; wherein the values ​​corresponding to the fourth feature of at least two datasets are included in multiple twelfth values ​​or at least one set of fourth values, and at least two datasets can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature cannot be used for training the same model; or, the first indication information is used to indicate the conditions that the values ​​corresponding to the fourth feature of at least two datasets that can be used for training the same model must satisfy.

[0052] For example, the fourth feature set includes one or more of the following: the number of receiving ports corresponding to the data in the dataset, the carrier frequency corresponding to the data in the dataset, the receiving bandwidth corresponding to the data in the dataset, the number of receiving sub-bands corresponding to the data in the dataset, the receiving sub-band granularity corresponding to the data in the dataset, and the number of bits of CSI feedback overhead corresponding to the data in the dataset.

[0053] In conjunction with the second aspect, in some implementations of the second aspect, receiving a first model includes: receiving a first model, a first dataset, and a ninth identifier, wherein the ninth identifier is used to identify the thirteenth value corresponding to the fifth feature of the first dataset, and the fifth feature belongs to a fifth feature set; receiving a second model includes: receiving a second model, a second dataset, and a tenth identifier, wherein the tenth identifier is used to identify the fourteenth value corresponding to the fifth feature of the second dataset; the first indication information includes multiple fifth feature identifiers, which are used to identify multiple fifteenth values ​​corresponding to the fifth feature, or the first indication information includes one or more fifth feature identifiers, each fifth feature identifier being used to identify a set of fifth values ​​corresponding to the fifth feature; wherein the values ​​corresponding to the fifth feature of at least two datasets are included in multiple fifteenth values ​​or at least one set of fifth values, and at least two datasets can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature can be used for training the same model; or, the first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature cannot be used for training the same model; or, the first indication information is used to indicate the conditions that the values ​​corresponding to the fifth feature of at least two datasets that can be used for training the same model must satisfy.

[0054] For example, the fifth feature set includes one or more of the following: the length of the historical time window corresponding to the data in the dataset, the length of the prediction time window corresponding to the data in the dataset, and the moving speed of the device corresponding to the data in the dataset.

[0055] In conjunction with the second aspect, in some implementations of the second aspect, the first indication information indicates the third rule and the fourth rule, with the third rule having a higher priority than the fourth rule. The third rule and the fourth rule belong to the following rules used to determine whether datasets corresponding to at least two models can be used or cannot be used for training the same model: rules related to the model's identifier, rules related to the dataset's identifier, rules related to the values ​​corresponding to the features included in the third feature set of the dataset, rules related to the values ​​corresponding to the features included in the fourth feature set of the dataset, and rules related to the values ​​corresponding to the features included in the fifth feature set of the dataset. The method further includes: determining, according to the third rule, that the first dataset and the second dataset can be used for training the same model, or determining that the first dataset and the second dataset cannot be used for training the same model.

[0056] Based on the above technical solution, if the first indication information indicates multiple rules, the second device can prioritize the rule with higher priority to determine whether the first dataset and the second dataset can be used for training the same model. Thus, if the second device determines that the first dataset and the second dataset cannot be used for training the same model based on the rule with higher priority, it is not necessary to determine whether the first dataset and the second dataset can be used for training the same model based on the rule with lower priority.

[0057] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes determining, according to the third rule, that the first dataset and the second dataset can be used for training the same model, or determining, according to the fourth rule, that the first dataset and the second dataset cannot be used for training the same model.

[0058] It should be understood that the methods provided in the second aspect correspond to those in the first aspect. The descriptions and technical effects of the various implementation methods in the second aspect can be found in the relevant descriptions in the first aspect, and will not be repeated here.

[0059] Thirdly, a communication method is provided, which may include: receiving a first model and an eleventh identifier, wherein the eleventh identifier is associated with the first model, and the first model is used for model training; receiving second model parameters and a twelfth identifier, wherein the twelfth identifier is associated with the second model, and the second model is used for model training; determining second information based on first information associated with the eleventh identifier and the twelfth identifier, wherein the second information includes one or more of the following: the first model parameters corresponding to the first model and the second model parameters corresponding to the second model can be used for training the same model, or the first model parameters and the second model parameters cannot be used for training the same model; or the first dataset corresponding to the first model and the second dataset corresponding to the second model can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model.

[0060] This method can be executed by a second device, which can be replaced by a device on the terminal device side, a device on the network device side, or a device on the core network element side. Further description of the device on the terminal device side, the device on the network device side, or the device on the core network element side can be found in the first aspect above.

[0061] In conjunction with the third aspect, in some implementations of the third aspect, the model structure of the first model is the same as that of the second model, and the second information includes one or more of the following: the parameters of the first model and the parameters of the second model can be used for training the same model, or the parameters of the first model and the parameters of the second model cannot be used for training the same model; the first dataset and the second dataset can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model; or, the model structure of the first model and the model structure of the second model are different, and the second information includes one of the following: the first dataset and the second dataset can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model.

[0062] In conjunction with the third aspect, in some implementations of the third aspect, the eleventh identifier includes the identifier of the first model, and the twelfth identifier includes the identifier of the second model; the first information includes the identifiers of multiple models, and the model parameters and / or datasets corresponding to the multiple models can be used for training the same model; or, the first information is used to indicate that the model parameters and / or datasets corresponding to models with different identifiers can be used for training the same model; or, the first information is used to indicate that the model parameters and / or datasets corresponding to models with different identifiers cannot be used for training the same model; or, the first information is used to indicate the conditions that the identifiers of at least two models whose corresponding model parameters and / or datasets can be used for training the same model must satisfy.

[0063] In conjunction with the third aspect, in some implementations of the third aspect, the eleventh identifier includes a first identifier, which is used to identify a first value corresponding to a first feature of the first model; the twelfth identifier includes a second identifier, which is used to identify a second value corresponding to a first feature of the second model; the first feature belongs to a first feature set; the first information includes multiple first feature identifiers, which are used to identify multiple third values ​​corresponding to the first feature, or the first information includes one or more first feature identifiers, each first feature identifier being used to identify a set of first values ​​corresponding to the first feature; wherein, the values ​​corresponding to the first features of at least two models are contained in multiple third values ​​or at least one set of first values, and the model parameters corresponding to at least two models can be used for training the same model; or, the first information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the first features can be used for training the same model; or, the first information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the first features cannot be used for training the same model; or, the first information is used to indicate the conditions that the values ​​corresponding to the first features of at least two models must satisfy for the corresponding model parameters to be used for training the same model.

[0064] In conjunction with the third aspect, in some implementations of the third aspect, the first feature set includes one or more of the following: learning rate, retention probability to prevent overfitting, total number of training epochs, batch size, size of training dataset, neural network dimension, and number of neural network layers.

[0065] In conjunction with the third aspect, in some implementations of the third aspect, the eleventh identifier includes a third identifier, which is used to identify the fourth value corresponding to the second feature of the first model; the twelfth identifier includes a fourth identifier, which is used to identify the fifth value corresponding to the second feature of the second model; the second feature belongs to a second feature set; the first information includes multiple second feature identifiers, which are used to identify multiple sixth values ​​corresponding to the second feature; or, the first information includes one or more second feature identifiers, each second feature identifier being used to identify a set of second values ​​corresponding to the second feature; wherein, the values ​​corresponding to the second features of at least two models are contained in multiple sixth values ​​or at least one set of second values, and the model parameters corresponding to at least two models can be used for training the same model; or, the first information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the second features can be used for training the same model; or, the first information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the second features cannot be used for training the same model; or, the first information is used to indicate the conditions that the values ​​corresponding to the second features of at least two models must satisfy for the corresponding model parameters to be used for training the same model.

[0066] In conjunction with the third aspect, in some implementations of the third aspect, the second feature set includes one or more of the following: the configuration of the historical time window length corresponding to the data in the training dataset corresponding to the model, the configuration of the prediction time window length corresponding to the data in the training dataset corresponding to the model, and the function for compressing CSI and / or predicting CSI.

[0067] In conjunction with the third aspect, in some implementations of the third aspect, the first information indicates the first rule and the second rule, with the first rule having a higher priority than the second rule. The first rule and the second rule belong to the following rules used to determine whether the model parameters corresponding to at least two models can be used or cannot be used for training the same model: rules related to the model's identifier, rules related to the values ​​corresponding to the features included in the first feature set of the model, and rules related to the values ​​corresponding to the features included in the second feature set of the model. The second information is determined based on the first information associated with the eleventh and twelfth identifiers, including: determining, based on the eleventh identifier, the twelfth identifier, and the first rule, that the first model parameters and the second model parameters can be used for training the same model, or determining that the first model parameters and the second model parameters cannot be used for training the same model.

[0068] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes determining, based on the eleventh identifier, the twelfth identifier, and the first rule, that the first model parameters and the second model parameters can be used for training the same model.

[0069] In conjunction with the third aspect, in some implementations of the third aspect, the eleventh identifier includes the identifier of the first dataset, and the twelfth identifier includes the identifier of the second dataset; the first information includes the identifiers of multiple datasets, which can be used to train the same model; or, the first information is used to indicate that datasets with different identifiers can be used to train the same model; or, the first information is used to indicate that datasets with different identifiers cannot be used to train the same model; or, the first information is used to indicate the conditions that the identifiers of at least two datasets that can be used to train the same model must satisfy.

[0070] In conjunction with the third aspect, in some implementations of the third aspect, the eleventh identifier includes the fifth identifier, which is used to identify the seventh value corresponding to the third feature of the first dataset; the twelfth identifier includes the sixth identifier, which is used to identify the eighth value corresponding to the third feature of the second dataset; the third feature belongs to the third feature set; the first information includes multiple third feature identifiers, which are used to identify multiple ninth values ​​corresponding to the third feature, or the first information includes one or more third feature identifiers, each of which is used to identify a set of third values ​​corresponding to the third feature; wherein, the values ​​corresponding to the third feature of at least two datasets are contained in multiple ninth values ​​or at least one set of third values, and at least two datasets can be used for training the same model; or, the first information is used to indicate that datasets with different values ​​corresponding to the third feature can be used for training the same model; or, the first information is used to indicate that datasets with different values ​​corresponding to the third feature cannot be used for training the same model; or, the first information is used to indicate the conditions that the values ​​corresponding to the third feature of at least two datasets that can be used for training the same model must satisfy.

[0071] In conjunction with the third aspect, in some implementations of the third aspect, the third feature set includes one or more of the following: whether the data was acquired during training with a quantizer, the absolute transmit power distribution range corresponding to the data, the receive power distribution range corresponding to the data, the receive noise power distribution range corresponding to the data, the time delay spread distribution range corresponding to the data, the angle spread distribution range corresponding to the data, the interference level distribution range corresponding to the data, and the SINR distribution range corresponding to the data.

[0072] In conjunction with the third aspect, in some implementations of the third aspect, the eleventh identifier includes the seventh identifier, which is used to identify the tenth value corresponding to the fourth feature of the first dataset; the twelfth identifier includes the eighth identifier, which is used to identify the eleventh value corresponding to the fourth feature of the second dataset; the fourth feature belongs to the fourth feature set; the first information includes multiple fourth feature identifiers, which are used to identify multiple twelfth values ​​corresponding to the fourth feature; or, the first information includes one or more fourth feature identifiers, each of which is used to identify a set of fourth values ​​corresponding to the fourth feature; wherein, the values ​​corresponding to the fourth feature of at least two datasets are contained in multiple twelfth values ​​or at least one set of fourth values, and at least two datasets can be used for training the same model; or, the first information is used to indicate that datasets with different values ​​corresponding to the fourth feature can be used for training the same model; or, the first information is used to indicate that datasets with different values ​​corresponding to the fourth feature cannot be used for training the same model; or, the first information is used to indicate the conditions that the values ​​corresponding to the fourth feature of at least two datasets that can be used for training the same model must satisfy.

[0073] In conjunction with the third aspect, in some implementations of the third aspect, the fourth feature set includes one or more of the following: the number of transmission ports corresponding to the data in the dataset, the carrier frequency corresponding to the data in the dataset, the transmission bandwidth corresponding to the data in the dataset, the number of transmission subbands corresponding to the data in the dataset, the granularity of the transmission subbands corresponding to the data in the dataset, and the number of bits of CSI feedback overhead corresponding to the data in the dataset.

[0074] In conjunction with the third aspect, in some implementations of the third aspect, the eleventh identifier includes the ninth identifier, which is used to identify the thirteenth value corresponding to the fifth feature of the first dataset; the twelfth identifier includes the tenth identifier, which is used to identify the fourteenth value corresponding to the fifth feature of the second dataset; the fifth feature belongs to the fifth feature set; the first information includes multiple fifth feature identifiers, which are used to identify multiple fifteenth values ​​corresponding to the fifth feature; or, the first information includes one or more fifth feature identifiers, each fifth feature identifier being used to identify a set of fifth values ​​corresponding to the fifth feature; wherein, the values ​​corresponding to the fifth feature of at least two datasets are contained in multiple fifteenth values ​​or at least one set of fifth values, and at least two datasets can be used for training the same model; or, the first information is used to indicate that datasets with different values ​​corresponding to the fifth feature can be used for training the same model; or, the first information is used to indicate that datasets with different values ​​corresponding to the fifth feature cannot be used for training the same model; or, the first information is used to indicate the conditions that the values ​​corresponding to the fifth feature of at least two datasets that can be used for training the same model must satisfy.

[0075] In conjunction with the third aspect, in some implementations of the third aspect, the fifth feature set includes one or more of the following: the length of the historical time window corresponding to the data in the dataset, the length of the prediction time window corresponding to the data in the dataset, and the moving speed of the device corresponding to the data in the dataset.

[0076] In conjunction with the third aspect, in some implementations of the third aspect, the first information indicates the third rule and the fourth rule, with the third rule having a higher priority than the fourth rule. The third rule and the fourth rule belong to the following rules used to determine whether datasets corresponding to at least two models can be used or not used for training the same model: rules related to the model's identifier, rules related to the dataset's identifier, rules related to the values ​​corresponding to the features included in the third feature set of the dataset, rules related to the values ​​corresponding to the features included in the fourth feature set of the dataset, and rules related to the values ​​corresponding to the features included in the fifth feature set of the dataset. Based on the first information associated with the eleventh and twelfth identifiers, the second information is determined, including: determining, based on the eleventh identifier, the twelfth identifier, and the third rule, that the first dataset and the second dataset can be used for training the same model, or determining that the first dataset and the second dataset cannot be used for training the same model.

[0077] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes determining that the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the third rule. Alternatively, the method may also determine that the first dataset and the second dataset can be used for training the same model based on the fourth rule, or determine that the first dataset and the second dataset cannot be used for training the same model.

[0078] It should be understood that the descriptions and technical effects of the various implementation methods in the third aspect can be found in the relevant descriptions of the first or second aspect, and will not be repeated here.

[0079] Fourthly, an apparatus is provided. This apparatus may include functional modules corresponding to each of the methods / operations / steps / actions described in any possible implementation of the first aspect, or may include functional modules corresponding to each of the methods / operations / steps / actions described in any of the second aspects, or may include functional modules corresponding to each of the methods / operations / steps / actions described in any of the third aspects. The module may be a hardware circuit, software, or a combination of hardware circuitry and software implementation.

[0080] In one design, the device may include a processing module and a communication module. The communication module is used to perform the sending and receiving actions performed by the first device in the method described in the first aspect above, while the processing module is used to perform processing-related actions performed by the first device in the method described in the first aspect above.

[0081] In one design, the device can be a network device, or a device, module, circuit, or chip configured in the network device, or a device that can be used in conjunction with the network device, such as an intelligent network element with a radio access network (RAN) intelligent controller (RIC) deployed thereon.

[0082] In one design, the device may include a processing module and a communication module. The communication module is used to perform the sending and receiving actions performed by the second device in the method described in the second or third aspect above, while the processing module is used to perform processing-related actions performed by the second device in the method described in the second or third aspect above.

[0083] In one design, the device can be a terminal device, or a device, module, circuit, or chip configured in the terminal device, or a device that can be used in conjunction with the terminal device, such as an OTT host or cloud server.

[0084] Fifthly, an apparatus is provided, comprising a processor and a storage medium storing instructions which, when executed by the processor, cause a method as described in the first aspect or any possible implementation thereof to be implemented, or cause a method as described in the second aspect or any possible implementation thereof to be implemented, or cause a method as described in the third aspect or any possible implementation thereof to be implemented.

[0085] A sixth aspect provides an apparatus comprising a processing circuit for processing data and / or information such that a method as in the first aspect or any possible implementation thereof is implemented, or a method as in the second aspect or any possible implementation thereof is implemented, or a method as in the third aspect or any possible implementation thereof is implemented.

[0086] The processing circuit may include one or more processors, or all or part of the circuitry in one or more processors used for control or processing functions.

[0087] Optionally, the apparatus may further include a memory for storing programs or instructions, and the processor for running the programs or instructions to implement the methods as in the first aspect or any possible implementation thereof, or to implement the methods as in the second aspect or any possible implementation thereof, or to implement the methods as in the third aspect or any possible implementation thereof.

[0088] Optionally, the device may also include the transceiver circuit, or an input / output interface.

[0089] In a seventh aspect, a chip is provided, including processing circuitry for running a program or instructions to cause the method as described in the first aspect or any possible implementation thereof to be implemented, or to cause the method as described in the second aspect or any possible implementation thereof to be implemented, or to cause the method as described in the third aspect or any possible implementation thereof to be implemented.

[0090] Optionally, the chip may further include a memory for storing programs or instructions.

[0091] Optionally, the chip may also include transceiver circuitry, or input / output interfaces.

[0092] Eighth aspect, a computer-readable storage medium is provided, the computer-readable storage medium including instructions that, when executed by a processor, cause the method as in the first aspect or any possible implementation of the first aspect to be implemented, or cause the method as in the second aspect or any possible implementation of the second aspect to be implemented, or cause the method as in the third aspect or any possible implementation of the third aspect to be implemented.

[0093] Ninth aspect, a computer program product is provided, the computer program product comprising computer program code or instructions, which, when executed, cause the method of the first aspect and any possible implementation thereof to be implemented, or cause the method of the second aspect and any possible implementation thereof to be implemented, or cause the method of the third aspect and any possible implementation thereof to be implemented.

[0094] In a tenth aspect, a communication system is provided, the communication system including means for performing the first aspect and any possible implementation thereof, or including means for performing the second aspect and any possible implementation thereof, or including means for performing the third aspect and any possible implementation thereof.

[0095] It should be understood that the fourth to tenth aspects of this application correspond to the technical solutions of the first to third aspects of this application, and the beneficial effects obtained by each aspect and the corresponding feasible implementation are similar, and will not be repeated here. Attached Figure Description

[0096] Figure 1 is a schematic diagram of a communication system applicable to the communication method of this application embodiment;

[0097] Figure 2 is a schematic diagram of another communication system applicable to the communication method of this application embodiment;

[0098] Figure 3 is a schematic diagram of a possible application framework in a communication system;

[0099] Figure 4 is a schematic diagram of another possible application framework in a communication system;

[0100] Figure 5 is a schematic diagram of CSI feedback using an auto-encoder (AE) model provided in an embodiment of this application;

[0101] Figure 6 is a schematic diagram of a deep neural network (DNN);

[0102] Figure 7 shows an example of a neuron structure;

[0103] Figure 8 is a schematic diagram of the data set docking between the network side and the terminal side provided in an embodiment of this application;

[0104] Figure 9 is a schematic diagram of model interoperability between the network side and the terminal side provided in an embodiment of this application;

[0105] Figure 10 is a schematic flowchart of the communication method provided in an embodiment of this application;

[0106] Figure 11 is a schematic flowchart of the communication method provided in an embodiment of this application;

[0107] Figure 12 is a schematic diagram of the identifier associated with the model provided in an embodiment of this application;

[0108] Figure 13 is a schematic flowchart of the communication method provided in an embodiment of this application;

[0109] Figure 14 is a schematic block diagram of a communication device provided in an embodiment of this application;

[0110] Figure 15 is another schematic block diagram of the communication device provided in an embodiment of this application;

[0111] Figure 16 is a schematic block diagram of the AI ​​processor provided in an embodiment of this application. Detailed Implementation

[0112] The technical solutions in this application will now be described with reference to the accompanying drawings.

[0113] To facilitate understanding of the embodiments of this application, the following points will be explained first:

[0114] First, in this application, the terminal side can also be referred to as the UE (user equipment) side, including: terminal equipment, components deployed in the terminal equipment (such as circuits or chips inside the terminal equipment), equipment deployed outside the terminal equipment (such as OTT hosts or cloud servers), or components deployed in equipment outside the terminal equipment (such as circuits or chips inside the equipment). The network (NW) side includes: network equipment communicating with the terminal equipment, components deployed in the network equipment (such as circuits or chips with near real-time RAN intelligent control functions inside the network equipment), equipment deployed outside the network equipment (such as intelligent network elements, for example, intelligent network elements with near real-time RAN intelligent control functions), or components deployed in the intelligent network element (such as circuits or chips inside the intelligent network element). The network equipment may include: access network equipment, core network equipment, or operation administration and maintenance (OAM) equipment.

[0115] Second, in this application, the indication includes direct indication (also known as explicit indication) and indirect indication (also known as implicit indication). Directly indicating information A means including information A; indirectly indicating information A can mean indicating information A through the correspondence between information A and information B and by directly indicating information B; or by indicating information A through a preset rule that can be used to determine A based on B and by directly indicating information B. The correspondence between information A and information B, and the preset rule, can be predefined, pre-stored, pre-burned, or pre-configured.

[0116] Third, in this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates an "or" relationship between the preceding and following related objects, but it does not exclude the possibility of indicating an "and" relationship; the specific meaning can be understood in context. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c; a and b; a and c; b and c; or a and b and c. Here, a, b, and c can be single or multiple.

[0117] Fourth, in this application, the use of prefixes such as "first" and "second" is merely for the purpose of distinguishing and describing different things belonging to the same category, and does not constrain the order, size, or quantity of things. For example, "first information" and "second information" are simply different information, and do not limit the quantity of information, the order of transmission, or the relationship of priority.

[0118] Fifth, in this application, "send" and "receive" indicate the direction of signal transmission. For example, "send information to XX" can be understood as the destination of the information being XX, which can include direct transmission via the air interface or indirect transmission by other units or modules via the air interface. "Receive information from YY" can be understood as the source of the information being YY, which can include direct reception from YY via the air interface or indirect reception from YY by other units or modules via the air interface. "Send" can also be understood as the "output" of the chip interface, and "receive" can also be understood as the "input" of the chip interface. In other words, sending and receiving can occur between devices, such as between a terminal device and a computing node, or within a device, such as between components, modules, chips, software modules, or hardware modules within the device via a bus, wiring, or interface.

[0119] Sixth, in the embodiments of this application, "when," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a time, nor do they require the device to make a judgment action when it is implemented, nor do they mean that there are other limitations.

[0120] Seventh, in this application, the words "example," "exemplarily," "for example," or "such as" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "example," "exemplarily," "for example," or "such as" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "example," "exemplarily," "for example," or "such as" is intended to present the relevant concepts in a specific manner.

[0121] The technical solutions provided in this application can be applied to various communication systems, such as: 5th generation (5G) or new radio (NR) systems, long term evolution (LTE) systems, LTE frequency division duplex (FDD) systems, LTE time division duplex (TDD) systems, wireless local area network (WLAN) systems, satellite communication systems, future communication systems, or integrated systems of multiple systems. The technical solutions provided in this application can also be applied to device-to-device (D2D) communication, vehicle-to-everything (V2X) communication, machine-to-machine (M2M) communication, machine-type communication (MTC), and Internet of Things (IoT) communication systems or other communication systems.

