Model performance monitoring method, device, and storage medium

Abstract

The present disclosure relates to a model performance monitoring method, a device, and a storage medium. The method comprises: acquiring first information, the first information comprising second information and / or third information, the second information being used for indicating the feedback performance of a codebook, the third information comprising fourth information and fifth information, the fourth information being channel state information (CSI) at a first time point, the fifth information being CSI at the first time point recovered by a second partial model of a first model, the first model comprising a first partial model and the second partial model, the first partial model being used for compressing CSI, and the second partial model being used for performing CSI recovery on the basis of the CSI compressed by the first partial model; and on the basis of the first information, determining the performance of the first model. In other words, by means of acquiring the CSI at the first time point and the CSI recovered by the second partial model, the performance of the first model can be monitored, thereby reducing signaling overhead during model performance monitoring.

Description

Model performance monitoring method, device and storage medium TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of communication, and particularly relates to a model performance monitoring method, device and storage medium. BACKGROUND

[0002] With the progress of communication technology, an artificial intelligence (AI) model is introduced in a communication system. Through the AI model, feedback overhead of a terminal can be reduced or feedback accuracy of channel status information (CSI) can be improved. However, the effect of the AI model depends on the performance of the AI model. Therefore, how to monitor the performance of the AI model becomes a problem to be solved urgently.

[0003] SUMMARY

[0004] Embodiments of the present disclosure provide a model performance monitoring method, device and storage medium.

[0005] According to a first aspect of embodiments of the present disclosure, a model performance monitoring method is provided, executed by a terminal device, and the method comprises:

[0006] obtaining first information, the first information comprising second information and / or third information; the second information being used for indicating feedback performance of a codebook, the third information comprising fourth information and fifth information, the fourth information being channel status information (CSI) at a first time, and the fifth information being the CSI at the first time recovered by a second part model of a first model; the first model comprising a first part model and the second part model, the first part model being used for compressing processing of the CSI, and the second part model being used for CSI recovery according to the CSI compressed processed by the first part model;

[0007] determining performance of the first model according to the first information.

[0008] According to a second aspect of embodiments of the present disclosure, a model performance monitoring method is provided, executed by a network device, and the method comprises:

[0009] obtain first information, the first information comprising second information and / or third information; the second information being used for indicating feedback performance of a codebook, the third information comprising fourth information and fifth information, the fourth information being channel state information (CSI) at a first time, the fifth information being the CSI at the first time recovered by a second part model of a first model; the first model comprising a first part model and the second part model, the first part model being used for compression processing of the CSI, the second part model being used for CSI recovery according to the CSI compressed by the first part model;

[0010] determine performance of the first model according to the first information.

[0011] According to a third aspect of embodiments of the present disclosure, a terminal device is provided, comprising:

[0012] obtain first information, the first information comprising second information and / or third information; the second information being used for indicating feedback performance of a codebook, the third information comprising fourth information and fifth information, the fourth information being channel state information (CSI) at a first time, the fifth information being the CSI at the first time recovered by a second part model of a first model; the first model comprising a first part model and the second part model, the first part model being used for compression processing of the CSI, the second part model being used for CSI recovery according to the CSI compressed by the first part model;

[0013] determine performance of the first model according to the first information.

[0014] According to a fourth aspect of embodiments of the present disclosure, a network device is provided, comprising:

[0015] obtain first information, the first information comprising second information and / or third information; the second information being used for indicating feedback performance of a codebook, the third information comprising fourth information and fifth information, the fourth information being channel state information (CSI) at a first time, the fifth information being the CSI at the first time recovered by a second part model of a first model; the first model comprising a first part model and the second part model, the first part model being used for compression processing of the CSI, the second part model being used for CSI recovery according to the CSI compressed by the first part model;

[0016] determine performance of the first model according to the first information.

[0017] According to a fifth aspect of embodiments of the present disclosure, a communication device is provided, comprising:

[0018] One or more processors; wherein the communication device can be configured to perform the optional implementation of the first aspect or the second aspect.

[0019] According to a sixth aspect of the embodiments of the present disclosure, a communication system is provided, including a terminal device and a network device, wherein the terminal device is configured to perform the method described in the optional implementation of the first aspect, and the network device is configured to perform the method described in the optional implementation of the second aspect.

[0020] According to a seventh aspect of the embodiments of the present disclosure, a storage medium is provided, which stores instructions, when the instructions run on a communication device, cause the communication device to perform the method described in the optional implementation of the first aspect or the second aspect.

[0021] The technical solutions provided by the embodiments of the present disclosure can have the following beneficial effects: obtaining first information, the first information including second information and / or third information; the second information is used to indicate the feedback performance of a codebook, the third information including fourth information and fifth information, the fourth information being channel state information CSI at a first time, and the fifth information being the CSI at the first time recovered by a second part model of a first model; the first model including a first part model and the second part model, the first part model being used to compress the CSI, and the second part model being used to recover the CSI according to the CSI compressed by the first part model; and determining the performance of the first model according to the first information. That is, by obtaining the CSI at the first time and the CSI recovered by the second part model, the performance of the first model can be monitored, and the signaling overhead during model performance monitoring is reduced.