[0122] In a communication system, one network element can send signals to or receive signals from another network element. These signals can include information, signaling, or data. The term "network element" can also be replaced by an entity, network entity, device, communication equipment, communication module, node, communication node, etc. This disclosure uses a network element as an example. For instance, a communication system can include at least one terminal device and at least one network device. The network device can send downlink signals to the terminal device, and / or the terminal device can send uplink signals to the network device. It is understood that the terminal device in this disclosure can be replaced by a first network element, and the network device can be replaced by a second network element, both performing the corresponding communication methods described in this disclosure.

[0123] Figure 1 is a schematic diagram of a communication system applicable to the communication method of this application embodiment. As shown in Figure 1, the communication system 100A may include at least one access network device, such as access network device 110 shown in Figure 1; the communication system 100A may also include at least one terminal device, such as terminal device 120 and terminal device 130 shown in Figure 1. Access network device 110 and terminal devices (such as terminal device 120 and terminal device 130) can communicate via a wireless link. The communication devices in this communication system, for example, access network device 110 and terminal device 120, can communicate via multi-antenna technology.

[0124] In wireless communication networks, such as mobile communication networks, the services supported by the networks are becoming increasingly diverse, thus requiring increasingly diverse demands. For example, networks need to support ultra-high speeds, ultra-low latency, and / or massive connectivity. This characteristic makes network planning, network configuration, and / or resource scheduling increasingly complex. Furthermore, as network functions become more powerful, such as supporting higher spectrum levels, supporting higher-order multiple-input multiple-output (MIMO) technologies, supporting beamforming (BF), and supporting beam management, network energy efficiency has become a hot research topic. These new demands, new scenarios, and new characteristics bring unprecedented challenges to network planning, operation, and efficient operation. To meet these challenges, artificial intelligence (AI) technology can be introduced into wireless communication networks to achieve network intelligence. To support AI technology in wireless networks, AI nodes may also be introduced.

[0125] Figure 2 is a schematic diagram of another communication system applicable to the communication method of this application embodiment. Compared with the communication system 100A shown in Figure 1, the communication system 100B shown in Figure 2 further includes an AI network element 140. The AI ​​network element 140 is used to perform AI-related operations, such as building training datasets or training AI models. The AI ​​network element can also be simply referred to as an intelligent network element.

[0126] In one possible implementation, access network device 110 can send data related to the training of the AI ​​model to AI network element 140, which then constructs a training dataset and trains the AI ​​model. For example, the data related to the training of the AI ​​model may include data reported by terminal devices. AI network element 140 can send the results of operations related to the AI ​​model to access network device 110, which then forwards them to the terminal devices. For example, the results of operations related to the AI ​​model may include at least one of the following: a trained AI model, model evaluation results, or test results. Exemplarily, a portion of the trained AI model may be deployed on access network device 110, and another portion on terminal devices 120 and / or 130. Alternatively, the trained AI model may be deployed on access network device 110. Or, the trained AI model may be deployed on terminal devices 120 and / or 130.

[0127] It should be understood that Figure 2 is only used as an example of the AI ​​network element 140 being directly connected to the access network device 110. In other scenarios, the AI ​​network element 140 can also be connected to the terminal device. Alternatively, the AI ​​network element 140 can be connected to both the access network device 110 and the terminal device simultaneously. Alternatively, the AI ​​network element 140 can also be connected to the access network device 110 through a third-party network element. This application embodiment does not limit the connection relationship between the AI ​​network element and other network elements. For example, the AI ​​network element 140 can also be set as a module in the access network device and / or the terminal device, for example, in the access network device 110 or the terminal device shown in Figure 1.

[0128] It should be noted that Figures 1 and 2 are simplified schematic diagrams for ease of understanding. For example, the communication system may also include other devices, such as wireless relay devices and / or wireless backhaul devices, which are not shown in Figures 1 and 2. In practical applications, the communication system may include multiple access network devices and multiple terminal devices. The embodiments of this application do not limit the number of access network devices and terminal devices included in the communication system.

[0129] In the embodiments of this application, the terminal device may also be referred to as UE, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent, or user equipment.

[0130] Terminal devices can be devices that provide voice / data, such as handheld devices with wireless connectivity, in-vehicle devices, etc. Currently, examples of terminals include: mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to a wireless modem, wearable devices, terminal devices in 5G networks, or future public land mobile communication networks. Terminal devices in a network (PLMN), etc., are not limited to this in the embodiments of this application.

[0131] By way of example and not limitation, in this embodiment, the terminal device can also be a wearable device. Wearable devices, also known as wearable smart devices, are a general term for devices that utilize wearable technology to intelligently design and develop everyday wearables, such as glasses, gloves, watches, clothing, and shoes. Wearable devices are portable devices that are worn directly on the body or integrated into the user's clothing or accessories. Wearable devices are not merely hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are feature-rich, large in size, and can achieve complete or partial functions without relying on a smartphone, such as smartwatches or smart glasses, as well as those that focus on a specific type of application function and require the use of other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0132] In this embodiment, the device for implementing the functions of the terminal device can be the terminal device itself, or it can be any device capable of supporting the terminal device in implementing those functions, such as a chip system. This device can be installed in or used in conjunction with the terminal device. In this embodiment, the chip system can be composed of chips or may include chips and other discrete components. This embodiment only uses the terminal device as an example to illustrate the device for implementing the functions of the terminal device, and does not constitute a limitation on the solution of this embodiment.

[0133] The access network device in this application embodiment can be a device used to communicate with terminal devices. This access network device can also be called a network device, such as a base station. In this application embodiment, the access network device can refer to a RAN node (or device) that connects the terminal device to the wireless network. A base station can broadly encompass, or be replaced by, various names such as: NodeB, evolved NodeB (eNB), next-generation NodeB (gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), master station, auxiliary station, motor slide retainer (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), radio unit (RU), positioning node, etc. A base station can be a macro base station, micro base station, relay node, donor node, or similar entities, or combinations thereof. A base station can also refer to a communication module, modem, or chip installed within the aforementioned equipment or apparatus. A base station can also be a mobile switching center, a device that performs base station functions in D2D, V2X, and M2M communications, or a device that performs base station functions in future communication systems. A base station can support networks using the same or different access technologies. Optionally, a RAN node can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network equipment in V2X technology can be a roadside unit (RSU). The embodiments of this application do not limit the specific technology or equipment form used in the access network equipment.

[0134] Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move depending on the location of the mobile base station. In other examples, a helicopter or drone can be configured as a device to communicate with another base station.

[0135] In some deployments, the access network equipment mentioned in the embodiments of this application may be a device including a CU, or a DU, or a device including both CU and DU, or a device with a control plane CU node (central unit-control plane (CU-CP)) and a user plane CU node (central unit-user plane (CU-UP)) and a DU node. For example, the access network equipment may include gNB-CU-CP, gNB-CU-UP, and gNB-DU.

[0136] In some deployments, multiple RAN nodes collaborate to assist terminals in achieving wireless access, with different RAN nodes each implementing some of the base station's functions. For example, RAN nodes can be CUs, DUs, CU-CPs, CU-UPs, or RUs. CUs and DUs can be configured separately or included in the same network element, such as a BBU. RUs can be included in radio frequency equipment or radio frequency units, such as RRUs, AAUs, or RRHs.

[0137] RAN nodes can support one or more types of fronthaul interfaces, each corresponding to a DU and RU with different functions. If the fronthaul interface between the DU and RU is a common public radio interface (CPRI), the DU is configured to implement one or more baseband functions, and the RU is configured to implement one or more radio frequency functions. If the fronthaul interface between the DU and RU is another type of interface, relative to CPRI, some downlink and / or uplink baseband functions, such as, for downlink, precoding, digital beamforming (BF), or one or more of inverse fast Fourier transform (IFFT) / cyclic prefix addition (CP), are moved from the DU to the RU; and for uplink, one or more of digital beamforming (BF), or fast Fourier transform (FFT) / cyclic prefix removal (CP), are moved from the DU to the RU. In one possible implementation, the interface can be an enhanced common public radio interface (eCPRI). Under the eCPRI architecture, the segmentation between DU and RU differs, corresponding to different categories (Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, and F.

[0138] Taking eCPRI Cat A as an example, for downlink transmission, layer mapping is used as the dividing line. DU is configured to implement one or more functions preceding layer mapping (i.e., coding, rate matching, scrambling, modulation, and layer mapping), while other functions following layer mapping (e.g., resource element (RE) mapping, digital beamforming (BF), or one or more functions in inverse fast Fourier transform (IFFT) / adding CP) are moved to RU. For uplink transmission, de-RE mapping is used as the dividing line. DU is configured to implement one or more functions preceding de-mapping (i.e., decoding, rate matching de-matching, descrambling, demodulation, inverse discrete Fourier transform (IDFT), channel equalization, and de-RE mapping), while other functions following de-mapping (e.g., digital BF or FFT / removing CP) are moved to RU. Understandably, the functional descriptions of the DU and RU corresponding to various types of eCPRI can be found in the eCPRI protocol, and will not be elaborated here.

[0139] In one possible design, the processing unit in the BBU used to implement baseband functions is called the baseband high (BBH) unit, and the processing unit in the RRU / AAU / RRH used to implement baseband functions is called the baseband low (BBL) unit.

[0140] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an open RAN (ORAN) architecture, CU can also be called open CU (open-CU, O-CU), DU can also be called open DU (open-DU, O-DU), CU-CP can also be called open CU-CP (open-CU-CP) O-CU-CP, CU-UP can also be called open CU-UP (open-CU-UP, O-CU-UP), and RU can also be called open RU (open-RU, O-RU). Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software modules and hardware modules.

[0141] In this embodiment, the apparatus for implementing the functions of a network device can be a network device itself; it can also be an apparatus capable of supporting the network device in implementing those functions, such as a chip system, hardware circuit, software module, or a hardware circuit plus a software module. This apparatus can be installed in the network device or used in conjunction with the network device. In this embodiment, the example of a network device being used to implement the functions of a network device is provided only and does not constitute a limitation on the solutions described in this embodiment.

[0142] Network devices and / or terminal devices can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; and they can also be deployed in the air on airplanes, balloons, and satellites. This application does not limit the scenario in which the network devices and terminal devices are located. Furthermore, terminal devices and network devices can be hardware devices, or software functions running on dedicated hardware or general-purpose hardware, such as virtualization functions instantiated on a platform (e.g., a cloud platform), or entities that include dedicated or general-purpose hardware devices and software functions. This application does not limit the specific form of the terminal devices and network devices.

[0143] Optionally, the AI ​​node can be deployed in one or more of the following locations within the communication system: access network equipment, terminal equipment, or core network elements. Alternatively, the AI ​​node can also be deployed independently, for example, in a location other than any of the aforementioned devices, such as in the host or cloud server of an over-the-top (OTT) system. The AI ​​node can communicate with other devices in the communication system, which can be, for example, one or more of the following: access network equipment, terminal equipment, or core network elements.

[0144] It is understood that this application does not limit the number of AI nodes. For example, when there are multiple AI nodes, they can be divided based on function, such as different AI nodes being responsible for different functions.

[0145] It can also be understood that AI nodes can be independent devices, or they can be integrated into the same device to achieve different functions. Alternatively, they can be network elements in hardware devices, software functions running on dedicated hardware, or virtualization functions instantiated on a platform (e.g., a cloud platform). This application does not limit the specific form of the aforementioned AI nodes.

[0146] AI nodes can be AI network elements or AI modules.

[0147] Figure 3 illustrates a possible application framework in a communication system. As shown in Figure 3, network elements in the communication system are connected via interfaces (e.g., next-generation (NG) interfaces, Xn interfaces) or air interfaces. The NG interface is the interface between the radio access network and the 5G core network. The Xn interface is the interface between access network devices, and the air interface is the interface between access network devices and terminal devices. These network element nodes, such as core network devices, access network nodes (RAN nodes), terminals, or one or more devices in the OAM, are equipped with one or more AI modules (only one is shown in Figure 3 for clarity). The access network node can be a single RAN node or can include multiple RAN nodes, for example, including CU and DU. The CU and / or DU can also be equipped with one or more AI modules. Optionally, the CU can be further divided into CU-CP and CU-UP. One or more AI models are configured in the CU-CP and / or CU-UP.

[0148] The AI ​​module is used to implement corresponding AI functions. AI modules deployed in different network elements can be the same or different. Depending on the parameter configuration, the AI ​​module can implement different functions. The AI ​​module model can be configured based on one or more of the following parameters: structural parameters (e.g., at least one of the following: number of neural network layers, neural network width, inter-layer connections, neuron weights, neuron activation function, or bias in the activation function), input parameters (e.g., type and / or dimension of input parameters), or output parameters (e.g., type and / or dimension of output parameters). The bias in the activation function can also be referred to as the neural network bias.

[0149] An AI module can have one or more models. A model can infer an output, which includes one or more parameters. The learning, training, or inference processes of different models can be deployed on different nodes or devices, or they can be deployed on the same node or device.

[0150] Network devices can be network devices equipped with one or more AI modules. These network devices can include one or more devices from the core network equipment, access network nodes (RAN nodes), or operation administration and maintenance (OAM) systems shown in Figure 3. For example, the AI ​​module can be a RAN intelligent controller (RIC) as shown in Figure 4, such as a near-real-time RIC (near-RT RIC) or a non-real-time RIC (non-RT RIC). For instance, a near-real-time RIC is located in a RAN node (e.g., in a CU or DU), while a non-real-time RIC is located in OAM, a cloud server, a core network device, or other access network devices. The RIC can obtain subsets from multiple terminal devices from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU), reassemble them into a training dataset, and train based on this dataset. Exemplarily, near-real-time RICs and non-real-time RICs can also be configured as separate network elements, and the access network device can be either a near-real-time RIC or a non-real-time RIC.

[0151] Figure 4 illustrates another possible application framework in a communication system. In addition to access network nodes (CU, DU, and RU are shown in the figure) and terminals, the communication system shown in Figure 4 also includes an RIC (Regulator-Integrated Circuit). For example, the RIC could be the AI ​​module shown in Figure 3, which can be used to implement AI-related functions. The RIC includes near-real-time RICs and non-real-time RICs. Non-real-time RICs primarily process non-real-time information, such as data that is not sensitive to latency, with latency on the order of seconds. Real-time RICs primarily process near-real-time information, such as data that is relatively sensitive to latency, with latency on the order of tens of milliseconds.

[0152] The near real-time RIC is used for model training and inference. For example, it can be used to train an AI model and then use that AI model for inference. The near real-time RIC can obtain network-side and / or terminal-side information from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU) and / or terminals. This information can be used as training data or inference data. Optionally, the near real-time RIC can deliver inference results to RAN nodes and / or terminals. Optionally, inference results can be exchanged between CU and DU, and / or between DU and RU. For example, the near real-time RIC delivers the inference result to the DU, and the DU sends it to the RU.

[0153] The non-real-time RIC is also used for model training and inference. For example, it can be used to train an AI model and then use that model for inference. The non-real-time RIC can obtain network-side and / or terminal-side information from RAN nodes (e.g., one or more of CU, CU-CP, CU-UP, DU, or RU) and / or terminals. This information can be used as training data or inference data, and the inference results can be delivered to the RAN nodes and / or terminals. Optionally, inference results can be exchanged between CU and DU, and / or between DU and RU; for example, the non-real-time RIC delivers the inference results to the DU, which then forwards them to the RU.

[0154] The near real-time RIC and non-real-time RIC can also be set up as separate network elements. Optionally, the near real-time RIC and non-real-time RIC can also be part of other devices. For example, the near real-time RIC can be set in the RAN node (e.g., in CU, DU), while the non-real-time RIC can be set in the OAM, cloud server, core network device, or other network device.

[0155] To facilitate understanding of the embodiments of this application, the terms involved in this application will be briefly explained below.

[0156] 1. AI Model: A function model that maps an input of a certain dimension to an output of a certain dimension. Its parameters can be obtained through machine learning (ML). For example, f(x) = ax 2 +b is a quadratic function model, which can be viewed as an AI model. a and b correspond to the parameters of the model and can be obtained through machine learning training.

[0157] 2. Autoencoder (AE) model: This generally refers to a network structure consisting of two AI models, such as an encoder and a decoder. Each model can be an independent AI model. AE models are also called bilateral models, two-end models, collaborative models, etc. The encoder and decoder of an AE are usually trained together and can be used in a coordinated manner.

[0158] For example, CSI feedback can be implemented based on an AI model of AE. Figure 5 is a schematic diagram of CSI feedback using an AE model provided in an embodiment of this application.

[0159] As shown in the figure, the terminal side compresses the target CSI using an encoder, while the network side reconstructs the CSI using a decoder. For example, the terminal side can use the target CSI (i.e., V) as input to the encoder, which compresses the target CSI to obtain the compressed CSI (i.e., C). The terminal side can quantize the compressed CSI to obtain CSI feedback information. The terminal side can then send this CSI feedback information to the network side, for example, through a CSI report.

[0160] The network side can first dequantize the CSI feedback information to obtain compressed CSI with quantization loss (i.e., The network side can use this compressed CSI as input to the decoder, which can then decompress the CSI to obtain the reconstructed CSI. ).

[0161] The quantizer used to perform quantization can be predefined, such as protocol predefined, or it can be indicated by the network side; this application does not limit this.

[0162] In another implementation, the encoder on the terminal side can also compress and quantize the target CSI, outputting CSI feedback information. The decoder on the network side can also dequantize and decompress the CSI feedback information to obtain the reconstructed CSI.

[0163] It should be noted that the encoder on the terminal side can be deployed inside the terminal device or in other devices outside the terminal device, such as the aforementioned OTT host or cloud server; the decoder on the network side can be deployed inside the network device or in other devices outside the network device, such as the aforementioned smart network element.

[0164] It should be understood that although the figure shows an encoder and a decoder, this is only a model division from a functional perspective. The encoder can also be called the first model, and the decoder can also be called the second model. Furthermore, this application does not limit the number of models included in the AE model.

[0165] It should also be understood that the AE model is only one possible model for achieving the above functions and should not constitute any limitation on this application. The AE model can also be replaced by other AI models that can achieve the same or similar functions.

[0166] In this embodiment, the meaning of CSI is broader than that in traditional schemes. It is not limited to channel quality indication (CQI), precoding matrix indicator (PMI), rank indicator (RI), or CSI-RS resource indicator (CRI). It can also be one or more of the following: channel response information (such as channel response matrix), channel matrix, channel feature matrix, precoding matrix, reference signal receiving power (RSRP), signal to interference plus noise ratio (SINR), the identity (ID) of the optimal beam, or the ID of the top K beams. The signal to interference plus noise ratio can also be called the signal-to-interference-plus-noise ratio. The optimal beam refers to the beam that maximizes the received or transmitted energy. For example, if the receiver uses different receiving beams to receive signals, the optimal beam may include the beam with the largest RSRP among the signals received from multiple different receiving beams. For example, if the transmitter uses different transmit beams to send signals, the optimal beam can include the beam with the highest RSRP (Receiving Power Ratio) of the signal when it arrives at the receiver. Similarly, the optimal top K beams refer to the top K beams that maximize the received or transmitted energy.

[0167] 3. Model Development: Model development refers to building and training models using AI technology to solve specific inference tasks. For example, model development may include one or more of the following: model training, model adaptation, or model enhancement. Model training includes one or more of the following: initial model training, model retraining, model fine-tuning, and model updating. Model adaptation refers to adapting the model format, i.e., converting the received model parameters and / or model files into a model format executable by the local device for inference on that device. Model enhancement refers to enhancing the model's functionality through model structure optimization and training, such as adding other functions or neural networks to the model's existing functions, so that the enhanced model can be applied to more scenarios or richer needs. Model deployment refers to applying the developed model to real-world scenarios.

[0168] Model development can also be represented as offline engineering in the standard.

[0169] It should be understood that the terms such as model development, model training, model adaptation, and model enhancement used in this document are introduced only to facilitate understanding of offline projects. In specific implementations, the operations involved by these terms may not be clearly distinguished. The operations involved by each term may include, but are not limited to, the contents listed in this document. Alternatively, offline projects may also include some of the operations listed above, or they may include other operations. This application does not limit these operations.

[0170] 4. Model Training: By selecting an appropriate loss function, the model parameters are trained using optimization algorithms to minimize the difference between the model's predicted values ​​and the ground truth (or target values, labels).

[0171] The model training methods involved in this application include supervised learning, self-supervised learning, and knowledge distillation. These methods will be explained in detail below.

[0172] Supervised learning, also known as supervised instruction, involves using machine learning algorithms to learn the mapping relationship between sample values ​​and labels based on collected sample values ​​and labels. This learned mapping relationship is then expressed using a machine learning model. The process of training the machine learning model is the process of learning this mapping relationship. For example, in signal detection, the noisy received signal is the sample, and the corresponding real constellation point is the label. Machine learning aims to learn the mapping relationship between samples and labels through training, that is, to enable the machine learning model to learn a signal detector. During training, the model parameters are optimized by calculating the error between the model's predicted values ​​and the real labels. Once the mapping relationship is learned, it can be used to predict the sample label of each new sample. The mapping relationship learned in supervised learning can include linear and nonlinear mappings. Based on the type of label, the learning task can be divided into classification tasks and regression tasks.

[0173] Self-supervised learning is a type of unsupervised learning. Unsupervised learning relies on collected sample values ​​to allow algorithms to discover inherent patterns within the samples themselves. Self-supervised learning uses the samples themselves as supervisory signals; that is, the model learns the mapping relationship from sample to sample. During training, model parameters are optimized by calculating the error between the model's predicted values ​​and the actual samples. Self-supervised learning can be used in signal compression and decompression recovery applications; common algorithms include autoencoders and generative adversarial networks.

[0174] Knowledge distillation: Generally, large models are often single complex networks or collections of networks, possessing excellent performance and generalization ability, while small models, due to their smaller network size, have limited expressive power. Therefore, the knowledge learned by the large model can be used to guide the training of the small model; this process is called knowledge distillation. Knowledge distillation can enable small models to achieve performance comparable to large models, but with fewer parameters and shorter inference latency, thus achieving model compression and acceleration. Furthermore, directly training small models with massive amounts of data often does not yield good performance, while training large models with massive amounts of data and then using the large model to perform knowledge distillation on the small model can achieve better continuation results. In addition, knowledge distillation can also be used to integrate and transfer datasets from different domains.

[0175] Knowledge distillation employs a teacher-student model, where a teacher model assists in training a student model. The teacher model is a complex, large model, while the student model is a simple, small model. Because the teacher model has strong learning capabilities, it can transfer the knowledge it learns to the relatively weaker student model, thereby enhancing the student model's generalization ability. The complex, cumbersome, but effective teacher model remains offline, simply acting as a mentor; the flexible and lightweight student model is the one actually deployed for prediction tasks.

[0176] 5. Model Files and Model Parameters: Model files and / or model parameters can be used to define the model. Model files can indicate the model structure, which may include, but is not limited to, feedforward neural networks (FNNs), convolutional neural networks (CNNs), and recurrent neural networks (RNNs). The model file can have a fixed format, such as a standard predefined format or a format pre-negotiated by both ends of the connection. Model parameters can refer to parameters in the neural network model, such as, but not limited to, the number of layers in the neural network, the type and weights of neurons in each layer, etc. This application does not limit the method of distributing the parameters of the reference model.

[0177] Take DNN as an example. The idea behind DNN comes from the neuronal structure of the brain. Each neuron can perform a weighted summation operation on its inputs and then use the result of the weighted summation operation to generate an output through a nonlinear function. Figure 6 shows an example of a neuron structure. The input of the neuron shown in Figure 6 is x = [x0 x1 … x N-1 The weights corresponding to the inputs are w = [w0 w1 … w] N-1The bias of the weighted summation is b. The nonlinear function f() can take many forms; for example, the nonlinear function f() can be the maximum value function max{0, x}. Then the effect of a neuron's execution is... Where N is a positive integer; n is a positive integer greater than or equal to 0 and less than or equal to (N-1).

[0178] A DNN typically has multiple neural network layers, including an input layer, one or more hidden layers, and an output layer. Generally, the first layer is the input layer, the last layer is the output layer, and the layers in between are hidden layers. Each layer contains multiple neurons. Layers are fully connected; that is, any neuron in the i-th layer is connected to any neuron in the (i+1)-th layer. The input layer processes the received numerical values ​​(i.e., the DNN's input) through neurons and then passes them to the hidden layers. Similarly, the hidden layers pass the computation results to the final output layer, producing the DNN's output. Figure 7 shows an example of a DNN. The DNN model shown in Figure 7 has three neural network layers: an input layer, a hidden layer, and an output layer.

[0179] It should be understood that the examples in conjunction with Figures 6 and 7 above are shown for ease of understanding only and should not constitute any limitation on this application. This application does not limit the structure and parameters used in the AI ​​model.

[0180] One of the model structure or model parameters can be predefined, while the other can be sent by the sender (e.g., the network side). Alternatively, both the model structure and model parameters can be sent by the sender (e.g., the network side). This application does not impose any restrictions on this.

[0181] In this embodiment of the application, sending a model may refer to sending a model file and / or model parameters, and receiving a model may refer to receiving a model file and / or model parameters.

[0182] 6. Two-way connection: This refers to the connection between the sender (e.g., the network side) and the receiver (e.g., the terminal side). It can be a connection of datasets or a connection of models. A dataset is used in machine learning for model training, validation, and testing; the quantity and quality of the dataset affect the effectiveness of machine learning. In this embodiment, the dataset can be used for model training.

[0183] The following section uses CSI compressed inference tasks as an example to explain the connection between the dataset and the model.

[0184] Dataset integration mainly refers to the sender providing a dataset to the receiver for model training. Figure 8 is a schematic diagram of dataset integration between the network side and the terminal side.

[0185] As shown in Figure 8, the network can first acquire the dataset through joint training. For example, the network can use a virtual CSI compression model (e.g., an encoder) to compress the target CSI, and then use a CSI reconstruction model (e.g., a decoder) to decompress the target CSI. That is, the target CSI is the input to the CSI compression model, and the output of the CSI compression model can be the compressed CSI. This compressed CSI can be the input to the CSI reconstruction model, and the CSI reconstruction model can output the reconstructed CSI. In this way, the dataset can include one or more of the following: target CSI or reconstructed CSI.

[0186] Optionally, the compressed CSI can be further quantized to obtain CSI feedback information, which can be used for CSI report submission in specific implementations. Accordingly, the dataset may also include one or more of the following: target CSI, CSI feedback information, or reconstructed CSI. The quantizer used to quantize the compressed CSI can be predefined, such as protocol predefined, or it can be issued by the sender. Optionally, the dataset also includes a quantizer. The quantizer can also be replaced by a quantization method. Optionally, the quantizer can be included in the CSI compression model; in other words, the CSI compression model can also have the function of quantizing the compressed CSI.

[0187] Optionally, if the compressed CSI has been quantized to obtain CSI feedback information, the CSI feedback information can be dequantized before being input into the CSI reconstruction model to obtain the input of the CSI reconstruction model. The dequantizer used to dequantize the CSI feedback information can be predefined, such as protocol predefined, or it can be issued by the sender. Optionally, the dequantizer can be included in the CSI reconstruction model; in other words, the CSI reconstruction model can also have the function of dequantizing the CSI feedback information.

[0188] The network side can distribute this dataset to the terminal side. The terminal side can then train a model based on the received dataset to obtain a CSI compressed model for the terminal side.

[0189] Model interoperability mainly refers to the sender providing model files and / or model parameters to the receiver for model deployment, development (e.g., model training), and deployment. In this embodiment, the sender can provide model files and / or model parameters to the receiver for model training. Figure 9 is a schematic diagram of model interoperability between the network side and the terminal side.

[0190] As shown in Figure 9, the network side can first obtain the CSI compressed model through joint training. The specific process of network-side joint training can be found in the explanation above combined with Figure 8, and will not be repeated here. The network side can then distribute the CSI compressed model or CSI reconstructed model obtained through joint training to the terminal side. For example, the network side can send the model file and / or model parameters of the CSI compressed model to the terminal side, or send the model file and / or model parameters of the CSI reconstructed model to the terminal side. The terminal side can then train the model based on the received model file and / or model parameters.

[0191] It should be understood that the CSI compression model shown above in conjunction with Figures 8 and 9 can be an encoder, and the CSI reconstruction model can be a decoder.

[0192] 7. Dataset: Data used for model training, validation, or testing in machine learning. The quantity and quality of the data affect the effectiveness of machine learning. Training data can include the input of the AI ​​model, or the input and target output of the AI ​​model. The target output is the target value of the AI ​​model's output, also known as the ground truth, comparison ground truth, label, or labeled sample.

[0193] 8. AI CSI Model:

[0194] AI CSI models / functions / tasks / scenarios include one or more of the following: AI CSI spatial frequency domain compression, AI CSI spatial time frequency domain compression, AI CSI prediction, AI CSI multi-slot spatial time frequency domain compression, AI CSI prediction + single-slot CSI compression, or AI CSI prediction + multi-slot CSI compression.

[0195] AI CSI spatial-frequency domain compression can be understood as the joint compression of CSI in both the spatial and frequency domains based on an AI model. AI CSI spatial-temporal-frequency domain compression can be understood as the joint compression of CSI in the spatial, frequency, and time domains based on an AI model, or, based on an AI model utilizing time-domain correlation, the joint compression of CSI in both the spatial and frequency domains. AI CSI prediction can be understood as predicting CSI based on an AI model. AI CSI multi-slot spatial-temporal-frequency domain compression can be understood as the joint compression of CSI in the spatial, frequency, and time domains based on an AI model utilizing time-domain correlation. AI CSI prediction + single-slot CSI compression can be understood as predicting CSI based on an AI model and compressing the predicted and / or measured single-slot CSI. AI CSI prediction + multi-slot CSI compression can be understood as predicting CSI based on an AI model and compressing the predicted and / or measured multi-slot CSI. Multi-slot CSI compression and / or spatial-temporal-frequency domain compression can utilize the time-domain correlation of CSI.

[0196] For example, the AI-based CSI feedback process is as follows: The CSI is input to the CSI compression model on the terminal device side. The terminal device side CSI compression model processes the input CSI and outputs CSI feedback bits. These CSI feedback bits are then input to the CSI reconstruction model on the network device side. The network device side CSI reconstruction model processes the input CSI feedback bits and outputs the reconstructed CSI. The CSI input to the terminal device side CSI compression model may include measured CSI and / or predicted CSI; the CSI feedback bits output by the terminal device side CSI compression model may include CSI feedback information; and the reconstructed CSI output by the network device side CSI reconstruction model may include the recovery result of the CSI feedback information.

[0197] It should be understood that the time slot mentioned in the embodiments of this application can be replaced with a time unit or moment. A time unit or moment can be understood as any one or more of a time slot, subframe, frame, or orthogonal frequency division multiplexing (OFDM) symbol. For example, the moment corresponding to A represents the time slot, subframe, frame, or OFDM symbol in which A is located, or the first time slot, subframe, frame, or OFDM symbol in which A is located, or the last time slot, subframe, frame, or OFDM symbol in which A is located.

[0198] In practical implementation, it is necessary to provide models and / or datasets to devices for training models and / or deployment modes for model training. Therefore, how to transmit models and / or datasets to reduce air interface overhead has become a technical problem that urgently needs to be solved.

[0199] In view of this, this application provides a communication method that enables the receiving end of a model to determine whether the model parameters and / or datasets corresponding to two received models can be used for training the same model, thereby facilitating the satisfaction of training requirements for different models while transmitting fewer model parameters and / or datasets.

[0200] The communication method provided in this application embodiment will be described below with reference to the accompanying drawings, taking the interaction between various devices as an example.

[0201] It should be noted that the first or second device in the following embodiments can be replaced by a device on the terminal device side, a device on the network device side, or a device on the core network (CN) side.

[0202] The devices on the terminal device side can include the terminal device itself, the communication module within the terminal device, or the circuits or chips within the terminal device responsible for communication functions (such as a modem chip, also known as a baseband chip, or a SoC chip or SIP chip containing a modem core, etc.), or can be included in the AI ​​entities on the terminal device side. The AI ​​entities on the terminal device side can be the terminal device itself, or they can be AI entities that serve the terminal device, such as servers, such as OTT servers or cloud servers.

[0203] Network device-side equipment can include the network device itself, communication modules within the network device, or circuits or chips within the network device responsible for communication functions (such as modem chips, also known as baseband chips, or system-on-a-chip (SoC) chips or SIP chips containing modem cores, etc.), or it can include AI entities on the network device side. AI entities on the network device side can be the network device itself, or AI entities serving the network device, such as RICs, OAMs, or servers, such as OTT servers or cloud servers.

[0204] The equipment on the core network side may include the core network element itself, functional modules within the core network element, or circuits or chips within the core network element, or may include AI entities on the core network side. The AI ​​entities on the core network side can be the core network element itself, or AI entities serving the core network element, such as servers, like OTT servers or cloud servers. The method provided in this application will be described in detail below with reference to the accompanying drawings.

[0205] It should also be noted that the model parameters mentioned in this application may include all the model parameters, or may include only some of the model parameters. If the model parameters include only some of the model parameters, then the model parameters may be predefined, or indicated by the first device to the second device. For example, if the model parameters include only some of the model parameters, then the model parameters may include weights.

[0206] It should also be noted that the dataset corresponding to the model mentioned in this application may include the entire dataset corresponding to the model, or it may include a portion of the dataset corresponding to the model. If the dataset corresponding to the model includes a portion of the dataset corresponding to the model, then the dataset corresponding to the model may be predefined, or it may be indicated by the first device to the second device. For example, if the dataset corresponding to the model includes some of the parameters corresponding to the model, the dataset corresponding to the model may include the training dataset corresponding to the model.

[0207] Figure 10 is a schematic flowchart of a communication method 1000 provided in an embodiment of this application. The steps of method 1000 will be described in detail below.

[0208] S1010, the second device sends the first request message.

[0209] Accordingly, the first device receives the first request message.

[0210] The first request message is used to request at least one dataset for model training. Alternatively, the first request message is used to request at least one model for model training.

[0211] For example, the first request message is used to request at least one dataset for a model having a first function, and the at least one dataset is used to train the model having the first function. Alternatively, the first request message is used to request a model having the first function. The model having the first function may be used for predicting CSI, or for beam management, or for CSI compression, or for localization, etc., and this application does not limit this to any particular function.

[0212] For example, if a model with a first function is used for CSI compression, the first function can be one of the following: spatial frequency domain compression, spatial temporal frequency domain compression, CSI prediction + single-slot CSI compression, or CSI prediction + multi-slot CSI compression. Further descriptions of these functions can be found in the terminology section above. Table 1 shows the information used by models with different first functions.

[0213] Table 1

[0214] Wherein, if the model with the first function is deployed on the terminal device side, the target CSI refers to the output of the model with the first function, and the historically obtained CSI used by the UE is the input of the model with the first function. If the model with the first function is deployed on the network device side, the target CSI refers to the input of the model with the first function, and the historically obtained CSI used by the network is the input of the model with the first function.

[0215] As shown in Table 1, case 0 corresponds to a model with spatial frequency domain compression function, case 2 corresponds to a model with spatial temporal frequency domain compression function, and case 3 corresponds to a model with CSI prediction + single-slot CSI compression function or with CSI prediction + multi-slot CSI compression function.

[0216] It should be noted that S1010 is an optional step.

[0217] S1020, the first device sends the first model.

[0218] Correspondingly, the second device receives the first model.

[0219] The first model is used for model training. In S1020, the first model sent by the first device may include a first model file and / or first model parameters corresponding to the first model. For more details about the model file and model parameters, please refer to the terminology section above.

[0220] Optionally, in S1020, the first device also sends a first model structure identifier, which is used to identify the model structure of the first model.

[0221] Optionally, in S1020, the first device also sends a first dataset corresponding to the first model, which is used for model training.

[0222] S1030, the first device sends the second model.

[0223] Correspondingly, the second device receives the second model.

[0224] The second model is used for model training. In S1030, the second model sent by the first device may include the second model file and / or the second model parameters corresponding to the second model.

[0225] Optionally, in S1030, the first device also sends a second model structure identifier, which is used to identify the model structure of the second model.

[0226] Optionally, in S1030, the first device also sends a second dataset corresponding to the second model, which is used for model training.

[0227] For example, if method 1000 executes S1010, the first device may, in response to the first request message, send the first model and the second model to the second device.

[0228] For example, if method 1000 does not execute S1010, the first device may actively send the first model and the second model to the second device.

[0229] S1040, the first device sends the first instruction information.

[0230] Correspondingly, the second device receives the first instruction information.

[0231] The first indication information is used to determine the second information, which includes one or more of the following: the first model parameters and the second model parameters can be used to train the same model, or the first model parameters and the second model parameters cannot be used to train the same model; or the first dataset corresponding to the first model and the second dataset corresponding to the second model can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

[0232] For example, if the model structure of the first model is the same as that of the second model, the second information includes one or more of the following: the parameters of the first model and the parameters of the second model can be used for training the same model, or the parameters of the first model and the parameters of the second model cannot be used for training the same model; or the first dataset and the second dataset can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model.

[0233] For example, if the model structure of the first model is different from that of the second model, the second information includes: the first dataset and the second dataset can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

[0234] Optionally, method 1000 may further include S1050 and / or S1060. For example, if the model structure of the first model is the same as the model structure of the second model, then method 1000 may include S1050 and S1060. Or, for example, if the model structure of the first model is different from the model structure of the second model, then method 1000 may include S1060.

[0235] S1050, the second device determines whether the first model parameters and the second model parameters can be used for training the same model.

[0236] For example, when the second device receives the first model, the second model, and the first instruction information, it can determine whether the parameters of the first model and the parameters of the second model can be used for training the same model based on the first instruction information.

[0237] Optionally, if the second device determines that the model structure of the first model is the same as the model structure of the second model, the second device may execute S1050. For example, if the second device receives a first model structure identifier and a second model structure identifier, the second device can determine whether the model structure of the first model is the same as the model structure of the second model based on the first model structure identifier and the second model structure identifier.

[0238] S1060, the second device determines whether the first dataset and the second dataset can be used to train the same model.

[0239] For example, when the second device receives the first dataset, the second dataset, and the first instruction information, it can determine whether the first dataset and the second dataset can be used for training the same model based on the first instruction information.

[0240] Optionally, method 1000 also includes S1070.

[0241] S1070, the second device performs model training.

[0242] In one possible implementation, if the second device determines that the first model parameters and the second model parameters can be used to train the same model, then the second device trains the same model based on the first model parameters and the second model parameters. For example, the second device trains the first model based on the first model parameters and the second model parameters.

[0243] In one possible implementation, if the second device determines that the first model parameters and the second model parameter set cannot be used to train the same model, then the second device trains different models based on the first model parameters and the second model parameters, respectively. For example, the second device trains the first model based on the first model parameters, and / or trains the second model based on the second model parameters.

[0244] In one possible implementation, if the second device determines that the first dataset and the second dataset can be used to train the same model, then the second device trains the same model based on the first dataset and the second dataset. For example, the second device trains the first model based on the first dataset and the second dataset.

[0245] In one possible implementation, if the second device determines that the first dataset and the second dataset cannot be used to train the same model, then the second device trains different models based on the first dataset and the second dataset, respectively. For example, the second device trains a first model based on the first dataset, and / or trains a second model based on the second dataset. Here, the first model and the second model are different.

[0246] In this embodiment of the application, if the first device sends multiple models to the second device, the first device can send instruction information to the second device so that the second device can determine whether the model parameters and / or datasets corresponding to at least two of the multiple models can be used for training the same model. This is beneficial to save the transmission overhead of model parameters and / or datasets, and enable the second device to perform different training tasks based on the received multiple models, thereby ensuring the performance of the trained or updated models.

[0247] For example, if the first dataset sent by the first device is used for training the first model, and the second dataset sent by the first device is used for training the second model, then if the second device determines, based on the received first instruction information, that the first dataset and the second dataset can be used to train the same model, it can train the first model based on the first dataset and the second dataset. This is equivalent to ensuring the performance of the first model after training while increasing the dataset used to train the first model.

[0248] For another example, if the second device can train the same model based on at least two previously received datasets after cell handover, thereby meeting the model requirements after cell handover, then the first device does not need to send a new model to the second device after the second device performs cell handover, thus saving data transmission requirements.

[0249] It should be understood that the second device determines the second information in different ways based on the first instruction information, depending on the different first instruction information.

[0250] The method by which the second device determines the second information based on different first instruction information is described below.

[0251] In implementation method 1, the first indication information is used to indicate one or more of the following: the first model parameters and the second model parameters can be used to train the same model, or the first model parameters and the second model parameters cannot be used to train the same model; or the first dataset and the second dataset can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

[0252] For example, if the first indication information is only used to indicate whether the first model parameters and the second model parameters can be used to train the same model, then the first indication information can be a 1-bit message. If the first indication information takes the first value, it means that the first indication information is used to indicate that the first model parameters and the second model parameters cannot be used to train the same model; if the first indication information takes the second value, it means that the first indication information is used to indicate that the first model parameters and the second model parameters can be used to train the same model. The first value is "0" and the second value is "1"; or, the first value is "1" and the second value is "0".

[0253] For example, if the first indication information is only used to indicate whether the first dataset and the second dataset can be used to train the same model, then the first indication information can be a 1-bit piece of information. If the first indication information takes the first value, it means that the first indication information is used to indicate that the first dataset and the second dataset cannot be used to train the same model; if the first indication information takes the second value, it means that the first indication information is used to indicate that the first dataset and the second dataset can be used to train the same model. The first value is "0" and the second value is "1"; or, the first value is "1" and the second value is "0".

[0254] For example, if the first indication information is used to indicate whether the first dataset and the second dataset can be used to train the same model, and to indicate whether the first model parameters and the second model parameters can be used to train the same model, then the first indication information can be a 2-bit information. If the first indication information takes the third value, it means that the first indication information is used to indicate that the first dataset and the second dataset cannot be used to train the same model, and to indicate that the first model parameters and the second model parameters cannot be used to train the same model; if the first indication information takes the fourth value, it means that the first indication information is used to indicate that the first dataset and the second dataset can be used to train the same model, and to indicate that the first model parameters and the second model parameters cannot be used to train the same model; if the first indication information takes the fifth value, it means that the first indication information is used to indicate that the first dataset and the second dataset cannot be used to train the same model, and to indicate that the first model parameters and the second model parameters can be used to train the same model; if the first indication information takes the fourth value, it means that the first indication information is used to indicate that the first dataset and the second dataset can be used to train the same model, and to indicate that the first model parameters and the second model parameters can be used to train the same model. For example, the third value is "00", the fourth value is "01", the fifth value is "10", and the sixth value is "11".

[0255] Implementation Method 2: The first indication information indicates rule #1. Rule #1 is related to the model's identifier. Rule #1 is used to determine whether the model parameters and / or datasets corresponding to at least two models can be used or cannot be used for training the same model.

[0256] It should be understood that if the first instruction information indicates rule #1, then in S1020, the first device also sends the identifier of the first model to the second device, and in S1030, the first device also sends the identifier of the second model to the second device.

[0257] For example, rule #1 is: if the first indication information includes the identifiers of multiple models, then the model parameters and / or datasets corresponding to the multiple models can be used for training the same model.

[0258] Based on rule #1 above, if the first indication information includes the identifier of the first model and the identifier of the second model, the second device determines, according to the first indication information, that the parameters of the first model and the parameters of the second model can be used for training the same model, and / or determines that the first dataset and the second dataset can be used for training the same model; if the first indication information does not include the identifier of the first model and / or the identifier of the second model, the second device determines, according to the first indication information, that the parameters of the first model and the parameters of the second model cannot be used for training the same model, and / or determines that the first dataset and the second dataset cannot be used for training the same model.

[0259] For example, rule #1 states: if the first indication information includes the identifiers of multiple models, then the model parameters and / or datasets corresponding to the multiple models cannot be used for training the same model.

[0260] Based on rule #1 above, if the first indication information includes the identifier of the first model and the identifier of the second model, then the second device determines, according to the first indication information, that the parameters of the first model and the parameters of the second model cannot be used for training the same model, and / or determines that the first dataset and the second dataset cannot be used for training the same model; if the first indication information does not include the identifier of the first model and / or the identifier of the second model, then the second device determines, according to the first indication information, that the parameters of the first model and the parameters of the second model can be used for training the same model, and / or determines that the first dataset and the second dataset can be used for training the same model.

[0261] For example, rule #1 states that model parameters and / or datasets corresponding to models with different identifiers can be used to train the same model. In other words, the first indication information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers can be used to train the same model.

[0262] Based on rule #1 above, the second device determines, according to the first instruction information, that the first model parameters and the second model parameters can be used for training the same model, and / or determines that the first dataset and the second dataset can be used for training the same model.

[0263] For example, rule #1 states that model parameters and / or datasets corresponding to models with different identifiers cannot be used to train the same model. In other words, the first indication information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers cannot be used to train the same model.

[0264] Based on rule #1 above, the second device determines, according to the first instruction information, that the first model parameters and the second model parameters cannot be used for training the same model, and / or that the first dataset and the second dataset cannot be used for training the same model.

[0265] For example, rule #1 is: the identifiers of at least two models whose corresponding model parameters and / or datasets can be used for training the same model must satisfy condition #1. In other words, the first indication information is used to indicate that the identifiers of at least two models whose corresponding model parameters and / or datasets can be used for training the same model must satisfy condition #1.

[0266] This application does not limit condition #1. For example, condition #1 is that the difference between the identifiers of two different models does not exceed threshold #1. Another example is that condition #1 is that the sum of the identifiers of two different models is a predefined value.

[0267] Based on rule #1 above, if the identifier of the first model and the identifier of the second model satisfy condition #1, then the second device determines that the parameters of the first model and the parameters of the second model can be used for training the same model, and / or, the first dataset and the second dataset can be used for training the same model; if the identifier of the first model and the identifier of the second model do not satisfy condition #1, then the second device determines that the parameters of the first model and the parameters of the second model cannot be used for training the same model, and / or, the first dataset and the second dataset cannot be used for training the same model.

[0268] Implementation method 3, the first indication information indicates rule #2, rule #2 is related to the values ​​of the features included in the first feature set of the model, rule #2 is used to determine whether the model parameters corresponding to at least two models can be used or cannot be used for training the same model.

[0269] For example, the first feature set of the model includes features related to one or more of the following: the model training configuration parameters corresponding to the model, the model training hyperparameters corresponding to the model, and the configuration parameters of the training dataset on which the model training is based.

[0270] For example, the first feature set includes one or more of the following: learning rate (hereinafter referred to as feature #1), dropout ratio (hereinafter referred to as feature #2), total number of training epochs (hereinafter referred to as feature #3), batch size (hereinafter referred to as feature #4), training set size (hereinafter referred to as feature #5), neural network dimension, and number of neural network layers.

[0271] For example, the learning rate represents the magnitude of model weight adjustment in each iteration during AI model training, i.e., the model's responsiveness to estimation errors. The dropout ratio refers to the proportion of neurons (or their connections) that are randomly and temporarily discarded from the network during deep learning network training. The total number of training epochs refers to the number of times the model traverses the entire training dataset. The batch size refers to the number of samples the model processes in each training iteration. The size of the training dataset refers to the number of data samples used for model training. Neural network dimension typically refers to the number of neurons in the input, hidden, and output layers of a neural network, and the connections between them. The number of neural network layers refers to the number of layers in a neural network, including the input, hidden, and output layers.

[0272] It should be understood that if the first instruction information indicates rule #2, then in S1020, the first device also sends a first identifier to the second device, and in S1030, the first device also sends a second identifier to the second device. The first identifier is used to identify the first value corresponding to the first feature of the first model, and the second identifier is used to identify the second value corresponding to the first feature of the second model. The first feature belongs to the first feature set.