[0022] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following describes the drawings required for the embodiment description. The following drawings are only some embodiments of the present disclosure, and do not specifically limit the protection scope of the present disclosure.

[0024] FIG. 1A is a schematic diagram of an architecture of a communication system according to an embodiment of the present disclosure.

[0025] FIG. 1B is a schematic diagram of CSI compression feedback and recovery based on a bilateral AI / ML model according to an embodiment of the present disclosure.

[0026] FIG. 1C is a schematic diagram of CSI compression feedback based on a bilateral CSI prediction compression model according to an embodiment of the present disclosure.

[0027] FIG. 2A is an interaction diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0028] FIG. 2B is a schematic diagram of CSI transmission according to an embodiment of the present disclosure.

[0029] FIG. 2C is an interaction diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0030] FIG. 3A is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0031] FIG. 3B is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0032] FIG. 3C is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0033] FIG. 3D is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0034] FIG. 3E is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0035] FIG. 3F is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0036] FIG. 3G is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0037] FIG. 4A is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0038] FIG. 4B is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0039] FIG. 4C is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0040] FIG. 4D is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0041] FIG. 5A is an interaction diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0042] FIG. 5B is an interaction diagram of a model performance monitoring method according to an embodiment of the present disclosure.

[0043] FIG. 6A is a structural diagram of a terminal device according to an embodiment of the present disclosure.

[0044] FIG. 6B is a structural schematic diagram of a network device according to an embodiment of the present disclosure.

[0045] FIG. 7A is a structural schematic diagram of a communication device according to an embodiment of the present disclosure.

[0046] FIG. 7B is a structural schematic diagram of a chip according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0047] The present disclosure provides a model performance monitoring method, device and storage medium.

[0048] In a first aspect, a model performance monitoring method is provided. The method is performed by a terminal device and includes:

[0049] obtaining first information, the first information including second information and / or third information; the second information being used to indicate feedback performance of a codebook, the third information including fourth information and fifth information, the fourth information being channel state information (CSI) at a first time, and the fifth information being CSI at the first time recovered by a second part model of a first model; the first model including a first part model and the second part model, the first part model being used to compress CSI, and the second part model being used to recover CSI according to the compressed CSI of the first part model;

[0050] determining performance of the first model according to the first information.

[0051] In the above embodiment, by obtaining CSI at the first time and the CSI recovered by the second part model, performance of the first model can be monitored, and signaling overhead during model performance monitoring is reduced.

[0052] In some embodiments of the first aspect, the first part model is deployed on the terminal device, and the second part model is deployed on a network device.

[0053] In the above embodiment, the first model can be a bilateral model in which the first part model is deployed on the terminal device and the second part model is deployed on the network device.

[0054] In some embodiments of the first aspect, obtaining the fifth information includes:

[0055] receiving the fifth information sent by the network device.

[0056] In the above embodiment, the terminal device can receive the CSI recovered by the second part model deployed on the network device side.

[0057] In some embodiments of the first aspect, in some embodiments, the receiving the fifth information sent by the network device comprises at least one of the following:

[0058] receiving all fifth information of the first time sent by the network device at the same time;

[0059] receiving the fifth information of each time in the first time sent by the network device in a first order, the first order being a protocol agreement or being indicated by the network device.

[0060] In the above embodiments, the network device can send all fifth information at a time, or send fifth information of each time at a time.

[0061] In some embodiments of the first aspect, in some embodiments, the sending mode of the fifth information comprises at least one of the following: floating point quantization mode, codebook mode.

[0062] In the above embodiments, the fifth information can be sent in at least one of the floating point number mode or the codebook mode, so that the sending mode of the fifth information is more flexible.

[0063] In some embodiments of the first aspect, in some embodiments, the method further comprises:

[0064] obtaining sixth information, the sixth information being CSI of a second time, the second time being before the first time;

[0065] inputting the sixth information into the first part model to obtain seventh information;

[0066] sending the seventh information to the network device, the seventh information being used by the network device to obtain the fifth information.

[0067] In the above embodiments, the terminal device can perform compression processing on the CSI of the historical time, and send the CSI after compression processing to the network device, so that the network device can obtain the CSI of the first time.

[0068] In some embodiments of the first aspect, in some embodiments, the obtaining the sixth information comprises:

[0069] receiving a first resource sent by the network device at the second time;

[0070] obtaining the sixth information according to the first resource.

[0071] In the above embodiments, the network device can send a first resource to the terminal device, so that the terminal device measures the CSI of the historical time.