[0273] Example 2-1, Rule #2 is: if the first indication information includes at least one first feature identifier, and the first first feature identifier is used to identify at least one third value corresponding to the first feature, or if the first indication information includes one or more first feature identifiers, and each first feature identifier is used to identify a set of first values ​​corresponding to the first feature, then if the values ​​corresponding to the first features of at least two models are included in at least one third value or at least one set of first values, the model parameters corresponding to at least two models can be used for training the same model.

[0274] Based on rule #2 above, if at least one third value or at least one set of first values ​​includes the first value and the second value (equivalent to the first indication information including the first identifier and the second identifier), then the second device determines that the first model parameters and the second model parameters can be used for training the same model according to the first indication information; if at least one third value or at least one set of first values ​​does not include the first value and / or the second value (equivalent to the first indication information not including the first identifier and / or the second identifier), then the second device determines that the first model parameters and the second model parameters cannot be used for training the same model according to the first indication information.

[0275] Optionally, if the first value and the second value are the same, the second device can determine that the first model parameters and the second model parameters can be used for training the same model even if at least one third value or at least one set of first values ​​does not include the first value and / or the second value (equivalent to the first indication information not including the first identifier and / or the second identifier).

[0276] It should be noted that if the first feature set includes multiple features, then the first feature can correspond to one feature in the first feature set, or it can correspond to at least two features in the first feature set. If the first feature corresponds to at least two features in the first feature set (denoted as at least two features #a), then the value corresponding to the first feature includes the value corresponding to each of the at least two features #a. For example, if the first feature corresponds to the aforementioned features #1 and #2, then the value corresponding to the first feature includes the value corresponding to feature #1 and the value corresponding to feature #2.

[0277] It should also be noted that if the first feature corresponds to at least two features #a in the first feature set, then the aforementioned first value includes the first value #a corresponding to each of the at least two features #a, the second value includes the second value #a corresponding to each of the at least two features #a, the third value includes the third value #a corresponding to each of the at least two features #a, and the first value set includes the first value set #a corresponding to each of the at least two features #a. At least one third value or at least one first value set includes the first value, meaning that at least one third value or at least one first value set includes the first value #a corresponding to each of the at least two features #a; at least one third value or at least one first value set does not include the first value #a corresponding to at least one of the at least two features #a. Similarly, at least one third value or at least one set of first values ​​includes a second value, meaning that at least one third value includes a second value #a corresponding to each of at least two features #a; at least one third value or at least one set of first values ​​does not include a second value #a, meaning that at least one third value or at least one set of first values ​​does not include a second value #a corresponding to at least one of at least two features #a.

[0278] It should also be noted that if the first feature corresponds to at least two features #a in the first feature set, then the identifier used to identify the value corresponding to the first feature (e.g., the first identifier, the second identifier, or the first feature identifier) ​​can be a single identifier, or the identifier used to identify the value corresponding to the first feature can include at least two identifiers #a, each of which is used to identify the value corresponding to each of the at least two features #a. For example, if the first feature corresponds to the aforementioned features #1 and #2, then the first identifier is identifier #x in Table 2 below, or it can include identifier #x1 (an example of identifier #a) and identifier #x2 (an example of identifier #a) shown in Table 2 below. Identifier #x is used to identify the first value #a corresponding to feature #1 as 0.01, and is used to identify the first value #a corresponding to feature #2 as 0.1. Identifier #x1 is used to identify the first value #a corresponding to feature #1 as 0.05, and identifier #x2 is used to identify the first value #a corresponding to feature #2 as 0.2.

[0279] Table 2

[0280] It should also be noted that if the first feature corresponds to at least two features #a in the first feature set, and the identifier used to identify the value corresponding to the first feature is an identifier, then the correspondence between the first feature and the at least two features #a can be predefined or preconfigured, or indicated by the first device to the second device. In other words, the correspondence between an identifier used to identify the value corresponding to the first feature and the at least two features #a can be predefined or preconfigured, or indicated by the first device to the second device.

[0281] It should also be noted that if the first feature corresponds to at least two features #a in the first feature set, and the first identifier includes at least two first identifiers #a, and the at least two first identifiers #a are respectively used to identify the first value #a corresponding to each feature #a in the at least two features #a, then the first indication information including the first identifier means that the first indication information includes at least two first identifiers #a. Similarly, if the second identifier includes at least two second identifiers #a, and the at least two second identifiers #a are respectively used to identify the second value #a corresponding to each feature #a in the at least two features #a, then the first indication information including the second identifier means that the first indication information includes at least two second identifiers #a.

[0282] Example 2-2, Rule #2 is: If the first indication information includes at least one feature identifier #a, and the at least one feature identifier #a is used to identify at least one value #a or at least one set of values ​​#a corresponding to the first feature, then if the values ​​corresponding to the first features of at least two models are included in at least one value #a or at least one set of values ​​#a, the model parameters corresponding to at least two models cannot be used for training the same model.

[0283] Based on rule #2 above, if at least one value #a or at least one set of values ​​#a includes a first value and a second value (equivalent to the first indication information including a first identifier and a second identifier), then the second device determines, according to the first indication information, that the first model parameters and the second model parameters cannot be used for training the same model; if at least one value #a or at least one set of values ​​#a does not include the first value and / or the second value (equivalent to the first indication information not including the first identifier and / or the second identifier), then the second device determines, according to the first indication information, that the first model parameters and the second model parameters can be used for training the same model.

[0284] Example 2-3, Rule #2 states: Model parameters corresponding to models with different values ​​for the first feature can be used to train the same model. In other words, the first indication information is used to indicate that model parameters corresponding to models with different values ​​for the first feature can be used to train the same model.

[0285] Based on rule #2 above, the second device determines, according to the first instruction information, that the first dataset and the second dataset can be used for training the same model.

[0286] It should be noted that if the first feature corresponds to at least two features #a in the first feature set, then the different values ​​of the first feature of the two models mean that at least one of the at least two features #a of the two models has a different value.

[0287] Example 2-4, Rule #2 states: Model parameters corresponding to models with different values ​​for the first feature cannot be used to train the same model. In other words, the first indication information is used to indicate that model parameters corresponding to models with different values ​​for the first feature cannot be used to train the same model.

[0288] Based on rule #2 above, if the first value and the second value are different, the second device determines that the first model parameters and the second model parameters cannot be used for training the same model; if the first value and the second value are the same, the second device determines that the first model parameters and the second model parameters can be used for training the same model.

[0289] It should be noted that if the first feature corresponds to at least two features #a in the first feature set, then the first feature of the two models has the same value, which means that each feature #a in at least two features #a of the two models has the same value.

[0290] Example 2-5, Rule #2 states: The values ​​of the first features of at least two models that can be used to train the same model must satisfy condition #2. In other words, the first indication information indicates that the condition #2 must be satisfied for the values ​​of the first features of at least two models that can be used to train the same model.

[0291] This application does not limit condition #2. For example, condition #2 is that the values ​​corresponding to the first features of two different models are the same. Another example is that condition #2 is that the difference between the values ​​corresponding to the first features of two different models does not exceed threshold #2.

[0292] Based on rule #2 above, if the first value and the second value satisfy condition #2, then the second device determines that the first model parameter and the second model parameter can be used for training the same model; if the first value and the second value do not satisfy condition #2, then the second device determines that the first model parameter and the second model parameter cannot be used for training the same model.

[0293] The following example, which includes the first feature set including features #1 to #5, illustrates how rule #2 is determined.

[0294] For example, rule #2 is related to Table 3 below.

[0295] Table 3

[0296] It should be understood that the different values ​​corresponding to each feature shown in Table 3 are merely examples, and this application does not limit them.

[0297] For example, if the first feature is feature #1 and the first indication information indicates rule #2 in Example 2-2 above, then the first indication information includes at least one feature identifier #a, which is identifier #1 to identifier #3. Identifier #1 is used to identify the value corresponding to feature #1 as 0.01, identifier #2 is used to identify the value corresponding to feature #1 as 0.05, and identifier #3 is used to identify the value corresponding to feature #1 as 0.1.

[0298] For another example, if the first feature corresponds to one or more of features #1 to features 5, then the first indication information may include rule #2 in Examples 2-4 above.

[0299] Optionally, if the first feature set also includes features other than the first feature, for example, if the first feature set also includes feature #p, and feature #p is different from the first feature, then rule #2 also includes rules related to the value corresponding to feature #p. Correspondingly, in S1020, the first device also sends an identifier #p1 to the second device, which identifies the value #p1 corresponding to feature #p of the first model; and in S1030, the first device also sends an identifier #p2 to the second device, which identifies the value #p2 corresponding to feature #p of the second model.

[0300] The rules related to the value of feature #p can be found in the description of the rules related to the value of the first feature above. For example, rule #2 includes the following rules related to the value of feature #p: the first indication information includes at least one feature identifier #p, and the at least one feature identifier #p is used to identify at least one value #p3 corresponding to feature #p. Then, if the value of feature #p of at least two models is included in at least one value #p3, the model parameters of at least two models can be used for training the same model.

[0301] It should be noted that if rule #2 includes a rule related to the value corresponding to the first feature (denoted as rule #2-1) and a rule related to the value corresponding to feature #p (denoted as rule #2-2), then the second device determines that the first model parameter and the second model parameter can be used for training the same model, meaning that the second device determines that the first model parameter and the second model parameter can be used for training the same model according to rule #2-1 and rule #2-2; the second device determines that the first model parameter and the second model parameter cannot be used for training the same model, meaning that the second device determines that the first model parameter and the second model parameter cannot be used for training the same model according to rule #2-1 and / or rule #2-2.

[0302] Implementation method 4: The first indication information indicates rule #3. Rule #3 is related to the values ​​of the features included in the second feature set of the model. Rule #3 is used to determine whether the model parameters corresponding to at least two models can be used or cannot be used for training the same model.

[0303] For example, the second feature set of the model is related to one or more of the following: the temporal features of the data in the training dataset corresponding to the model, the AI ​​function corresponding to the model, and the functions supported by the model.

[0304] For example, the second feature set includes one or more of the following: the configuration of the historical time window length corresponding to the data in the training dataset corresponding to the model (hereinafter referred to as feature #6), the configuration of the prediction time window length corresponding to the data in the training dataset corresponding to the model (hereinafter referred to as feature #7), and the function for compressing CSI and / or predicting CSI (hereinafter referred to as feature #8).

[0305] In this context, the historical time window corresponding to the data in the dataset refers to the time window used to acquire the model's input. For example, if the model corresponding to the data in the dataset is used for CSI compression, then the historical time window could be the number of historical CSIs used by the model in performing one inference task. The prediction time window corresponding to the data in the dataset refers to the prediction time window corresponding to the prediction task performed by the model. If the model corresponding to the data in the dataset is used for CSI prediction and / or compression, then the prediction time window could be the number of CSIs obtained by the model in performing one prediction task.

[0306] It should be understood that if the first instruction information indicates rule #3, then in S1020, the first device also sends a third identifier to the second device, and in S1030, the first device also sends a fourth identifier to the second device. The third identifier is used to identify the fourth value corresponding to the second feature of the first model, and the fourth identifier is used to identify the fifth value corresponding to the second feature of the second model. The second feature belongs to the second feature set.

[0307] Optionally, the third identifier can be the identifier of the first dataset, that is, the identifier of the first dataset can be used to identify the values ​​corresponding to the features included in the second feature set of the first model, or in other words, the identifier of the first dataset can be used to identify the functions supported by the first model. The fourth identifier can be the identifier of the second dataset, that is, the identifier of the second dataset can be used to identify the values ​​corresponding to the features included in the second feature set of the second model, or in other words, the identifier of the second dataset can be used to identify the functions supported by the second model.

[0308] Example 3-1, Rule #3 is: if the first indication information includes at least one second feature identifier, and the at least one second feature identifier is used to identify at least one sixth value corresponding to the second feature, or if the first indication information includes one or more second feature identifiers, and each second feature identifier is used to identify a set of second values ​​corresponding to the first feature, then if the values ​​corresponding to the second features of at least two models are included in at least one sixth value or at least one set of second values, the model parameters corresponding to at least two models can be used for training the same model.

[0309] Based on rule #3 above, if at least one sixth value or at least one set of second values ​​includes the fourth and fifth values ​​(equivalent to the first indication information including the third and fourth identifiers), then the second device determines that the first model parameters and the second model parameters can be used for training the same model according to the first indication information; if at least one sixth value or at least one set of second values ​​does not include the fourth and / or fifth values ​​(equivalent to the first indication information not including the third and / or fourth identifiers), then the second device determines that the first model parameters and the second model parameters cannot be used for training the same model according to the first indication information.

[0310] Optionally, if the fourth value is the same as the fifth value, then even if at least one sixth value or at least one set of second values ​​does not include the fourth value and / or the fifth value (equivalent to the first indication information not including the third identifier and / or the fourth identifier), the second device can still determine that the first model parameters and the second model parameters can be used for training the same model.

[0311] It should be noted that if the second feature set includes multiple features, the second feature can correspond to one feature in the second feature set, or it can correspond to at least two features in the second feature set. If the second feature corresponds to at least two features in the second feature set (denoted as at least two features #b), then the value corresponding to the second feature includes the value corresponding to each of the at least two features #b. For example, if the second feature corresponds to features #6 and #7 mentioned above, then the value corresponding to the second feature includes the value corresponding to feature #6 and the value corresponding to feature #7.

[0312] It should also be noted that if the second feature corresponds to at least two features #b in the second feature set, then the aforementioned fourth value includes the fourth value #b corresponding to each of the at least two features #b, the fifth value includes the fifth value #b corresponding to each of the at least two features #b, the sixth value includes the sixth value #b corresponding to each of the at least two features #b, and the second value set includes the second value set #b corresponding to each of the at least two features #b. At least one sixth value or at least one second value set includes the fourth value, meaning that at least one sixth value or at least one second value set includes the fourth value #b corresponding to each of the at least two features #b; at least one sixth value or at least one second value set does not include the fourth value #b corresponding to at least one of the at least two features #b. Similarly, at least one sixth value or at least one set of second values ​​including a fifth value means that at least one sixth value or at least one set of second values ​​includes the fifth value #b corresponding to each of the at least two features #b; at least one sixth value or at least one set of second values ​​not including a fifth value means that at least one sixth value or at least one set of second values ​​does not include the fifth value #b corresponding to at least one of the at least two features #b.

[0313] It should also be noted that if the second feature corresponds to at least two features #b in the second feature set, then the identifier used to identify the value corresponding to the second feature (e.g., a third identifier, a fourth identifier, or a second feature identifier) ​​can be a single identifier. Alternatively, the identifier used to identify the value corresponding to the second feature can include at least two identifiers #b, each of which is used to identify the value corresponding to each of the at least two features #b. For a more detailed description of the identifier used to identify the value corresponding to the second feature, please refer to the description of the identifier used to identify the value corresponding to the first feature in Implementation Method 3 above.

[0314] It should also be noted that if the second feature corresponds to at least two features #b in the second feature set, and the third identifier includes at least two third identifiers #b, and the at least two third identifiers #b are respectively used to identify the fourth value #b corresponding to each of the at least two features #b, then the first indication information including the third identifier means that the first indication information includes at least two third identifiers #b. Similarly, if the fourth identifier includes at least two fourth identifiers #b, and the at least two fourth identifiers #b are respectively used to identify the fifth value #b corresponding to each of the at least two features #b, then the first indication information including the fourth identifier means that the first indication information includes at least two fourth identifiers #b.

[0315] Example 3-2, Rule #3 is: If the first indication information includes at least one feature identifier #b, and the at least one feature identifier #b is used to identify at least one value #b or at least one set of values ​​#b corresponding to the second feature, then if the values ​​corresponding to the second features of at least two models are included in at least one value #b or at least one set of values ​​#b, the model parameters corresponding to at least two models cannot be used for training the same model.

[0316] Based on rule #3 above, if at least one value #b or at least one set of values ​​#b includes a fourth value and a fifth value (equivalent to the first indication information including a third identifier and a fourth identifier), then the second device determines, according to the first indication information, that the first model parameters and the second model parameters cannot be used for training the same model; if at least one value #b or at least one set of values ​​#b does not include a fourth value and / or a fifth value (equivalent to the first indication information not including a third identifier and / or a fourth identifier), then the second device determines, according to the first indication information, that the first model parameters and the second model parameters can be used for training the same model.

[0317] Example 3-3, Rule #3 states: Model parameters corresponding to models with different values ​​for the second feature can be used to train the same model. In other words, the first indication information is used to indicate that model parameters corresponding to models with different values ​​for the second feature can be used to train the same model.

[0318] Based on rule #3 above, the second device determines, according to the first instruction information, that the first model parameters and the second model parameters can be used for training the same model.

[0319] It should be noted that if the second feature corresponds to at least two features #b in the second feature set, then the different values ​​of the second feature in the two models mean that at least one of the at least two features #b in the two models has a different value.

[0320] Example 3-4, Rule #3 states: Model parameters corresponding to models with different values ​​for the second feature cannot be used to train the same model. In other words, the first indication is used to indicate that model parameters corresponding to models with different values ​​for the second feature cannot be used to train the same model.

[0321] Based on rule #3 above, if the fourth value is different from the fifth value, the second device determines that the first model parameters and the second model parameters cannot be used for training the same model; if the fourth value is the same as the fifth value, the second device determines that the first model parameters and the second model parameters can be used for training the same model.

[0322] It should be noted that if the second feature corresponds to at least two features #b in the second feature set, then the second feature of the two models has the same value, which means that each feature #b in at least two features #b of the two models has the same value.

[0323] Example 3-5, Rule #3 states: The values ​​of the second features of at least two models that can be used to train the same model must satisfy condition #3. In other words, the first indication information indicates that the condition that the values ​​of the second features of at least two models that can be used to train the same model must satisfy is condition #3.

[0324] This application does not limit condition #3. For example, condition #3 is that the values ​​corresponding to the second features of two different models are the same. Another example is that condition #3 is that the difference between the values ​​corresponding to the second features of two different models does not exceed threshold #3.

[0325] Based on rule #3 above, if the fourth and fifth values ​​satisfy condition #3, the second device determines that the first model parameters and the second model parameters can be used for training the same model; if the fourth and fifth values ​​do not satisfy condition #3, the second device determines that the first model parameters and the second model parameters cannot be used for training the same model.

[0326] The following example, which includes the second feature set including features #6 to #8, illustrates how rule #3 is determined.

[0327] For example, rule #3 is related to Table 4 below.

[0328] Table 4

[0329] It should be understood that the different values ​​corresponding to each feature shown in Table 4 are merely examples, and this application does not limit them.

[0330] For example, if the second feature corresponds to one or more features from feature #6 to feature 8, then the first indication information may include rule #4 in Example 3-3 above.

[0331] Optionally, if the second feature set also includes features other than the second feature, for example, if the second feature set also includes feature #q, and feature #q is different from the second feature, then rule #3 also includes rules related to the value corresponding to feature #q. Correspondingly, in S1020, the first device also sends identifier #q1 to the second device, identifier #q1 being used to identify the value #q1 corresponding to feature #q of the first model; and in S1030, the first device also sends identifier #q2 to the second device, identifier #q2 being used to identify the value #q2 corresponding to feature #q of the second model.

[0332] The rules related to the value of feature #q can be found in the description of the rules related to the value of the second feature above. For example, rule #3 includes the following rules related to the value of feature #q: the first indication information includes at least one feature identifier #q, and the at least one feature identifier #q is used to identify at least one value #q3 corresponding to feature #q. Then, if the value of feature #q of at least two models is included in at least one value #q3, the model parameters of at least two models can be used for training the same model.

[0333] It should be noted that if rule #3 includes a rule related to the value of the second feature (denoted as rule #3-1) and a rule related to the value of feature #q (denoted as rule #3-2), then the second device determines that the first model parameter and the second model parameter can be used for training the same model, meaning that the second device determines that the first model parameter and the second model parameter can be used for training the same model according to rule #3-1 and rule #3-2; the second device determines that the first model parameter and the second model parameter cannot be used for training the same model, meaning that the second device determines that the first model parameter and the second model parameter cannot be used for training the same model according to rule #3-1 and / or rule #3-2.

[0334] Implementation method 5: The first instruction information indicates at least two of the above rules #1 to #3.

[0335] It should be understood that if the first instruction information indicates at least two of the rules #1 to #3, then in S1020, the first device may also send an eleventh identifier associated with the first model to the second device, the eleventh identifier including an identifier related to the rule indicated by the first instruction information, and in S1020, the first device may also send a twelfth identifier associated with the second model to the second device, the twelfth identifier including an identifier related to the rule indicated by the first instruction information.

[0336] For example, if the first instruction information includes the above-mentioned rule #1 and rule #2, then the eleventh identifier may include the identifier of the first model and the above-mentioned first identifier, and the twelfth identifier may include the identifier of the second model and the above-mentioned second identifier.

[0337] For example, if the first instruction information indicates the above rules #1 to #3, then the eleventh identifier may include the identifier of the first model, the above first identifier and the above third identifier, and the twelfth identifier may include the identifier of the second model, the above second identifier and the above fourth identifier.

[0338] It should be understood that if the first instruction information indicates at least two of the rules #1 to #3, the second device determines whether the first model parameters and the second model parameters can be used for training the same model based on at least one of the rules indicated by the first instruction information.

[0339] For example, the second device determines that the first model parameters and the second model parameters can be used for training the same model based on the first instruction information, including: the second device determines that the first model parameters and the second model parameters can be used for training the same model based on each rule indicated by the first instruction information.

[0340] For example, the second device determines that the first model parameters and the second model parameters cannot be used for training the same model based on the first instruction information, including: the second device determining that the first model parameters and the second model parameters cannot be used for training the same model based on at least one rule indicated by the first instruction information.

[0341] Optionally, the multiple rules indicated by the first indication information may have different priorities. In this case, the first device will prioritize determining whether the first model parameters and the second model parameters can be used for training the same model based on the rule with higher priority.

[0342] For example, if the first instruction indicates a first rule and a second rule, and the first and second rules belong to rules #1 to #3 above, with the first rule having a higher priority than the second rule, then the second device prioritizes determining whether the first model parameters and the second model parameters can be used for training the same model based on the first rule. If the second device determines that the first model parameters and the second model parameters cannot be used for training the same model based on the first rule, then the second device does not need to determine whether the first model parameters and the second model parameters can be used for training the same model based on the second rule. If the second device determines that the first model parameters and the second model parameters can be used for training the same model based on the second rule, then the second device continues to determine whether the first model parameters and the second model parameters can be used for training the same model based on the second rule. If the second rule is the rule with the lowest priority, and the second device determines that the first model parameters and the second model parameters can be used for training the same model based on the second rule, then the second device ultimately determines that the first model parameters and the second model parameters can be used for training the same model.

[0343] This application does not limit the priority order of rules #1 to #3. For example, the priority order of rules #1 to #3 from high to low is: rule 1, rule #2, and rule #3.

[0344] For example, if the first instruction information indicates at least two of the rules #1 to #3, the second device can determine the rules indicated by the first instruction information based on the received eleventh identifier and / or twelfth identifier. For instance, if the eleventh identifier includes the identifier of the first model and the aforementioned first identifier, then the second device can determine that the first instruction information indicates the aforementioned rules #1 and #2.

[0345] For example, if the first instruction information indicates at least two rules from rule #1 to rule #3, the rules indicated by the first instruction information are sorted according to their priority. For instance, if the first instruction information indicates rule #1 and rule #2, and rule #1 has a higher priority than rule #2, then rule #1 indicated by the first instruction information is placed before rule #2. Alternatively, if the first instruction information indicates at least two rules from rule #1 to rule #3, the rules indicated by the first instruction information are sorted according to a predefined or preconfigured order. Alternatively, if the first instruction information indicates at least two rules from rule #1 to rule #3, the first instruction information also indicates an identifier corresponding to each rule, thereby enabling the second device to determine which rules the first instruction information specifically indicates. The identifier corresponding to each rule from rule #1 to rule #3 may be predefined or preconfigured, or may be indicated by the first device to the second device; this application does not limit this.

[0346] For example, if the first indication information indicates at least two of rules #1 to #3, the different rules indicated by the first indication information can be carried in the same message, field, or information element, or the different rules indicated by the first indication information can be carried in different messages, fields, or information elements. For instance, if rule #1 and rule #2 indicated by the first indication information are carried in the same message, then in step 1040, the first device can send message #1, which indicates rule #1 and rule #2. As another example, if rule #1 and rule #2 indicated by the first indication information are carried in message #1, rule #1 and rule #2 can also be carried in the same field (or information element) of message #1, or in different fields (or information elements) of message #1. For example, message #1 may include field #1, which indicates rule #1 and rule #2. Alternatively, message #1 may include field #2 and field #3, where field #2 indicates rule #1 and field #3 indicates rule #2. For another example, if the first instruction information indicates rule #1 and rule #2 in different messages, then in S1040, the first device can send message #2 and message #3, where message #2 is used to indicate rule #1 and message #3 is used to indicate rule #2.

[0347] In this embodiment, for models with different features, the first device can send different forms of indication information, enabling the second device to determine whether model parameters corresponding to at least two models can be used for training the same model. This facilitates different training tasks based on the received model parameters of at least two models while saving model parameter transmission overhead, ensuring the performance of the trained or updated model. For example, if the value corresponding to the second feature of the first model is the same as the value corresponding to the second feature of the second model, the first device can send first indication information indicating rule #2, so that the second device can determine whether the first model parameters and the second model parameters can be used for training the same model based on the values ​​corresponding to the first and second features of the first and second models.

[0348] Implementation method 6: The first indication information indicates rule #4. Rule #4 is related to the identifier of the dataset. Rule #4 is used to determine whether the datasets corresponding to at least two models can be used or cannot be used for training the same model.

[0349] It should be understood that if the first instruction information indicates rule #4, then in S1020, the first device also sends the identifier of the first dataset to the second device, and in S1030, the first device also sends the identifier of the second dataset to the second device.

[0350] For example, rule #4 states: if the first indication information includes the identifiers of multiple datasets, then multiple datasets can be used to train the same model.

[0351] Based on rule #4 above, if the first indication information includes the identifier of the first dataset and the identifier of the second dataset, the second device determines that the first dataset and the second dataset can be used for training the same model according to the first indication information; if the first indication information does not include the identifier of the first dataset and / or the identifier of the second dataset, the second device determines that the first dataset and the second dataset cannot be used for training the same model according to the first indication information.

[0352] For example, rule #4 states: if the first indication information includes the identifiers of multiple datasets, then the multiple datasets cannot be used to train the same model.

[0353] Based on rule #4 above, if the first indication information includes the identifier of the first dataset and the identifier of the second dataset, the second device determines that the first dataset and the second dataset cannot be used for training the same model based on the first indication information; if the first indication information does not include the identifier of the first dataset and / or the identifier of the second dataset, the second device determines that the first dataset and the second dataset can be used for training the same model based on the first indication information.

[0354] For example, rule #4 states: Datasets with different labels can be used to train the same model. In other words, the first indication is used to indicate that datasets with different labels can be used to train the same model.

[0355] Based on rule #4 above, the second device determines, according to the first instruction information, that the first dataset and the second dataset can be used for training the same model.

[0356] For example, rule #4 states: Datasets with different labels cannot be used to train the same model. In other words, the first indication is used to indicate that datasets with different labels cannot be used to train the same model.

[0357] Based on rule #4 above, the second device determines, according to the first instruction information, that the first dataset and the second dataset cannot be used for training the same model.

[0358] For example, rule #4 states that the identifiers of at least two datasets that can be used to train the same model must satisfy condition #4. In other words, the first indication information is used to indicate that the identifiers of at least two datasets that can be used to train the same model must satisfy condition #4.

[0359] This application does not limit condition #4. For example, condition #4 is that the difference between the identifiers of two different datasets does not exceed threshold #4. Another example is that condition #4 is that the sum of the identifiers of two different datasets is a predefined value.

[0360] Based on rule #4 above, if the identifiers of the first dataset and the second dataset satisfy condition #4, then the second device determines that the first dataset and the second dataset can be used for training the same model; if the identifiers of the first dataset and the second dataset do not satisfy condition #4, then the second device determines that the first dataset and the second dataset cannot be used for training the same model.

[0361] Implementation method 7, the first indication information indicates rule #5, rule #5 is related to the values ​​of the features included in the third feature set of the dataset, rule #5 is used to determine whether the datasets corresponding to at least two models can be used or cannot be used for training the same model.

[0362] For example, the third feature set of the dataset includes features related to one or more of the following: different preprocessing methods of the data in the dataset, different quantization methods of the data in the dataset, and different channel characteristics of the data in the dataset.

[0363] For example, the third feature set includes one or more of the following: whether the data was acquired during training with a quantizer (hereinafter referred to as feature #9), the range of transmit power distribution corresponding to the data (hereinafter referred to as feature #10), the range of receive power distribution corresponding to the data, the range of receive noise power distribution corresponding to the data, the range of delay spread distribution corresponding to the data (hereinafter referred to as feature #11), the range of angle spread distribution corresponding to the data, the range of interference level distribution corresponding to the data, and the range of signal to interference plus noise ratio (SINR) distribution corresponding to the data (hereinafter referred to as feature #12).

[0364] In this context, the data obtained by training the encoder with a quantizer and the decoder with a dequantizer refers to the data acquired during the training process with the quantizer. For example, if the data in the dataset is obtained through channel measurements, then the corresponding transmit power distribution range is the transmit power distribution range corresponding to the transmit power used to transmit the reference signal. The corresponding receive power distribution range, receive noise power distribution range, delay spread distribution range, angle spread distribution range, interference level distribution range, and SINR distribution range are obtained by the receiving reference signal device through channel measurements. As another example, if the data in the dataset is generated through channel modeling, then the corresponding transmit power distribution range is the transmit power distribution range corresponding to the transmit power configured for the channel model. The corresponding receive power distribution range, receive noise power distribution range, delay spread distribution range, angle spread distribution range, interference level distribution range, and SINR distribution range are obtained through channel modeling.

[0365] It should be understood that if the first instruction information indicates rule #5, then in S1020, the first device also sends a fifth identifier to the second device, and in S1030, the first device also sends a sixth identifier to the second device. The fifth identifier is used to identify the seventh value corresponding to the third feature of the first dataset, and the sixth identifier is used to identify the eighth value corresponding to the third feature of the second dataset. The third feature belongs to the third feature set.

[0366] Example 5-1, Rule #5 is: if the first indication information includes at least one third feature identifier, and the at least one third feature identifier is used to identify at least one ninth value corresponding to the third feature, or if the first indication information includes one or more third feature identifiers, and each third feature identifier is used to identify a set of third values ​​corresponding to the first feature, then if the values ​​corresponding to the third features of at least two datasets are included in at least one ninth value or at least one set of third values, then at least two datasets can be used for training the same model.

[0367] Based on rule #5 above, if at least one ninth value or at least one set of third values ​​includes the seventh and eighth values ​​(equivalent to the first indication information including the fifth and sixth identifiers), the second device determines that the first dataset and the second dataset can be used for training the same model according to the first indication information; if at least one ninth value or at least one set of third values ​​does not include the seventh and / or the eighth value (equivalent to the first indication information not including the fifth and / or the sixth identifier), the second device determines that the first dataset and the second dataset cannot be used for training the same model according to the first indication information.

[0368] Optionally, if the seventh value is the same as the eighth value, then even if at least one ninth value or at least one third value set does not include the seventh value and / or the eighth value (equivalent to the first indication information not including the fifth identifier and / or the sixth identifier), the second device can still determine that the first dataset and the second dataset can be used for training the same model.

[0369] It should be noted that if the third feature set includes multiple features, the third feature can correspond to one feature in the third feature set, or it can correspond to at least two features in the third feature set. If the third feature corresponds to at least two features in the third feature set (denoted as at least two features #c), then the value corresponding to the third feature includes the value corresponding to each of the at least two features #c. For example, if the third feature corresponds to features #9 and #10 mentioned above, then the value corresponding to the first feature includes the value corresponding to feature #9 and the value corresponding to feature #10.

[0370] It should also be noted that if the third feature corresponds to at least two features #c in the third feature set, then the aforementioned seventh value includes the seventh value #c corresponding to each of the at least two features #c, the eighth value includes the eighth value #c corresponding to each of the at least two features #c, the ninth value includes the ninth value #c corresponding to each of the at least two features #c, and the third value set includes the third value set #c corresponding to each of the at least two features #c. At least one ninth value or at least one third value set includes the seventh value, meaning that at least one ninth value or at least one third value set includes the seventh value #c corresponding to each of the at least two features #c; at least one ninth value or at least one third value set does not include the seventh value #c corresponding to at least one of the at least two features #c. Similarly, at least one ninth value or at least one set of third values ​​includes an eighth value, meaning that at least one ninth value or at least one set of third values ​​includes the eighth value #c corresponding to each of at least two features #c; at least one ninth value or at least one set of third values ​​does not include an eighth value #c, meaning that at least one ninth value or at least one set of third values ​​does not include the eighth value #c corresponding to at least one of at least two features #c.

[0371] It should also be noted that if the third feature corresponds to at least two features #c in the third feature set, then the identifier used to identify the value corresponding to the third feature (e.g., the fifth identifier, the sixth identifier, or the third feature identifier) ​​can be a single identifier. Alternatively, the identifier used to identify the value corresponding to the third feature can include at least two identifiers #c, each of which is used to identify the value corresponding to each of the at least two features #a. For a more detailed description of the identifier used to identify the value corresponding to the third feature, please refer to the description of the identifier used to identify the value corresponding to the first feature in implementation method 3 above.

[0372] It should also be noted that if the third feature corresponds to at least two features #c in the third feature set, and the identifier used to identify the value corresponding to the third feature is an identifier, then the correspondence between the third feature and the at least two features #c is predefined or preconfigured, or indicated by the first device to the second device. In other words, the correspondence between an identifier used to identify the value corresponding to the third feature and the at least two features #c is predefined or preconfigured, or indicated by the first device to the second device.

[0373] It should also be noted that if the third feature corresponds to at least two features #c in the third feature set, and the fifth identifier includes at least two fifth identifiers #c, and the at least two fifth identifiers #c are used to identify the seventh value #c corresponding to each of the at least two features #c, then the first indication information including the fifth identifier means that the first indication information includes at least two fifth identifiers #c. Similarly, if the sixth identifier includes at least two sixth identifiers #c, and the at least two sixth identifiers #c are used to identify the eighth value #c corresponding to each of the at least two features #c, then the first indication information including the sixth identifier means that the first indication information includes at least two sixth identifiers #c.

[0374] Example 5-2, Rule #5 states: If the first indication information includes at least one feature identifier #c, and the at least one feature identifier #c is used to identify at least one value #c or at least one set of values ​​#c corresponding to the third feature, then if the values ​​corresponding to the third feature of at least two datasets are included in at least one value #c or at least one set of values ​​#cc, then at least two datasets cannot be used for training the same model.

[0375] Based on rule #5 above, if at least one value #c or at least one set of values ​​#c includes the seventh value and the eighth value (equivalent to the first indication information including the fifth identifier and the sixth identifier), then the second device determines that the first dataset and the second dataset cannot be used for training the same model according to the first indication information; if at least one value #c or at least one set of values ​​#c does not include the seventh value and / or the eighth value (equivalent to the first indication information not including the fifth identifier and / or the sixth identifier), then the second device determines that the first dataset and the second dataset can be used for training the same model according to the first indication information.

[0376] Example 5-3, Rule #5 states: Datasets with different values ​​corresponding to the third feature can be used to train the same model. In other words, the first indication is used to indicate that datasets with different values ​​corresponding to the third feature can be used to train the same model.

[0377] Based on rule #2 above, the second device determines, according to the first instruction information, that the first dataset and the second dataset can be used for training the same model.

[0378] It should be noted that if the third feature corresponds to at least two features #c in the third feature set, then the different values ​​of the third feature in the two datasets mean that at least one of the at least two features #c in the two datasets has a different value.

[0379] Example 5-4, Rule #5 states: Datasets with different values ​​corresponding to the third feature cannot be used to train the same model. In other words, the first indication is used to indicate that datasets with different values ​​corresponding to the third feature cannot be used to train the same model.

[0380] Based on rule #5 above, if the seventh value is different from the eighth value, the second device determines that the first dataset and the second dataset cannot be used for training the same model; if the seventh value is the same as the eighth value, the second device determines that the first dataset and the second dataset can be used for training the same model.

[0381] It should be noted that if the third feature corresponds to at least two features #c in the third feature set, then the values ​​corresponding to the third feature in the two datasets are the same, which means that the values ​​corresponding to each feature #c in at least two features #c in the two datasets are the same.

[0382] Example 5-5, Rule #5 states: The values ​​corresponding to the third features of at least two datasets that can be used to train the same model must satisfy condition #5. In other words, the first indication information is used to indicate that the value corresponding to the third feature of at least two datasets that can be used to train the same model must satisfy condition #5.

[0383] This application does not limit condition #5. For example, condition #5 is that the values ​​corresponding to the third feature of two different datasets are the same. Another example is that condition #5 is that the difference between the values ​​corresponding to the third feature of two different datasets does not exceed threshold #5.

[0384] Based on rule #5 above, if the seventh and eighth values ​​satisfy condition #5, the second device determines that the first and second datasets can be used for training the same model; if the seventh and eighth values ​​do not satisfy condition #5, the second device determines that the first and second datasets cannot be used for training the same model.

[0385] The following example, using the third feature set including features #9 to #12, illustrates how rule #5 is determined.

[0386] For example, rule #5 is related to Table 5 below.

[0387] Table 5

[0388] It should be understood that the different values ​​corresponding to each feature shown in Table 5 are merely examples, and this application does not limit them.

[0389] For example, if the third feature is feature #11 and the first indication information indicates rule #5 in Example 5-1 above, then the first indication information includes at least one third feature identifier as identifier #4 to identifier #6. Identifier #4 is used to identify the value corresponding to feature #11 as [0,10]ns, identifier #5 is used to identify the value corresponding to feature #11 as [10,20]ns, and identifier #6 is used to identify the value corresponding to feature #11 as [20,30]ns.

[0390] For another example, if the third feature is feature #9, then the first indication information may include rule #5 from example 5-4 above.

[0391] For example, if the third feature is feature #11, then the first indication information may include rule #5 from example 5-3 above.

[0392] For another example, if the third feature corresponds to features #9 and #11, then the first indication information may include rule #5 from example 5-4 above.

[0393] For example, if the third feature is feature #11 and feature #12, then the first indication information may include rule #5 in example 5-3 above.

[0394] Optionally, if the third feature set includes features other than the third feature, for example, if the third feature set also includes feature #r, and feature #r is different from the third feature, then rule #5 also includes rules related to the value corresponding to feature #r. Correspondingly, in S1020, the first device also sends identifier #r1 to the second device, identifier #r1 being used to identify the value #r1 corresponding to feature #r in the first dataset; and in S1030, the first device also sends identifier #r2 to the second device, identifier #r2 being used to identify the value #r2 corresponding to feature #r in the second dataset.

[0395] The rules related to the value of feature #r can be found in the description of the rules related to the value of the third feature above. For example, rule #5 includes the following rules related to the value of feature #r: the first indication information includes at least one feature identifier #r, and the at least one feature identifier #r is used to identify at least one value #r3 corresponding to feature #r. If the value of feature #r in at least two datasets is included in at least one value #r3, then at least two datasets can be used to train the same model.

[0396] It should be noted that if rule #5 includes: a rule related to the value of the third feature (denoted as rule #5-1) and a rule related to the value of feature #r (denoted as rule #5-2), then the second device determines that the first dataset and the second dataset can be used for training the same model, meaning that the second device determines that the first dataset and the second dataset can be used for training the same model according to rule #5-1 and rule #5-2; the second device determines that the first dataset and the second dataset cannot be used for training the same model, meaning that the second device determines that the first dataset and the second dataset cannot be used for training the same model according to rule #5-1 and / or rule #5-2.

[0397] Implementation method 8, the first indication information indicates rule #6, rule #6 is related to the values ​​corresponding to the features included in the fourth feature set of the dataset, rule #6 is used to determine whether at least two datasets can be used or cannot be used for training the same model.

[0398] For example, the fourth feature set of the dataset includes features related to one or more of the following: the transmission configuration corresponding to the data in the dataset, and / or the reception configuration corresponding to the data in the dataset, and the format of the data in the dataset. For instance, if the data in the dataset is obtained through channel measurements, then the transmission configuration corresponding to the data in the dataset is a configuration for transmitting a reference signal, and the reception configuration corresponding to the data in the dataset is a configuration for receiving a reference signal. As another example, if the data in the dataset is generated through channel modeling, then the transmission or reception configuration corresponding to the data in the dataset is a configuration of the channel model.

[0399] For example, the fourth feature set includes one or more of the following: the number of transmission ports corresponding to the data in the dataset (hereinafter referred to as feature #13), the carrier frequency corresponding to the data in the dataset (hereinafter referred to as feature #14), the transmission bandwidth corresponding to the data in the dataset (hereinafter referred to as feature #15), the number of transmission subbands corresponding to the data in the dataset (hereinafter referred to as feature #16), the granularity of the transmission subbands corresponding to the data in the dataset (hereinafter referred to as feature #17), and the number of bits of CSI feedback overhead corresponding to the data in the dataset (hereinafter referred to as feature #18).

[0400] In this dataset, the number of transmitting ports, carrier frequency, transmission bandwidth, number of transmitting sub-bands, and granularity of the transmitting sub-bands correspond to the transmitting and / or receiving configurations of the data. For example, if the data is obtained through channel measurements, then the number of transmitting ports is the number of ports used to transmit reference signals, the carrier frequency is the carrier frequency used to transmit and / or receive reference signals, the transmission bandwidth is the transmission bandwidth used to transmit reference signals, the number of transmitting sub-bands is the number of sub-bands used to transmit reference signals, and the granularity of the transmitting sub-bands is the sub-band granularity used to transmit reference signals.

[0401] It should be understood that if the first instruction information indicates rule #6, then in S1020, the first device also sends a seventh identifier to the second device, and in S1030, the first device also sends an eighth identifier to the second device. The seventh identifier is used to identify the tenth value corresponding to the fourth feature of the first dataset, and the eighth identifier is used to identify the eleventh value corresponding to the fourth feature of the second dataset. The fourth feature belongs to the fourth feature set.

[0402] Example 6-1, Rule #6 states: If the first indication information includes at least one fourth feature identifier, and the at least one fourth feature identifier is used to identify at least one twelfth value corresponding to the fourth feature, or if the first indication information includes one or more fourth feature identifiers, and each fourth feature identifier is used to identify a set of fourth values ​​corresponding to the first feature, then if the values ​​corresponding to the fourth feature of at least two datasets are included in at least one twelfth value or at least one set of fourth values, then at least two datasets can be used for training the same model.

[0403] Based on rule #6 above, if at least one twelfth value or at least one fourth value set includes the tenth and eleventh values ​​(equivalent to the first indication information including the seventh and eighth identifiers), then the second device determines that the first dataset and the second dataset can be used for training the same model according to the first indication information; if at least one twelfth value or at least one fourth value set does not include the tenth and / or eleventh values ​​(equivalent to the first indication information not including the seventh and / or eighth identifiers), then the second device determines that the first dataset and the second dataset cannot be used for training the same model according to the first indication information.

[0404] Optionally, if the tenth value and the eleventh value are the same, then even if at least one twelfth value or at least one fourth value set does not include the tenth value and / or the eleventh value (equivalent to the first indication information not including the seventh identifier and / or the eighth identifier), the second device can still determine that the first dataset and the second dataset can be used for training the same model.

[0405] For a more detailed description of rule #6 in Example 6-1, please refer to the description of rule #5 in Example 5-1 above.

[0406] Example 6-2, Rule #6 is: If the first indication information includes at least one feature identifier #d, and the at least one feature identifier #d is used to identify at least one value #d or at least one set of values ​​#d corresponding to the fourth feature, then if the values ​​corresponding to the fourth feature of at least two datasets are included in at least one value #d or at least one set of values ​​#d, then at least two datasets cannot be used for training the same model.

[0407] Based on rule #3 above, if at least one value #d or at least one set of values ​​#d includes the tenth and eleventh values ​​(equivalent to the first indication information including the seventh and eighth identifiers), then the second device determines that the first dataset and the second dataset cannot be used for training the same model according to the first indication information; if at least one value #d or at least one set of values ​​#d does not include the tenth and / or eleventh values ​​(equivalent to the first indication information not including the seventh and / or eighth identifiers), then the second device determines that the first dataset and the second dataset can be used for training the same model according to the first indication information.

[0408] Example 6-3, Rule #6 states: Datasets with different values ​​corresponding to the fourth feature can be used to train the same model. In other words, the first indication is used to indicate that datasets with different values ​​corresponding to the fourth feature can be used to train the same model.

[0409] Based on rule #6 above, the second device determines, according to the first instruction information, that the first dataset and the second dataset can be used for training the same model.

[0410] It should be noted that if the fourth feature corresponds to at least two features #d in the fourth feature set, then the different values ​​corresponding to the fourth feature mean that at least one of the at least two features #d included in the fourth feature set has a different value.

[0411] Example 6-4, Rule #6 states: Datasets with different values ​​corresponding to the fourth feature cannot be used to train the same model. In other words, the first indication is used to indicate that datasets with different values ​​corresponding to the fourth feature cannot be used to train the same model.

[0412] Based on rule #6 above, if the tenth value and the eleventh value are different, the second device determines that the first dataset and the second dataset cannot be used for training the same model; if the tenth value and the eleventh value are the same, the second device determines that the first dataset and the second dataset can be used for training the same model.

[0413] It should be noted that if the fourth feature corresponds to at least two features #d in the fourth feature set, then the values ​​corresponding to the fourth feature in the two datasets are the same, which means that the values ​​corresponding to each feature #d in at least two features #d in the two datasets are the same.

[0414] Example 6-5, Rule #6 states: The values ​​corresponding to the fourth feature of at least two datasets that can be used to train the same model must satisfy condition #6. In other words, the first indication information is used to indicate that the value corresponding to the fourth feature of at least two datasets that can be used to train the same model must satisfy condition #6.

[0415] This application does not limit condition #6. For example, condition #6 is that the values ​​corresponding to the fourth feature in two different datasets are the same. Another example is that condition #6 is that the difference between the values ​​corresponding to the fourth feature in two different datasets does not exceed threshold #6.

[0416] Based on rule #6 above, if the tenth and eleventh values ​​satisfy condition #6, the second device determines that the first and second datasets can be used for training the same model; if the tenth and eleventh values ​​do not satisfy condition #6, the second device determines that the first and second datasets cannot be used for training the same model.

[0417] The following example, using the fourth feature set including features #13 to #18, illustrates how rule #6 is determined.

[0418] For example, rule #6 is related to Table 6 below.

[0419] Table 6

[0420] It should be understood that the different values ​​corresponding to each feature shown in Table 6 are merely examples, and this application does not limit them.

[0421] For example, if the fourth feature is feature #13 and the first indication information indicates rule #6 in Example 6-1 above, then the first indication information includes at least one fourth feature identifier, which is identifier #7 to identifier #10. Identifier #7 is used to identify the value corresponding to feature #13 as port 16, identifier #8 is used to identify the value corresponding to feature #13 as port 32, identifier #9 is used to identify the value corresponding to feature #13 as port 64, and identifier #10 is used to identify the value corresponding to feature #13 as port 256.

[0422] For another example, if the fourth feature corresponds to one or more features from feature #13 to feature #18, then the first indication information may include rule #6 in example 6-3 above.

[0423] Optionally, if the fourth feature set also includes features other than the fourth feature, for example, if the fourth feature set also includes feature #s, and feature #s is different from the fourth feature, then rule #6 also includes rules related to the value corresponding to feature #s. Correspondingly, in S1020, the first device also sends identifier #s1 to the second device, identifier #s1 being used to identify the value #s1 corresponding to feature #s in the first dataset; and in S1030, the first device also sends identifier #s2 to the second device, identifier #s2 being used to identify the value #s2 corresponding to feature #s in the second dataset.

[0424] The rules related to the value of feature #s can be found in the description of the rules related to the value of the fourth feature above. For example, rule #6 includes the following rules related to the value of feature #s: the first indication information includes at least one feature identifier #s, and the at least one feature identifier #s is used to identify at least one value #s3 corresponding to feature #s. If the value of feature #s in at least two datasets is included in at least one value #s3, then at least two datasets can be used to train the same model.

[0425] It should be noted that if rule #6 includes a rule related to the value of the fourth feature (denoted as rule #6-1) and a rule related to the value of feature #s (denoted as rule #6-2), then the second device determines that the first dataset and the second dataset can be used for training the same model, meaning that the second device determines that the first dataset and the second dataset can be used for training the same model according to rule #6-1 and rule #6-2; the second device determines that the first dataset and the second dataset cannot be used for training the same model, meaning that the second device determines that the first dataset and the second dataset cannot be used for training the same model according to rule #6-1 and / or rule #6-2.

[0426] Implementation method 9, the first indication information indicates rule #7, rule #7 is related to the values ​​corresponding to the features included in the fifth feature set of the dataset, rule #7 is used to determine whether at least two datasets can be used or cannot be used for training the same model.

[0427] For example, the fifth feature set of the dataset is related to one or more of the following: the temporal features of the data in the dataset, the AI ​​function corresponding to the data in the dataset, the function supported by the data in the dataset, the model corresponding to the data in the dataset, and the model supported by the data in the dataset. Here, the model corresponding to the data in the dataset refers to the model obtained by training the model based on the dataset.

[0428] For example, the fifth feature set includes one or more of the following: the length of the historical time window corresponding to the data in the dataset (hereinafter referred to as feature #19), the length of the prediction time window corresponding to the data in the dataset (hereinafter referred to as feature #20), and the moving speed of the device corresponding to the data in the dataset (hereinafter referred to as feature #21).

[0429] The historical time windows and prediction time windows corresponding to the data in the dataset can be referred to the description in Implementation Method 4 above. The devices corresponding to the dataset include the transmitting device and / or receiving device corresponding to the dataset. For example, if the dataset is obtained through channel measurement, then the transmitting device corresponding to the dataset is a device for transmitting reference signals, and the receiving device corresponding to the dataset is a device for receiving reference signals.

[0430] It should be understood that if the first instruction information indicates rule #7, then in S1020, the first device also sends a ninth identifier to the second device, and in S1030, the first device also sends a tenth identifier to the second device. The ninth identifier is used to identify the thirteenth value corresponding to the fifth feature of the first dataset, and the tenth identifier is used to identify the fourteenth value corresponding to the fifth feature of the second dataset. The fifth feature belongs to the fifth feature set.

[0431] Optionally, the ninth identifier can be the identifier of the first model, that is, the identifier of the first model can be used to identify the thirteenth value corresponding to the fifth feature of the first dataset. The tenth identifier can be the identifier of the second model, that is, the identifier of the second model can be used to identify the fourteenth value corresponding to the fifth feature of the second dataset.

[0432] Example 7-1, Rule #7 states: If the first indication information includes at least one fifth feature identifier, and the at least one fifth feature identifier is used to identify at least one fifteenth value corresponding to the fifth feature, or if the first indication information includes one or more fifth feature identifiers, and each fifth feature identifier is used to identify a set of fifth values ​​corresponding to the first feature, then if the values ​​corresponding to the fifth features of at least two datasets are included in at least one fifteenth value or at least one set of fifth values, then at least two datasets can be used for training the same model.

[0433] Based on rule #7 above, if at least one fifteenth value or at least one fifth value set includes the thirteenth and fourteenth values ​​(equivalent to the first indication information including the ninth and tenth identifiers), then the second device determines that the first dataset and the second dataset can be used for training the same model according to the first indication information; if at least one fifteenth value or at least one fifth value set does not include the thirteenth and / or fourteenth values ​​(equivalent to the first indication information not including the ninth and / or tenth identifiers), then the second device determines that the first dataset and the second dataset cannot be used for training the same model according to the first indication information.

[0434] Optionally, if the thirteenth value is the same as the fourteenth value, then even if at least one fifteenth value or at least one fifth value set does not include the thirteenth value and / or the fourteenth value (equivalent to the first indication information not including the ninth identifier and / or the tenth identifier), the second device can still determine that the first dataset and the second dataset can be used for training the same model.

[0435] For a more detailed description of rule #7 in Example 7-1, please refer to the description of rule #5 in Example 5-1 above.

[0436] Example 7-2, Rule #7 states: If the first indication information includes at least one feature identifier #e, and the at least one feature identifier #e is used to identify at least one value #e or at least one set of values ​​#e corresponding to the fifth feature, then if the values ​​corresponding to the fifth feature of at least two datasets are included in at least one value #e or at least one set of values ​​#e, then at least two datasets cannot be used for training the same model.

[0437] Based on rule #7 above, if at least one value #e or at least one set of values ​​#e includes the thirteenth and fourteenth values ​​(equivalent to the first indication information including the ninth and tenth identifiers), then the second device determines, according to the first indication information, that the first dataset and the second dataset cannot be used for training the same model; if at least one value #e or at least one set of values ​​#e does not include the thirteenth and / or fourteenth values ​​(equivalent to the first indication information not including the ninth and / or tenth identifiers), then the second device determines, according to the first indication information, that the first dataset and the second dataset can be used for training the same model.

[0438] Example 7-3, Rule #7 states: Datasets with different values ​​corresponding to the fifth feature can be used to train the same model. In other words, the first indication is used to indicate that datasets with different values ​​corresponding to the fifth feature can be used to train the same model.

[0439] Based on rule #7 above, the second device determines, according to the first instruction information, that the first dataset and the second dataset can be used for training the same model.

[0440] It should be noted that if the fifth feature corresponds to at least two features #e in the fifth feature set, then the values ​​of the fifth feature in the two datasets are different, which means that at least one of the at least two features #e in the two datasets has a different value.

[0441] Example 7-4, Rule #7 states: Datasets with different values ​​corresponding to the fifth feature cannot be used to train the same model. In other words, the first indication is used to indicate that datasets with different values ​​corresponding to the fifth feature cannot be used to train the same model.

[0442] Based on rule #7 above, if the thirteenth value is different from the fourteenth value, the second device determines that the first dataset and the second dataset cannot be used for training the same model; if the thirteenth value is the same as the fourteenth value, the second device determines that the first dataset and the second dataset can be used for training the same model.

[0443] It should be noted that if the fifth feature corresponds to at least two features #e in the fifth feature set, then the values ​​corresponding to the fifth feature in the two datasets are the same, which means that the values ​​corresponding to each feature #e in at least two features #e in the two datasets are the same.

[0444] Example 7-5, Rule #7 states: The values ​​corresponding to the fifth features of at least two datasets that can be used to train the same model must satisfy condition #7. In other words, the first indication information is used to indicate that the condition that the values ​​corresponding to the fifth features of at least two datasets that can be used to train the same model must satisfy is condition #7.

[0445] This application does not limit condition #7. For example, condition #7 is that the values ​​corresponding to the fifth feature in two different datasets are the same. Another example is that condition #7 is that the difference between the values ​​corresponding to the fifth feature in two different datasets does not exceed threshold #7.

[0446] Based on rule #7 above, if the thirteenth and fourteenth values ​​satisfy condition #7, the second device determines that the first dataset and the second dataset can be used for training the same model; if the thirteenth and fourteenth values ​​do not satisfy condition #7, the second device determines that the first dataset and the second dataset cannot be used for training the same model.

[0447] The following example, using the fifth feature set including features #19 to #21, illustrates how rule #7 is determined.

[0448] For example, rule #7 is related to Table 7 below.

[0449] Table 7

[0450] It should be understood that the different values ​​corresponding to each feature shown in Table 7 are merely examples, and this application does not limit them.

[0451] For example, if the fifth feature corresponds to one or more features from feature #19 to feature #21, then the first indication information may include rule #7 in Example 7-4 above.

[0452] For another example, if the fifth feature is feature #19 and the first indication information indicates rule #7 in example 7-2 above, then the first indication information includes at least one fifth feature identifier as identifier #11 to identifier #13, identifier #11 is used to identify the value corresponding to feature #19 as 5, identifier #12 is used to identify the value corresponding to feature #19 as 10, and identifier #13 is used to identify the value corresponding to feature #19 as 50.

[0453] Optionally, if the fifth feature set also includes features other than the fifth feature, for example, if the fifth feature set also includes feature #t, and feature #t is different from the fifth feature, then rule #7 also includes rules related to the value corresponding to feature #t. Correspondingly, in S1020, the first device also sends identifier #t1 to the second device, identifier #t1 being used to identify the value #t1 corresponding to feature #t in the first dataset; and in S1030, the first device also sends identifier #t2 to the second device, identifier #t2 being used to identify the value #t2 corresponding to feature #t in the second dataset.

[0454] The rules related to the value of feature #t can be found in the description of the rules related to the value of the fifth feature above. For example, rule #7 includes the following rules related to the value of feature #t: the first indication information includes at least one feature identifier #t, and the at least one feature identifier #t is used to identify at least one value #t3 corresponding to feature #t. If the value of feature #t in at least two datasets is included in at least one value #t3, then at least two datasets can be used to train the same model.

[0455] It should be noted that if rule #7 includes a rule related to the value of the fifth feature (denoted as rule #7-1) and a rule related to the value of feature #t (denoted as rule #7-2), then the second device determines that the first dataset and the second dataset can be used for training the same model, meaning that the second device determines that the first dataset and the second dataset can be used for training the same model according to rule #7-1 and rule #7-2; the second device determines that the first dataset and the second dataset cannot be used for training the same model, meaning that the second device determines that the first dataset and the second dataset cannot be used for training the same model according to rule #7-1 and / or rule #7-2.

[0456] Implementation method 10: The first instruction information indicates at least two of the above rules #1, #4 to #7.

[0457] It should be understood that if the first instruction information indicates at least two of the above rules #1, #4 to #7, then in S1020, the first device may also send an eleventh identifier associated with the first model to the second device, the eleventh identifier including an identifier related to the rule indicated by the first instruction information, and in S1020, the first device may also send a twelfth identifier associated with the second model to the second device, the twelfth identifier including an identifier related to the rule indicated by the first instruction information.

[0458] For example, if the first indication information indicates the above rules #1 and #4, then the eleventh identifier may include the identifier of the first model and the identifier of the first dataset, and the twelfth identifier may include the identifier of the second model and the identifier of the second dataset.

[0459] For example, if the first instruction information indicates the above rules #1, #4 to #6, then the eleventh identifier may include the identifier of the first model, the identifier of the first dataset, the above fifth identifier and the above seventh identifier, and the twelfth identifier may include the identifier of the second model, the identifier of the second dataset, the above sixth identifier and the above eighth identifier.

[0460] It should be understood that if the first instruction information indicates at least two of the above rules #1, #4 to #7, the second device determines whether the first dataset and the second dataset can be used for training the same model based on at least one of the rules indicated by the first instruction information.

[0461] For example, the second device determines that the first dataset and the second dataset can be used for training the same model based on the first instruction information, including: the second device determines that the first dataset and the second dataset can be used for training the same model based on each rule indicated by the first instruction information.

[0462] For example, the second device determines that the first dataset and the second dataset cannot be used for training the same model based on the first instruction information, including: the second device determining that the first dataset and the second dataset cannot be used for training the same model based on at least one rule indicated by the first instruction information.

[0463] Optionally, the multiple rules indicated by the first indication information may have different priorities, in which case the first device will determine whether the first dataset and the second dataset can be used for training the same model based on the rule with higher priority.

[0464] For example, if the first instruction indicates rules three and four, and rules three and four belong to rules #1, #4 to #7 mentioned above, with rule three having a higher priority than rule four, then the second device prioritizes determining whether the first and second datasets can be used for training the same model based on rule three. If the second device determines that the first and second datasets cannot be used for training the same model based on rule three, then the second device does not need to determine whether the first and second datasets can be used for training the same model based on rule four. If the second device determines that the first and second datasets can be used for training the same model based on rule three, then the second device continues to determine whether the first and second datasets can be used for training the same model based on rule four. If rule four is the lowest priority rule, and the second device determines that the first and second datasets can be used for training the same model based on rule four, then the second device ultimately determines that the first and second datasets can be used for training the same model.

[0465] This application does not limit the priority order of rules #1, #4 to #7. For example, the priority order of rules #1, #4 to #7 from high to low is: rule 1, rule #4, rule #5, rule #6 and rule 7.

[0466] For example, if the first instruction information indicates at least two of the rules #1, #4 to #7, the second device can determine the rules indicated by the first instruction information based on the received eleventh identifier and / or twelfth identifier. For instance, if the eleventh identifier includes the identifier of the first model and the description of the first dataset, the second device can determine that the first instruction information indicates the aforementioned rules #1 and #4.

[0467] For example, if the first instruction information indicates at least two rules from rule #1 to rule #7, then the rules indicated by the first instruction information are sorted according to their priority. For instance, if the first instruction information indicates rule #1 and rule #4, and rule #1 has a higher priority than rule #4, then rule #1 indicated by the first instruction information is placed before rule #4. Alternatively, if the first instruction information indicates at least two rules from rule #1 to rule #7, then the rules indicated by the first instruction information are sorted according to a predefined or preconfigured order. Alternatively, if the first instruction information indicates at least two rules from rule #1 to rule #7, then the first instruction information also indicates an identifier corresponding to each rule, thereby enabling the second device to determine which rules the first instruction information specifically indicates. The identifier corresponding to each rule from rule #1 to rule #7 may be predefined or preconfigured, or may be indicated by the first device to the second device; this application does not limit this.

[0468] For example, if the first instruction information indicates at least two of rules #1, #4 to #7, then at least two of the different rules indicated by the first instruction information can be carried in the same message, field, or information element; or, at least two of the different rules indicated by the first instruction information can be carried in different messages, fields, or information elements. For instance, if rule #1 and rule #4 indicated by the first instruction information are carried in the same message, then in step 1040, the first device can send message #4, which indicates rule #1 and rule #4. As another example, if rule #1 and rule #4 indicated by the first instruction information are carried in message #4, rule #1 and rule #4 can also be carried in the same field (or information element) of message #4, or in different fields (or information elements) of message #4. For example, message #4 may include field #4, and field #1 indicates rule #1 and rule #4. Alternatively, message #4 may include field #5 and field #6, where field #5 indicates rule #1 and field #6 indicates rule #4. For another example, if the first instruction information indicates rules #1 and #4, which are carried in different messages, then in S1040, the first device can send messages #5 and #6, where message #5 indicates rule #1 and message #6 indicates rule #4. For instance, when at least two items from different rules are carried in the same field or information element, different values ​​of the same field or information element can each correspond to at least two items from the different rules, or the same value of the same field or information element can simultaneously correspond to at least two items from the different rules (i.e., a combined instruction method).

[0469] In this embodiment, for different types of datasets, datasets of different formats, and datasets with different configurations, the first device can send different forms of indication information. This allows the second device to determine whether at least two datasets among multiple datasets can be used to train the same model, thereby saving data transmission overhead and enabling the second device to perform different training tasks based on the received datasets, ensuring the performance of the trained or updated model. For example, if the value corresponding to the fourth feature of the first dataset is the same as the value corresponding to the fourth feature of the second dataset, the first device can send first indication information including the aforementioned rule #5, so that the second device can determine whether the first dataset and the second dataset can be used to train the same model based on the value corresponding to the third feature of the first dataset and the value corresponding to the third feature of the second dataset.

[0470] The above describes how the second device determines whether the model parameters and / or datasets corresponding to at least two models can be used for training the same model based on the first instruction information from the first device. The following describes, with reference to Figure 11, how the second device determines whether the model parameters and / or datasets corresponding to at least two models can be used for training the same model based on predefined first information.

[0471] Figure 11 is a schematic flowchart of a communication method 1100 provided in an embodiment of this application. The steps of method 1100 will be described in detail below.

[0472] S1110, the second device sends the first request message.

[0473] Accordingly, the first device receives the first request message.

[0474] For a more detailed description of S1110, please refer to S1010 in Method 1000 above.

[0475] S1120, the first device sends the first model and the eleventh identifier.

[0476] Correspondingly, the second device receives the first model and the eleventh identifier.

[0477] The first model is used for model training. The eleventh identifier is associated with the first model.

[0478] Optionally, the first device may also send the first dataset corresponding to the first model.

[0479] Optionally, the first device may also send the first model structure identifier corresponding to the first model.

[0480] For example, the eleventh identifier may include one or more of the following: the identifier of the first model, the identifier of the first dataset, the aforementioned first identifier, the aforementioned third identifier, the aforementioned fifth identifier, the aforementioned seventh identifier, or the aforementioned ninth identifier. For example, as shown in FIG12, in S1120, the first device sends model structure 1 (i.e., an example of the first model structure identifier), model ID 1 (i.e., an example of the first model identifier), dataset ID 1 (i.e., an example of the first dataset identifier), configuration 1 ID 1 (i.e., an example of the first identifier), and configuration 2 ID 1 (i.e., an example of the third identifier) ​​to the second device.

[0481] S1130, the first device sends the second model and the twelfth identifier.

[0482] Correspondingly, the second device receives the second model and the twelfth identifier.

[0483] The second model is used for model training. The twelfth identifier is associated with the second model.

[0484] Optionally, the first device may also send a second dataset corresponding to the second model.

[0485] Optionally, the first device may also send the second model structure identifier corresponding to the second model.

[0486] For example, the twelfth identifier may include one or more of the following: the identifier of the second model, the identifier of the second dataset, the aforementioned second identifier, the aforementioned fourth identifier, the aforementioned sixth identifier, the aforementioned eighth identifier, or the aforementioned tenth identifier. For example, as shown in FIG12, in S1130, the first device sends model structure 2 (i.e., an example of the second model structure identifier), model identifier 2 (i.e., model ID 2 (i.e., an example of the second model identifier), dataset identifier 2 (i.e., dataset ID 2 (i.e., an example of the second dataset identifier), configuration 1 identifier 2 (i.e., configuration 1 ID 2 (i.e., an example of the second identifier)) and configuration 2 identifier 2 (i.e., configuration 2 ID 2 (i.e., an example of the fourth identifier)) to the second device.

[0487] S1140, the first device sends the first information.

[0488] Correspondingly, the second device receives the first information.

[0489] The first piece of information is associated with the eleventh and twelfth identifiers. This first piece of information is used to determine one or more of the following: whether the first dataset and the second dataset can be used to train the same model, or whether the first model parameters and the second model parameters can be used to train the same model. Specifically, the first dataset and the first model parameters correspond to the first model, and the second dataset and the second model parameters correspond to the second model.

[0490] It should be understood that S1140 is an optional step. For example, if the first information is pre-configured or predefined by the protocol, then method 1100 may not include S1140.

[0491] S1150, the second device determines whether the first model parameters and the second model parameters can be used for training the same model.

[0492] For example, when the second device receives the first model, the second model, the eleventh identifier, and the twelfth identifier, it can determine whether the parameters of the first model and the parameters of the second model can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0493] Optionally, if the model structure of the first model is the same as that of the second model, then method 1100 executes S1150.

[0494] The following describes how the second device determines whether the first model parameters and the second model parameters can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0495] In implementation method a, the eleventh identifier includes the identifier of the first model, the twelfth identifier includes the identifier of the second model, and the first information indicates rule #1 as described in method 1000 above. The second device determines whether the parameters of the first model and the parameters of the second model can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information. This can be referred to implementation method 2 as described in method 1000 above.

[0496] In implementation method b, the eleventh identifier includes a first identifier, which is used to identify the first value corresponding to the first feature of the first model. The twelfth identifier includes a second identifier, which is used to identify the second value corresponding to the first feature of the second model. The first information indicates rule #2 as described in method 1000 above. The way the second device determines whether the parameters of the first model and the parameters of the second model can be used for training the same model based on the eleventh identifier, the twelfth identifier and the first information can refer to implementation method 3 as described in method 1000 above.

[0497] In implementation method c, the eleventh identifier includes the third identifier, which is used to identify the fourth value corresponding to the second feature of the first model. The twelfth identifier includes the fourth identifier, which is used to identify the fifth value corresponding to the second feature of the second model. The first information indicates rule #3 as described in method 1000 above. The second device determines whether the parameters of the first model and the parameters of the second model can be used for training the same model based on the eleventh identifier, the twelfth identifier and the first information. This can be referred to implementation method 4 as described in method 1000 above.

[0498] In implementation d, the first information indicates at least two rules from rule #1 to rule #3 in method 1000 above, the eleventh identifier includes an identifier related to the rule indicated by the first information, and the twelfth identifier includes an identifier related to the rule indicated by the first information. The second device determines whether the first model parameters and the second model parameters can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information. This can be referred to implementation 5 in method 1000 above.

[0499] S1160, the second device determines whether the first dataset and the second dataset can be used to train the same model.

[0500] For example, if the second device receives the first dataset, the second dataset, the eleventh identifier, and the twelfth identifier, it can determine whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0501] The following describes how the second device determines whether the first dataset and the second dataset can be used to train the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0502] In implementation method f, the eleventh identifier includes the identifier of the first model, the twelfth identifier includes the identifier of the second model, and the first information indicates rule #1 as described in method 1000 above. The second device determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information. This can be referred to implementation method 2 as described in method 1000 above.

[0503] In implementation method g, the eleventh identifier includes the identifier of the first dataset, the twelfth identifier includes the identifier of the second dataset, and the first information indicates rule #4 as described in method 1000 above. The second device determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information. This can be referred to implementation method 6 as described in method 1000 above.

[0504] In implementation method h, the eleventh identifier includes the fifth identifier, which is used to identify the seventh value corresponding to the third feature of the first dataset. The twelfth identifier includes the sixth identifier, which is used to identify the eighth value corresponding to the third feature of the second dataset. The first information indicates rule #5 as described in method 1000 above. The second device determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information. This can be referred to implementation method 7 as described in method 1000 above.

[0505] In implementation method i, the eleventh identifier includes the seventh identifier, which is used to identify the tenth value corresponding to the fourth feature of the first dataset. The twelfth identifier includes the eighth identifier, which is used to identify the eleventh value corresponding to the fourth feature of the second dataset. The first information indicates rule #6 as described in method 1000 above. The second device determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information. This can be referred to implementation method 8 as described in method 1000 above.

[0506] In implementation j, the eleventh identifier includes the ninth identifier, which is used to identify the thirteenth value corresponding to the fifth feature of the first dataset. The twelfth identifier includes the tenth identifier, and the sixth identifier is used to identify the fourteenth value corresponding to the fifth feature of the second dataset. The first information indicates rule #7 as described in method 1000 above. The second device determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information. This can be referred to implementation 9 as described in method 1000 above.

[0507] In implementation k, the first information indicates at least two of the rules #1, #4 to #7 described in method 1000 above, the eleventh identifier includes an identifier related to the rule indicated by the first information, and the twelfth identifier includes an identifier related to the rule indicated by the first information. The second device determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier and the first information. This can be referred to in implementation 10 described in method 1000 above.

[0508] S1170, the second device performs model training.

[0509] S1170 can be referenced from S1070 in Method 1000 above.

[0510] In this embodiment, the second device can determine whether the model parameters and / or datasets corresponding to at least two models out of multiple models can be used for training the same model based on the first information and the identifier associated with the dataset. This facilitates the second device performing different training tasks based on the received multiple models while saving data overhead, ensuring the performance of the trained or updated models. For example, if the second device receives the identifier of the first model and the identifier of the second model, the second device can determine whether the parameters of the first model and the parameters of the second model can be used for training the same model based on the rules associated with the model identifiers.

[0511] As described above, the first device can be replaced by at least one of a device (such as a terminal device, access network device or core network element) or an AI entity on the device side. If the first device is replaced by a device and an AI entity on the device side, the following steps in the above embodiments can be performed by the AI ​​entity of the device: the step of sending the dataset, the step of sending the identifier related to the dataset, and the step of sending the first instruction information (or the first information).

[0512] The second device can be replaced by at least one of a device (such as a terminal device, access network device or core network element) or an AI entity on the device side. If the first device is replaced by a device and an AI entity on the device side, the following steps in the above embodiments can be performed by the AI ​​entity of the device: the step of receiving the dataset, the step of receiving the identifier related to the dataset, and the step of receiving the first instruction information (or the first information).

[0513] For example, Figure 13 illustrates a schematic flowchart of the method provided in this application from the perspective of the interaction between intelligent network elements, network devices, terminal devices, and OTT servers. The intelligent network element can be an AI entity on the network device side. The OTT server can be an AI entity on the terminal device side.

[0514] As shown in Figure 13, method 1200 may include the following steps.

[0515] S1201, the OTT server sends request message #1.

[0516] Accordingly, the terminal device receives request message #1.

[0517] Request message #1 is used to request at least one model for model training. Further description of request message #1 can be found in S1010 of method 1000 above.

[0518] It should be noted that S1201 is an optional step. For example, if the device used to train the model is a terminal device, then method 1200 may not include S1201.

[0519] S1202, the terminal device sends request message #2.

[0520] Accordingly, the network device receives request message #2.

[0521] Request message #2 is used to request at least one model for model training. Further description of request message #2 can be found in S1010 of method 1000 above.

[0522] S1203, the network device sends request message #3.

[0523] Accordingly, the terminal device receives request message #3.

[0524] Request message #3 is used to request at least one model for model training. Further description of request message #3 can be found in S1010 of method 1000 above.

[0525] It should be noted that S1203 is an optional step. For example, if the device used to provide the model is a network device, then method 1200 may not include S1203.

[0526] Optionally, if the device used to provide the model is an intelligent network element, then method 1200 further includes S1204 and S1205.

[0527] S1204, the intelligent network element sends the first model and the eleventh identifier.

[0528] Accordingly, the network device receives the first model and the eleventh identifier.

[0529] The first model is used for model training. The eleventh identifier is associated with the first model. For more details on the eleventh identifier, please refer to Method 1100 above.

[0530] S1205, the intelligent network element sends the second model and the twelfth identifier.

[0531] Accordingly, the network device receives the second model and the twelfth identifier.

[0532] The second model is used for model training. The twelfth identifier is associated with the second model. For more details on the twelfth identifier, please refer to Method 1100 above.

[0533] S1206, the network device sends the first model and the eleventh identifier.

[0534] Accordingly, the terminal device receives the first model and the eleventh identifier.

[0535] S1207, the network device sends the second model and the twelfth identifier.

[0536] Accordingly, the terminal device receives the second model and the twelfth identifier.

[0537] Optionally, if the device used to train the model is an OTT server, then method 1200 further includes S1208 and S1209.

[0538] S1208, the terminal device sends the first model and the eleventh identifier.

[0539] Accordingly, the OTT server receives the first model and the eleventh identifier.

[0540] S1209, the terminal device sends the second model and the twelfth identifier.

[0541] Accordingly, the OTT server receives the second model and the twelfth identifier.

[0542] Furthermore, if the device used to train the model is a terminal device, then method 1200 executes S1210 and / or S1211; if the device used to train the model is an OTT server, then method 1200 executes S1212 and / or S1213.

[0543] S1210, the terminal device determines whether the first model parameters and the second model parameters can be used for training the same model.

[0544] For example, the terminal device determines whether the first model parameters and the second model parameters can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0545] S1211, the terminal device determines whether the first dataset and the second dataset can be used to train the same model.

[0546] For example, the terminal device determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0547] S1212, the OTT server determines whether the first model parameters and the second model parameters can be used to train the same model.

[0548] For example, the OTT server determines whether the first model parameters and the second model parameters can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0549] S1213, the OTT server determines whether the first dataset and the second dataset can be used to train the same model.

[0550] For example, the OTT server determines whether the first dataset and the second dataset can be used for training the same model based on the eleventh identifier, the twelfth identifier, and the first information.

[0551] For further description of S1210 to S1211 or S1212 to S1213, please refer to S1150 to S1160 in Method 1100 above.

[0552] In one possible implementation, if the device used to train the model is an OTT server, then method 1200 can execute S1210 and / or S1211. Furthermore, method 1200 may further include the following steps: the terminal device sends indication information #a to the OTT server, where indication information #a indicates whether the first model parameters and the second model parameters can be used for training the same model, and / or whether the first dataset and the second dataset can be used for training the same model. Furthermore, in S1212, the OTT server can determine whether the first model parameters and the second model parameters can be used for training the same model based on the indication information #a from the terminal device; and / or, in S1213, the OTT server can determine whether the first dataset and the second dataset can be used for training the same model based on the indication information #a from the terminal device.

[0553] In one possible implementation, if the device used to train the model is a terminal device, then method 1200 may execute S1212 and / or S1213. Furthermore, method 1200 may further include the following steps: the OTT server sends indication information #b to the terminal device, whereby indication information #b indicates whether the first model parameters and the second model parameters can be used for training the same model, and / or whether the first dataset and the second dataset can be used for training the same model. Furthermore, in S1210, the terminal device may determine whether the first model parameters and the second model parameters can be used for training the same model based on the indication information #b from the OTT server; and / or, in S1211, the terminal device may determine whether the first dataset and the second dataset can be used for training the same model based on the indication information #b from the OTT server.

[0554] The methods provided in the embodiments of this application have been described in detail above with reference to several accompanying drawings. The apparatus provided in the embodiments of this application will now be described with reference to the accompanying drawings.

[0555] Figures 14 and 15 are schematic block diagrams of possible apparatuses provided in embodiments of this application. These apparatuses can be used to implement the terminal-side or network-side functions in the above method embodiments, and thus can also achieve the beneficial effects of the above method embodiments.

[0556] Figure 14 is a schematic block diagram of an apparatus provided in an embodiment of this application. The apparatus 1300 shown in Figure 14 may include a processing module 1310 and a communication module 1320.

[0557] In one possible design, device 1300 can be used to implement the communication method implemented by the first device in any of the embodiments shown in Figures 10, 13 to 15. For example, processing module 1310 is used to implement the processing-related steps performed by the first device in each method embodiment; communication module 1320 is used to implement the sending and / or receiving steps performed by the first device in each method embodiment, such as sending a first model, sending a second model, and sending first indication information, or one or more of these steps.

[0558] For example, the communication module 1320 is configured to: send a first model, the first model being used for model training; send a second model, the second model being used for model training; and send first indication information, the first indication information being used to determine one or more of the following: the model parameters corresponding to the first model and the model parameters corresponding to the second model can be used for training the same model, or the model parameters corresponding to the first model and the model parameters corresponding to the second model cannot be used for training the same model; the first dataset corresponding to the first model and the second dataset corresponding to the second model can be used for training the same model, or the first dataset corresponding to the first model and the second dataset corresponding to the second model cannot be used for training the same model.

[0559] A more detailed description of the processing module 1310 and the communication module 1320 can be obtained directly from the relevant descriptions in the method embodiments shown in Figures 10 to 13, and will not be repeated here.

[0560] In another possible design, device 1300 can be used to implement the communication method implemented by the second device in any of the embodiments shown in Figures 10 to 13. For example, processing module 1310 is used to implement the processing-related steps performed by the second device in each method embodiment; communication module 1320 is used to implement the sending and / or receiving steps performed by the second device in each method embodiment, such as receiving one or more of the following: receiving a first model, receiving a second model, and receiving first indication information.

[0561] For example, the communication module 1320 is configured to: receive a first model, the first model being used for model training; receive a second model, the second model being used for model training; and receive first indication information, the first indication information being configured to determine one or more of the following: the model parameters corresponding to the first model and the model parameters corresponding to the second model can be used for training the same model, or the model parameters corresponding to the first model and the model parameters corresponding to the second model cannot be used for training the same model; the first dataset corresponding to the first model and the second dataset corresponding to the second model can be used for training the same model, or the first dataset corresponding to the first model and the second dataset corresponding to the second model cannot be used for training the same model.

[0562] A more detailed description of the processing module 1310 and the communication module 1320 can be obtained directly from the relevant descriptions in the method embodiments shown in Figures 10 to 13, and will not be repeated here.

[0563] It should be noted that the communication module can also be called a transceiver module, transceiver unit, transceiver, transceiver device, or transceiver apparatus, etc. The processing module can also be called a processor, processing board, processing unit, or processing apparatus, etc. Optionally, the communication module is used to execute the sending and receiving operations of the first or second communication device in the above method. The device in the communication module that implements the receiving function can be considered as the receiving module, and the device in the communication module that implements the sending function can be considered as the sending module; that is, the communication module can include both a receiving module and a sending module.

[0564] It should also be noted that, in one possible design, the aforementioned processing module and / or communication module can be implemented through virtual modules. For example, the processing module can be implemented through software functional units or virtual devices, and the communication module can be implemented through software functions or virtual devices. In another possible design, the processing module or communication module can also be implemented through physical devices. For example, if the device is implemented using a chip / chip circuit, the communication module can be an input / output circuit and / or a communication interface, performing input operations (corresponding to the aforementioned receiving operation) and output operations (corresponding to the aforementioned sending operation); the processing module can be an integrated processor, a microprocessor, or an integrated circuit.

[0565] The module division in this embodiment is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. Furthermore, the functional modules in the various examples of this embodiment can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0566] Figure 15 is a schematic diagram of the structure of a communication device provided in another embodiment of this application. As shown in Figure 15, the device 1400 includes a processing circuit 1410 and a communication circuit 1420. The processing circuit 1410 and the communication circuit 1420 are coupled to each other.

[0567] It can be understood that the processing circuit 1410 can be one or more processors, or it can be all or part of the circuits with processing functions in one or more processors.

[0568] Understandably, the communication circuit 1420 can be a transceiver or an input / output interface.

[0569] Optionally, the device 1400 may further include a memory 1430 for storing instructions executed by the processing circuit 1410, or storing input data required for the running instructions of the processing circuit 1410, or storing data generated after the running instructions of the processing circuit 1410.

[0570] It is understood that the memory 1430 may be located outside the processing circuit 1410, or inside the processing circuit 1410.

[0571] As an example, the processing circuit 1410 is used to implement the functions of the processing module 1310, and the communication circuit 1420 is used to implement the functions of the communication module 1320.

[0572] As an example, device 1400 can be a communication device or a chip used in a communication device.

[0573] When device 1400 is a communication device, the communication circuit can be a transceiver; when device 1400 is a chip, the communication circuit can be an input / output circuit, a bus, pins, or other types of communication interfaces. The input circuit in the input / output circuit can be used for receiving, and the output interface can be used for transmitting.

[0574] In one possible implementation, the apparatus 1400 is used to implement the various processes and steps corresponding to the first apparatus in the above method embodiments. In another possible implementation, the apparatus 1400 is used to implement the various processes and steps corresponding to the second apparatus in the above method embodiments.

[0575] It is understood that device 1400 may specifically be the first device or the second device in the above embodiments, or it may be a chip or a chip system. Correspondingly, the communication circuit may be the interface circuit of the chip, or an input / output circuit, which is not limited here. Specifically, device 1400 may be used to execute the various steps and / or processes corresponding to the first device or the second device in the above method embodiments.

[0576] When the aforementioned communication device is a chip or OTT device applied to the first device, the chip or OTT device of the first device implements the functions of the first device in the above method embodiments, for example, implementing the processing functions of the first device. The chip or OTT device of the first device receiving information from the second device can be understood as the information being first received by other modules (such as radio frequency modules or antennas) in the first device, and then sent by these modules to the chip or OTT device of the first device. The chip or OTT device of the first device sending information to the second device can be understood as the information being first sent by the chip or OTT device of the first device to other modules (such as radio frequency modules or antennas) in the first device, and then sent by these modules to the second device.

[0577] When the aforementioned communication device is a chip or OTT device applied to the second device, the chip or OTT device of the second device implements the functions of the second device in the above method embodiments, for example, implementing the processing functions of the second device. The chip or OTT device of the second device receiving information from the first device can be understood as the information being first received by other modules (such as radio frequency modules or antennas) in the second device, and then sent by these modules to the chip or OTT device of the second device. The chip or OTT device of the second device sending information to the first device can be understood as the information being first sent by the chip or OTT device of the second device to other modules (such as radio frequency modules or antennas) in the second device, and then sent by these modules to the first device.

[0578] This application also provides a computer program product that, when run on a processor, can implement the communication method executed by the first device or the communication method executed by the second device in the above method embodiments.

[0579] This application also provides a computer-readable storage medium containing computer instructions that, when executed on a processor, can implement the communication method executed by the first device or the communication method executed by the second device in the above method embodiments.

[0580] This application also provides a communication system, including the aforementioned first device and second device.

[0581] It is understood that the processor in the embodiments of this application may be any of the following devices or all or part of the circuitry used for processing functions: a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), processors for AI, field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor may be a microprocessor or any conventional processor.

[0582] For example, the processor used for AI can be one or more of the following: graphics processing unit (GPU), neural processing unit (NPU), tensor processing unit (TPU), and data processing unit (DPU).

[0583] For example, one possible implementation of a processor for AI could be the AI ​​processor 1500 shown in Figure 16.

[0584] As shown in Figure 16, the AI ​​processor 1500 may include one or more of the following: an AI core, a digital vision pre-processing (DVPP) module, a task scheduler (TS), an L3 cache, an AI CPU, a control CPU, an L2 cache, a universal serial bus (USB) interface, a network interface card (NIC), a peripheral component interconnect express (PCIe) interface (PCIe is a high-speed serial computer expansion bus standard), a double data rate (DDR) / high bandwidth memory (HBM) interface, a generation purpose input / output (GPIO) / inter-integrated circuit (I2C) bus, etc. It is understood that the specific meanings of these terms are well known to those skilled in the art and will not be elaborated upon here.

[0585] The terms “unit”, “module”, etc., used in this specification may be used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution.

[0586] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0587] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0588] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0589] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0590] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0591] In the above embodiments, the functions of each functional unit can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. This computer program product includes one or more computer instructions (programs). When the computer program instructions (programs) are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).

[0592] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0593] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A communication method, characterized in that, include: Receive a first model, which is used for model training; Receive the second model, which is used for model training; Receive first indication information, the first indication information being used to determine one or more of the following: The first model parameters corresponding to the first model and the second model parameters corresponding to the second model can be used to train the same model, or the first model parameters and the second model parameters cannot be used to train the same model. The first dataset corresponding to the first model and the second dataset corresponding to the second model can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

2. The method according to claim 1, characterized in that, The method further includes: The model structure of the first model is the same as that of the second model. Based on the first indication information, one or more of the following are determined: The first model parameters and the second model parameters can be used to train the same model, or the first model parameters and the second model parameters cannot be used to train the same model. The first dataset and the second dataset can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model; or, The model structure of the first model is different from that of the second model. Based on the first indication information, determine one of the following: The first dataset and the second dataset can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

3. The method according to claim 1 or 2, characterized in that, The receiving of the first model includes: Receive the first model and its identifier; The receiving of the second model includes: Receive the second model and its identifier; The first indication information includes identifiers for multiple models, and the model parameters and / or datasets corresponding to the multiple models can be used to train the same model; or, The first indication information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers can be used for training the same model; or, The first indication information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the corresponding model parameters and / or datasets must meet for the identification of at least two models that can be used to train the same model.

4. The method according to any one of claims 1 to 3, characterized in that, The first indication information is used to determine whether the first model parameters and the second model parameters can be used to train the same model, or to determine whether the first model parameters and the second model parameters can be used to train the same model. The receiving of the first model includes: The system receives the first model and a first identifier, wherein the first identifier is used to identify a first value corresponding to a first feature of the first model; and the first feature belongs to a first feature set. Receiving the second model includes: Receive the second model and the second identifier, wherein the second identifier is used to identify the second value corresponding to the first feature of the second model; The first indication information includes multiple first feature identifiers, which are used to identify multiple third values ​​corresponding to the first feature; or, the first indication information includes one or more first feature identifiers, each first feature identifier being used to identify a set of first values ​​corresponding to the first feature; wherein, the values ​​corresponding to the first feature of at least two models are included in the multiple third values ​​or at least one set of first values, and the model parameters corresponding to at least two models can be used for training the same model; or... The first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the first feature can be used for training the same model; or, The first indication information is used to indicate that model parameters corresponding to models with different values ​​for the first feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the values ​​of the first feature corresponding to the corresponding model parameters must satisfy for training at least two models that can be used for the same model.

5. The method according to claim 4, characterized in that, The first feature set includes one or more of the following: Learning rate, retention probability to prevent overfitting, total number of training epochs, batch size, size of training dataset, neural network dimension, number of neural network layers.

6. The method according to any one of claims 1 to 5, characterized in that, The first indication information is used to determine that the first model parameters and the second model parameters can be used to train the same model, and to determine that the first model parameters and the second model parameters cannot be used to train the same model. The receiving of the first model includes: The system receives the first model and a third identifier, wherein the third identifier is used to identify the fourth value corresponding to the second feature of the first model; the second feature belongs to the second feature set. The receiving of the second model includes: Receive the second model and the fourth identifier, wherein the fourth identifier is used to identify the fifth value corresponding to the second feature of the second model; The first indication information includes multiple second feature identifiers, which are used to identify multiple sixth values ​​corresponding to the second feature; or, the first indication information includes one or more second feature identifiers, each second feature identifier being used to identify a set of second values ​​corresponding to the second feature; wherein, the values ​​corresponding to the second feature of at least two models are included in the multiple sixth values ​​or at least one set of second values, and the model parameters corresponding to at least two models can be used for training the same model; or... The first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the second feature can be used to train the same model; or, The first indication information is used to indicate that model parameters corresponding to models with different values ​​for the second feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the values ​​of the second feature corresponding to the corresponding model parameters must satisfy for training at least two models that can be used for the same model.

7. The method according to claim 6, characterized in that, The second feature set includes one or more of the following: The configuration of the historical time window length corresponding to the data in the training dataset of the model, the configuration of the prediction time window length corresponding to the data in the training dataset of the model, and the function for compressing channel state information (CSI) and / or predicting CSI.

8. The method according to any one of claims 1 to 7, characterized in that, The first indication information indicates a first rule and a second rule, with the first rule having a higher priority than the second rule. The first rule and the second rule belong to the following rules used to determine whether the model parameters corresponding to at least two models can be used or cannot be used for training the same model: rules related to the model's identifier, rules related to the values ​​corresponding to the features included in the first feature set of the model, and rules related to the values ​​corresponding to the features included in the second feature set of the model. The method further includes: Based on the first rule, it is determined that the first model parameters and the second model parameters can be used to train the same model, or it is determined that the first model parameters and the second model parameters cannot be used to train the same model.

9. The method according to claim 8, characterized in that, The method further includes determining, based on the first rule, that the first model parameters and the second model parameters can be used for training the same model. The first model parameters and the second model parameters are determined to be usable for training the same model according to the second rule, or the first model parameters and the second model parameters are determined not to be usable for training the same model.

10. The method according to any one of claims 1 to 9, characterized in that, The first indication information is used to determine whether the first dataset and the second dataset can be used to train the same model, or to determine whether the first dataset and the second dataset can be used to train the same model. The receiving of the first model includes: Receive the first model, the first dataset, and the identifier of the first dataset; The receiving of the second model includes: Receive the second model, the second dataset, and the identifier of the second dataset; The first indication information includes identifiers of multiple datasets, which can be used to train the same model; or, The first indication information is used to indicate that datasets with different identifiers can be used to train the same model; or, The first indication information is used to indicate that datasets with different identifiers cannot be used to train the same model; or, The first indication information is used to indicate the conditions that the identifiers of at least two datasets that can be used to train the same model must meet.

11. The method according to any one of claims 1 to 10, characterized in that, The first indication information is used to determine whether the first dataset and the second dataset can be used to train the same model, or to determine whether the first dataset and the second dataset can be used to train the same model. The receiving of the first model includes: The system receives the first model, the first dataset, and a fifth identifier, wherein the fifth identifier is used to identify the seventh value corresponding to the third feature of the first dataset, and the third feature belongs to the third feature set. Receiving the second model includes: The system receives the second model, the second dataset, and a sixth identifier, wherein the sixth identifier is used to identify the eighth value corresponding to the third feature of the second dataset. The first indication information includes multiple third feature identifiers, which are used to identify multiple ninth values ​​corresponding to the third feature; or, the first indication information includes one or more third feature identifiers, each of which is used to identify a set of third values ​​corresponding to the third feature; wherein, the values ​​corresponding to the third feature in at least two datasets are included in the multiple ninth values ​​or at least one set of third values, and at least two datasets can be used for training the same model; or... The first indication information is used to indicate that datasets with different values ​​corresponding to the third feature can be used to train the same model; or, The first indication information is used to indicate that datasets with different values ​​corresponding to the third feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the value corresponding to the third feature must satisfy in at least two datasets that can be used to train the same model.

12. The method according to claim 11, characterized in that, The third feature set includes one or more of the following: Whether the data was acquired during training with a quantizer, the range of absolute received power distribution, the range of received power distribution, the range of received noise power distribution, the range of time delay spread, the range of angle spread, the range of interference level, and the range of signal-to-interference-plus-noise ratio (SINR).

13. The method according to any one of claims 1 to 12, characterized in that, The first indication information is used to indicate whether the first dataset and the second dataset can be used for training the same model, or to indicate that the first dataset and the second dataset cannot be used for training the same model; The receiving of the first model includes: The system receives the first model, the first dataset, and a seventh identifier, wherein the seventh identifier is used to identify the tenth value corresponding to the fourth feature of the first dataset, and the fourth feature belongs to the fourth feature set. The receiving of the second model includes: The system receives the second model, the second dataset, and an eighth identifier, wherein the eighth identifier is used to identify the eleventh value corresponding to the fourth feature of the second dataset. The first indication information includes multiple fourth feature identifiers, which are used to identify multiple twelfth values ​​corresponding to the fourth feature; or, the first indication information includes one or more fourth feature identifiers, each of which is used to identify a set of fourth values ​​corresponding to the fourth feature; wherein, the values ​​corresponding to the fourth feature in at least two datasets are included in the multiple twelfth values ​​or at least one set of fourth values, and at least two datasets can be used for training the same model; or... The first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature can be used to train the same model; or, The first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the value corresponding to the fourth feature of at least two datasets that can be used to train the same model must satisfy.

14. The method according to claim 13, characterized in that, The fourth feature set includes one or more of the following: The data in the dataset includes the number of receiving ports, the carrier frequency, the receiving bandwidth, the number of receiving sub-bands, the receiving sub-band granularity, and the number of bits of Channel State Information (CSI) feedback overhead.

15. The method according to any one of claims 1 to 14, characterized in that, The first indication information is used to determine whether the first dataset and the second dataset can be used for training the same model, or to determine whether the first dataset and the second dataset can be used for training the same model. The receiving of the first model includes: The system receives the first model, the first dataset, and the ninth identifier, wherein the ninth identifier is used to identify the thirteenth value corresponding to the fifth feature of the first dataset, and the fifth feature belongs to the fifth feature set. The receiving of the second model includes: The system receives the second model, the second dataset, and the tenth identifier, wherein the tenth identifier is used to identify the fourteenth value corresponding to the fifth feature of the second dataset. The first indication information includes multiple fifth feature identifiers, which are used to identify multiple fifteenth values ​​corresponding to the fifth feature; or, the first indication information includes one or more fifth feature identifiers, each of which is used to identify a set of fifth values ​​corresponding to the fifth feature; wherein, the values ​​corresponding to the fifth feature in at least two datasets are included in the multiple fifteenth values ​​or at least one set of fifth values, and at least two datasets can be used for training the same model; or... The first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature can be used to train the same model; or, The first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the values ​​corresponding to the fifth feature of at least two datasets that can be used to train the same model must satisfy.

16. The method according to claim 15, characterized in that, The fifth feature set includes one or more of the following: The length of the historical time window corresponding to the data in the dataset, the length of the prediction time window corresponding to the data in the dataset, and the moving speed of the device corresponding to the data in the dataset.

17. The method according to any one of claims 1 to 16, characterized in that, The first indication information indicates the third rule and the fourth rule, wherein the third rule has a higher priority than the fourth rule. The third rule and the fourth rule belong to the following rules used to determine whether the datasets corresponding to at least two models can be used or not used for training the same model: rules related to the model's identifier, rules related to the dataset's identifier, rules related to the values ​​corresponding to the features included in the third feature set of the dataset, rules related to the values ​​corresponding to the features included in the fourth feature set of the dataset, and rules related to the values ​​corresponding to the features included in the fifth feature set of the dataset. The method further includes: The third rule determines whether the first dataset and the second dataset can be used to train the same model, or whether the first dataset and the second dataset cannot be used to train the same model.

18. The method according to claim 17, characterized in that, The method further includes determining, based on the third rule, that the first dataset and the second dataset can be used for training the same model, and that: According to the fourth rule, it is determined that the first dataset and the second dataset can be used to train the same model, or that the first dataset and the second dataset cannot be used to train the same model.

19. A communication method, characterized in that, include: Send the first model, which is used for model training; Send the second model, which is used for model training; Send a first indication message, the first indication message being used to determine one or more of the following: The first model parameters corresponding to the first model and the second model parameters corresponding to the second model can be used to train the same model, or the first model parameters and the second model parameters cannot be used to train the same model. The first dataset corresponding to the first model and the second dataset corresponding to the second model can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

20. The method according to claim 19, characterized in that, The method further includes: The model structure of the first model is the same as that of the second model, and the first indication information is used to determine one or more of the following: The first model parameters and the second model parameters can be used to train the same model, or the first model parameters and the second model parameters cannot be used to train the same model. The first dataset and the second dataset can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model; or, The model structure of the first model is different from that of the second model, and the first indication information is used to determine one of the following: The first dataset and the second dataset can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

21. The method according to claim 19 or 20, characterized in that, The sending of the first model includes: Send the first model and its identifier; The sending of the second model includes: Send the second model and its identifier; The first indication information includes identifiers for multiple models, and the model parameters and / or datasets corresponding to the multiple models can be used to train the same model; or, The first indication information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers can be used for training the same model; or, The first indication information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the corresponding model parameters and / or datasets must meet for the identification of at least two models that can be used to train the same model.

22. The method according to any one of claims 19 to 21, characterized in that, The first indication information is used to determine whether the first model parameters and the second model parameters can be used to train the same model, or to determine whether the first model parameters and the second model parameters can be used to train the same model. The sending of the first model includes: Send the first model and the first identifier, wherein the first identifier is used to identify the first value corresponding to the first feature of the first model; the first feature belongs to the first feature set. Sending the second model includes: Send the second model and the second identifier, whereby the second identifier is used to identify the second value corresponding to the first feature of the second model; The first indication information includes multiple first feature identifiers, which are used to identify multiple third values ​​corresponding to the first feature; or, the first indication information includes one or more first feature identifiers, each first feature identifier being used to identify a set of first values ​​corresponding to the first feature; wherein, the values ​​corresponding to the first feature of at least two models are included in the multiple third values ​​or at least one set of first values, and the model parameters corresponding to at least two models can be used for training the same model; or... The first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the first feature can be used for training the same model; or, The first indication information is used to indicate that model parameters corresponding to models with different values ​​for the first feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the values ​​of the first feature corresponding to the corresponding model parameters must satisfy for training at least two models that can be used for the same model.

23. The method according to claim 22, characterized in that, The first feature set includes one or more of the following: Learning rate, retention probability to prevent overfitting, total number of training epochs, batch size, size of training dataset, neural network dimension, number of neural network layers.

24. The method according to any one of claims 19 to 23, characterized in that, The first indication information is used to determine that the first model parameters and the second model parameters can be used to train the same model, and to determine that the first model parameters and the second model parameters cannot be used to train the same model. The sending of the first model includes: Send the first model and the third identifier, wherein the third identifier is used to identify the fourth value corresponding to the second feature of the first model; the second feature belongs to the second feature set. The sending of the second model includes: Send the second model and the fourth identifier, wherein the fourth identifier is used to identify the fifth value corresponding to the second feature of the second model; The first indication information includes multiple second feature identifiers, which are used to identify multiple sixth values ​​corresponding to the second feature; or, the first indication information includes one or more second feature identifiers, each second feature identifier being used to identify a set of second values ​​corresponding to the second feature; wherein, the values ​​corresponding to the second feature of at least two models are included in the multiple sixth values ​​or at least one set of second values, and the model parameters corresponding to at least two models can be used for training the same model; or... The first indication information is used to indicate that the model parameters corresponding to models with different values ​​corresponding to the second feature can be used to train the same model; or, The first indication information is used to indicate that model parameters corresponding to models with different values ​​for the second feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the values ​​of the second feature corresponding to the corresponding model parameters must satisfy for training at least two models that can be used for the same model.

25. The method according to claim 24, characterized in that, The second feature set includes one or more of the following: The configuration of the historical time window length corresponding to the data in the training dataset of the model, the configuration of the prediction time window length corresponding to the data in the training dataset of the model, and the function for compressing channel state information (CSI) and / or predicting CSI.

26. The method according to any one of claims 19 to 25, characterized in that, The first indication information is used to determine whether the first dataset and the second dataset can be used to train the same model, or to determine whether the first dataset and the second dataset can be used to train the same model. The sending of the first model includes: Send the identifiers of the first model, the first dataset, and the first dataset; The sending of the second model includes: Send the identifiers of the second model, the second dataset, and the second dataset; The first indication information includes identifiers of multiple datasets, which can be used to train the same model; or, The first indication information is used to indicate that datasets with different identifiers can be used to train the same model; or, The first indication information is used to indicate that datasets with different identifiers cannot be used to train the same model; or, The first indication information is used to indicate the conditions that the identifiers of at least two datasets that can be used to train the same model must meet.

27. The method according to any one of claims 19 to 26, characterized in that, The first indication information is used to determine whether the first dataset and the second dataset can be used to train the same model, or to determine whether the first dataset and the second dataset can be used to train the same model. The sending of the first model includes: Send the first model, the first dataset, and the fifth identifier, wherein the fifth identifier is used to identify the seventh value corresponding to the third feature of the first dataset, and the third feature belongs to the third feature set; Sending the second model includes: Send the second model, the second dataset, and the sixth identifier, wherein the sixth identifier is used to identify the eighth value corresponding to the third feature of the second dataset; The first indication information includes multiple third feature identifiers, which are used to identify multiple ninth values ​​corresponding to the third feature; or, the first indication information includes one or more third feature identifiers, each of which is used to identify a set of third values ​​corresponding to the third feature; wherein, the values ​​corresponding to the third feature in at least two datasets are included in the multiple ninth values ​​or at least one set of third values, and at least two datasets can be used for training the same model; or... The first indication information is used to indicate that datasets with different values ​​corresponding to the third feature can be used to train the same model; or, The first indication information is used to indicate that datasets with different values ​​corresponding to the third feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the value corresponding to the third feature must satisfy in at least two datasets that can be used to train the same model.

28. The method according to claim 27, characterized in that, The third feature set includes one or more of the following: Whether the data was acquired during training with a quantizer, the range of absolute received power distribution, the range of received power distribution, the range of received noise power distribution, the range of time delay spread, the range of angle spread, the range of interference level, and the range of signal-to-interference-plus-noise ratio (SINR).

29. The method according to any one of claims 19 to 28, characterized in that, The first indication information is used to indicate whether the first dataset and the second dataset can be used for training the same model, or to indicate that the first dataset and the second dataset cannot be used for training the same model; The sending of the first model includes: Send the first model, the first dataset, and the seventh identifier, wherein the seventh identifier is used to identify the tenth value corresponding to the fourth feature of the first dataset, and the fourth feature belongs to the fourth feature set; The sending of the second model includes: Send the second model, the second dataset, and the eighth identifier, wherein the eighth identifier is used to identify the eleventh value corresponding to the fourth feature of the second dataset; The first indication information includes multiple fourth feature identifiers, which are used to identify multiple twelfth values ​​corresponding to the fourth feature; or, the first indication information includes one or more fourth feature identifiers, each of which is used to identify a set of fourth values ​​corresponding to the fourth feature; wherein, the values ​​corresponding to the fourth feature in at least two datasets are included in the multiple twelfth values ​​or at least one set of fourth values, and at least two datasets can be used for training the same model; or... The first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature can be used to train the same model; or, The first indication information is used to indicate that datasets with different values ​​corresponding to the fourth feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the value corresponding to the fourth feature of at least two datasets that can be used to train the same model must satisfy.

30. The method according to claim 29, characterized in that, The fourth feature set includes one or more of the following: The data in the dataset includes the number of receiving ports, the carrier frequency, the receiving bandwidth, the number of receiving sub-bands, the receiving sub-band granularity, and the number of bits of Channel State Information (CSI) feedback overhead.

31. The method according to any one of claims 19 to 30, characterized in that, The first indication information is used to determine whether the first dataset and the second dataset can be used for training the same model, or to determine whether the first dataset and the second dataset can be used for training the same model. The sending of the first model includes: Send the first model, the first dataset, and the ninth identifier, wherein the ninth identifier is used to identify the thirteenth value corresponding to the fifth feature of the first dataset, and the fifth feature belongs to the fifth feature set; The sending of the second model includes: Send the second model, the second dataset, and the tenth identifier, wherein the tenth identifier is used to identify the fourteenth value corresponding to the fifth feature of the second dataset; The first indication information includes multiple fifth feature identifiers, which are used to identify multiple fifteenth values ​​corresponding to the fifth feature; or, the first indication information includes one or more fifth feature identifiers, each of which is used to identify a set of fifth values ​​corresponding to the fifth feature; wherein, the values ​​corresponding to the fifth feature in at least two datasets are included in the multiple fifteenth values ​​or at least one set of fifth values, and at least two datasets can be used for training the same model; or... The first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature can be used to train the same model; or, The first indication information is used to indicate that datasets with different values ​​corresponding to the fifth feature cannot be used for training the same model; or, The first indication information is used to indicate the conditions that the values ​​corresponding to the fifth feature of at least two datasets that can be used to train the same model must satisfy.

32. The method according to claim 31, characterized in that, The fifth feature set includes one or more of the following: The length of the historical time window corresponding to the data in the dataset, the length of the prediction time window corresponding to the data in the dataset, and the moving speed of the device corresponding to the data in the dataset.

33. A communication method, characterized in that, include: Receive a first model and an eleventh identifier, wherein the eleventh identifier is associated with the first model, and the first model is used for model training; Receive a second model and a twelfth identifier, the twelfth identifier being associated with the second model, the second model being used for model training; Based on the first information associated with the eleventh identifier and the twelfth identifier, second information is determined, the second information including one or more of the following: the first model parameters corresponding to the first model and the second model parameters corresponding to the second model can be used for training the same model, or the first model parameters and the second model parameters cannot be used for training the same model; or the first dataset corresponding to the first model and the second dataset corresponding to the second model can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model.

34. The method according to claim 33, characterized in that, The model structure of the first model is the same as that of the second model. The second information includes one or more of the following: the parameters of the first model and the parameters of the second model can be used for training the same model, or the parameters of the first model and the parameters of the second model cannot be used for training the same model; the first dataset and the second dataset can be used for training the same model, or the first dataset and the second dataset cannot be used for training the same model. or, The model structure of the first model is different from that of the second model. The second information includes one of the following: the first dataset and the second dataset can be used to train the same model, or the first dataset and the second dataset cannot be used to train the same model.

35. The method according to claim 33 or 34, characterized in that, The eleventh identifier includes the identifier of the first model, and the twelfth identifier includes the identifier of the second model; The first information includes the identifiers of multiple models, and the model parameters and / or datasets corresponding to the multiple models can be used to train the same model; or, The first information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers can be used to train the same model; or, The first information is used to indicate that model parameters and / or datasets corresponding to models with different identifiers cannot be used for training the same model; or, The first information is used to indicate the conditions that the corresponding model parameters and / or datasets must meet for the identification of at least two models that can be used to train the same model.

36. The method according to any one of claims 33 to 35, characterized in that, The eleventh identifier includes a first identifier, which is used to identify a first value corresponding to a first feature of the first model; the twelfth identifier includes a second identifier, which is used to identify a second value corresponding to the first feature of the second model; the first feature belongs to a first feature set; The first information includes multiple first feature identifiers, which are used to identify multiple third values ​​corresponding to the first feature; or, the first information includes one or more first feature identifiers, each first feature identifier being used to identify a set of first values ​​corresponding to the first feature; wherein, the values ​​corresponding to the first feature of at least two models are included in the multiple third values ​​or at least one set of first values, and the model parameters corresponding to the at least two models can be used for training the same model; or... The first information is used to indicate that the model parameters corresponding to models with different values ​​of the first feature can be used to train the same model; or, The first information is used to indicate that model parameters corresponding to models with different values ​​of the first feature cannot be used for training the same model; or, The first information is used to indicate the conditions that the values ​​of the first feature corresponding to the corresponding model parameters must satisfy for training at least two models that can be used for training the same model.

37. The method according to claim 36, characterized in that, The first feature set includes one or more of the following: learning rate, retention probability to prevent overfitting, total number of training epochs, batch size, size of training dataset, neural network dimension, and number of neural network layers.

38. The method according to any one of claims 33 to 37, characterized in that, The eleventh identifier includes a third identifier, which is used to identify the fourth value corresponding to the second feature of the first model; the twelfth identifier includes a fourth identifier, which is used to identify the fifth value corresponding to the second feature of the second model; the second feature belongs to the second feature set. The first information includes multiple second feature identifiers, which are used to identify multiple sixth values ​​corresponding to the second feature; or, the first information includes one or more second feature identifiers, each second feature identifier being used to identify a set of second values ​​corresponding to the second feature; wherein, the values ​​corresponding to the second feature of at least two models are included in the multiple sixth values ​​or at least one set of second values, and the model parameters corresponding to the at least two models can be used for training the same model; or... The first information is used to indicate that the model parameters corresponding to models with different values ​​of the second feature can be used to train the same model; or, The first information is used to indicate that model parameters corresponding to models with different values ​​for the second feature cannot be used for training the same model; or, The first information is used to indicate the conditions that the values ​​of the second feature corresponding to the corresponding model parameters must satisfy for training at least two models that can be used for training the same model.

39. The method according to claim 38, characterized in that, The second feature set includes one or more of the following: the configuration of the historical time window length corresponding to the data in the training dataset corresponding to the model, the configuration of the prediction time window length corresponding to the data in the training dataset corresponding to the model, and the function for compressing CSI and / or predicting CSI.

40. The method according to any one of claims 33 to 39, characterized in that, The first information indicates the first rule and the second rule, wherein the first rule has a higher priority than the second rule. The first rule and the second rule belong to the following rules used to determine whether the model parameters corresponding to at least two models can be used or cannot be used for training the same model: rules related to the model's identifier, rules related to the values ​​corresponding to the features included in the first feature set of the model, and rules related to the values ​​corresponding to the features included in the second feature set of the model. The step of determining the second information based on the first information associated with the eleventh identifier and the twelfth identifier includes: Based on the eleventh identifier, the twelfth identifier, and the first rule, it is determined that the first model parameters and the second model parameters can be used for training the same model, or that the first model parameters and the second model parameters cannot be used for training the same model.

41. The method according to claim 40, characterized in that, The method further includes determining, based on the eleventh identifier, the twelfth identifier, and the first rule, that the first model parameters and the second model parameters can be used for training the same model. Based on the eleventh identifier, the twelfth identifier, and the second rule, it is determined that the first model parameters and the second model parameters can be used for training the same model, or that the first model parameters and the second model parameters cannot be used for training the same model.

42. The method according to any one of claims 33 to 41, characterized in that, The eleventh identifier includes the identifier of the first dataset, and the twelfth identifier includes the identifier of the second dataset; The first information includes the identifiers of multiple datasets, which can be used to train the same model; or, The first piece of information is used to indicate that datasets with different identifiers can be used to train the same model; or, The first piece of information is used to indicate that datasets with different identifiers cannot be used to train the same model; or, The first information is used to indicate the conditions that the identifiers of at least two datasets that can be used to train the same model must meet.

43. The method according to any one of claims 33 to 42, characterized in that, The eleventh identifier includes the fifth identifier, which is used to identify the seventh value corresponding to the third feature of the first dataset; the twelfth identifier includes the sixth identifier, which is used to identify the eighth value corresponding to the third feature of the second dataset; the third feature belongs to the third feature set; The first information includes multiple third feature identifiers, which are used to identify multiple ninth values ​​corresponding to the third feature; or, the first information includes one or more third feature identifiers, each third feature identifier being used to identify a set of third values ​​corresponding to the third feature; wherein, the values ​​corresponding to the third feature in at least two datasets are included in the multiple ninth values ​​or at least one set of third values, and the at least two datasets can be used for training the same model; or... The first information is used to indicate that datasets with different values ​​corresponding to the third feature can be used to train the same model; or, The first information is used to indicate that datasets with different values ​​corresponding to the third feature cannot be used for training the same model; or, The first information is used to indicate the conditions that the values ​​corresponding to the third feature in at least two datasets that can be used to train the same model must satisfy.

44. The method according to claim 43, characterized in that, The third feature set includes one or more of the following: whether the data was acquired during training with a quantizer, the absolute transmit power distribution range corresponding to the data, the receive power distribution range corresponding to the data, the receive noise power distribution range corresponding to the data, the time delay spread distribution range corresponding to the data, the angle spread distribution range corresponding to the data, the interference level distribution range corresponding to the data, and the SINR distribution range corresponding to the data.

45. The method according to any one of claims 33 to 44, characterized in that, The eleventh identifier includes the seventh identifier, which is used to identify the tenth value corresponding to the fourth feature of the first dataset. The twelfth identifier includes the eighth identifier, which is used to identify the eleventh value corresponding to the fourth feature of the second dataset. The fourth feature belongs to the fourth feature set. The first information includes multiple fourth feature identifiers, which are used to identify multiple twelfth values ​​corresponding to the fourth feature; or, the first information includes one or more fourth feature identifiers, each fourth feature identifier being used to identify a set of fourth values ​​corresponding to the fourth feature; wherein, the values ​​corresponding to the fourth feature in at least two datasets are included in the multiple twelfth values ​​or at least one set of fourth values, and the at least two datasets can be used for training the same model; or... The first information is used to indicate that datasets with different values ​​corresponding to the fourth feature can be used to train the same model; or, The first information is used to indicate that datasets with different values ​​corresponding to the fourth feature cannot be used for training the same model; or, The first information is used to indicate the conditions that the values ​​corresponding to the fourth feature in at least two datasets that can be used to train the same model must satisfy.

46. ​​The method according to claim 45, characterized in that, The fourth feature set includes one or more of the following: the number of transmission ports corresponding to the data in the dataset, the carrier frequency corresponding to the data in the dataset, the transmission bandwidth corresponding to the data in the dataset, the number of transmission subbands corresponding to the data in the dataset, the granularity of the transmission subbands corresponding to the data in the dataset, and the number of bits of CSI feedback overhead corresponding to the data in the dataset.

47. The method according to any one of claims 33 to 46, characterized in that, The eleventh identifier includes the ninth identifier, which is used to identify the thirteenth value corresponding to the fifth feature of the first dataset; the twelfth identifier includes the tenth identifier, which is used to identify the fourteenth value corresponding to the fifth feature of the second dataset; the fifth feature belongs to the fifth feature set. The first information includes multiple fifth feature identifiers, which are used to identify multiple fifteenth values ​​corresponding to the fifth feature; or, the first information includes one or more fifth feature identifiers, each fifth feature identifier being used to identify a set of fifth values ​​corresponding to the fifth feature; wherein, the values ​​corresponding to the fifth feature in at least two datasets are included in the multiple fifteenth values ​​or at least one set of fifth values, and the at least two datasets can be used for training the same model; or... The first information is used to indicate that datasets with different values ​​corresponding to the fifth feature can be used to train the same model; or, The first information is used to indicate that datasets with different values ​​corresponding to the fifth feature cannot be used for training the same model; or, The first information is used to indicate the conditions that the values ​​corresponding to the fifth feature in at least two datasets that can be used to train the same model must satisfy.

48. The method according to claim 47, characterized in that, The fifth feature set includes one or more of the following: the length of the historical time window corresponding to the data in the dataset, the length of the prediction time window corresponding to the data in the dataset, and the moving speed of the device corresponding to the data in the dataset.

49. The method according to any one of claims 33 to 48, characterized in that, The first information indicates the third rule and the fourth rule, wherein the third rule has a higher priority than the fourth rule. The third rule and the fourth rule belong to the following rules used to determine whether the datasets corresponding to at least two models can be used or not used for training the same model: rules related to the model's identifier, rules related to the dataset's identifier, rules related to the values ​​corresponding to the features included in the third feature set of the dataset, rules related to the values ​​corresponding to the features included in the fourth feature set of the dataset, and rules related to the values ​​corresponding to the features included in the fifth feature set of the dataset. The step of determining the second information based on the first information associated with the eleventh identifier and the twelfth identifier includes: Based on the eleventh identifier, the twelfth identifier, and the third rule, it is determined that the first dataset and the second dataset can be used for training the same model, or that the first dataset and the second dataset cannot be used for training the same model.

50. The method according to claim 49, characterized in that, The method further includes determining, based on the eleventh identifier, the twelfth identifier, and the third rule, that the first dataset and the second dataset can be used for training the same model. According to the fourth rule, it is determined that the first dataset and the second dataset can be used to train the same model, or that the first dataset and the second dataset cannot be used to train the same model.

51. A communication device, characterized in that, It includes functional modules for implementing the method as described in any one of claims 1 to 18, or includes functional modules for implementing the method as described in any one of claims 19 to 32, or includes functional modules for implementing the method as described in any one of claims 33 to 50.

52. A communication device, characterized in that, The device includes one or more processors and communication circuitry, the communication circuitry being used for at least one of inputting or outputting signals; the one or more processors are used to implement the method as described in any one of claims 1 to 18, or the one or more processors are used to implement the method as described in any one of claims 19 to 32, or the one or more processors are used to implement the method as described in any one of claims 33 to 50.

53. The communication device according to claim 52, characterized in that, The communication device also includes a memory.

54. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program or instructions that, when executed by a processor, cause the method as claimed in any one of claims 1 to 18 to be performed, or cause the method as claimed in any one of claims 19 to 32 to be performed, or cause the method as claimed in any one of claims 33 to 50 to be performed.

55. A computer program product, characterized in that, Includes a computer program that, when run, causes the method as described in any one of claims 1 to 18 to be performed, or causes the method as described in any one of claims 19 to 32 to be performed, or causes the method as described in any one of claims 33 to 50 to be performed.

56. A chip, characterized in that, The device includes a processor and a communication interface, wherein the processor reads instructions from a memory via the communication interface to execute the method as described in any one of claims 1 to 18, or to execute the method as described in any one of claims 19 to 32, or to execute the method as described in any one of claims 33 to 50.

Images