[0072] In some embodiments of the first aspect, in some embodiments, the obtaining the fourth information comprises at least one of the following:

[0073] The fourth information is measured;

[0074] The fourth information is predicted by a second model, the second model being used for predicting CSI;

[0075] The eighth information is measured, and the fourth information is predicted by the second model according to the eighth information.

[0076] In the above embodiments, the terminal device can measure the fourth information, can predict the fourth information by a model, or can obtain the fourth information by combining measurement and prediction.

[0077] In some embodiments of the first aspect, in some embodiments, the measuring the fourth information comprises:

[0078] A second resource sent by the network device at the first time point is received;

[0079] The fourth information is measured according to the second resource.

[0080] In the above embodiments, the terminal device can measure the fourth information according to the second resource sent by the network device.

[0081] In some embodiments of the first aspect, in some embodiments, the predicting the fourth information by the second model comprises:

[0082] Ninth information is obtained, the ninth information being CSI at a historical time point;

[0083] Tenth information is input into the second model to obtain eleventh information, the tenth information comprising at least one CSI at a historical time point and / or at least one CSI at the first time point predicted by the second model;

[0084] The fourth information is determined from the eleventh information according to first indication information, the first indication information being used for indicating the first time point.

[0085] In the above embodiments, the terminal device can predict the CSI at the first time point according to the CSI at the historical time point, and can predict the fourth information according to the CSI at the historical time point and part or all of the CSI at the first time point.

[0086] In some embodiments of the first aspect, in some embodiments, the measuring the eighth information and predicting the fourth information by the second model according to the eighth information comprises:

[0087] receive a third resource sent by the network device at a third time, the third time including the first N time points in the first time and / or historical time points before the first time;

[0088] measure the eighth information according to the third resource;

[0089] input the eighth information into the second model to obtain twelfth information;

[0090] determine the fourth information from the twelfth information according to first indication information, the first indication information being used to indicate the first time.

[0091] In the above embodiment, the terminal device can predict the CSI of the entire first time according to the CSI of part of the first time and the CSI of the historical time.

[0092] In combination with some embodiments of the first aspect, in some embodiments, the method further includes:

[0093] sending the fourth information to the network device.

[0094] In the above embodiment, the terminal device can report the fourth information to the network device, so that the network device monitors the performance of the first model according to the fourth information.

[0095] In combination with some embodiments of the first aspect, in some embodiments, the sending of the fourth information to the network device includes at least one of:

[0096] sending all the fourth information of the first time to the network device at the same time;

[0097] sending the fourth information of each time point in the first time to the network device in sequence according to a second order, the second order being a protocol agreement or an indication of the terminal device.

[0098] In the above embodiment, the terminal device can report all the fourth information at one time, or report the fourth information of each time point in sequence according to the second order.

[0099] In combination with some embodiments of the first aspect, in some embodiments, the sending of the fourth information to the network device includes at least one of:

[0100] sending the fourth information to the network device by a floating-point quantization method;

[0101] sending the fourth information to the network device by a codebook method;

[0102] The fourth information is compressed by a third partial model in a third model to obtain thirteenth information, and the thirteenth information is sent to the network device, the third model includes the third partial model and a fourth partial model, the third partial model is used for compressing the CSI, and the fourth partial model is used for recovering the compressed CSI.

[0103] In the above embodiment, the terminal device can report the fourth information in a floating-point quantization manner, can report the fourth information in a codebook manner, and can report the fourth information in a double-sided model manner, and the reporting manner is more flexible.

[0104] With reference to some embodiments of the first aspect, in some embodiments, the method further includes:

[0105] The second indication information sent by the network device is received, and the second indication information is used to indicate that the terminal device sends the fourth information.

[0106] In the above embodiment, the terminal device can report the fourth information according to the indication of the network device.

[0107] With reference to some embodiments of the first aspect, in some embodiments, a fourth model is further deployed on the terminal device, and the fourth model is used for recovering the compressed CSI.

[0108] In the above embodiment, the fourth model for recovering the CSI can also be deployed on the terminal device side, and the CSI can be recovered through the fourth model, so that the network device does not need to send the recovered CSI, signaling overhead is reduced, and system performance is improved.

[0109] With reference to some embodiments of the first aspect, in some embodiments, obtaining the fifth information includes:

[0110] The fifth information is obtained through the fourth model.

[0111] With reference to some embodiments of the first aspect, in some embodiments, the performance of the first model is determined according to the first information includes:

[0112] Fourteenth information is determined according to the first information, the fourteenth information is used to indicate the performance of the first model, and the fourteenth information includes at least one of the following: square of cosine similarity SGCS, normalized mean square error NMSE, block error rate BLER, and throughput.

[0113] In the above embodiment, the performance of the first model can be determined by various parameters, so that the indication manner of the performance is more flexible.

[0114] In some embodiments of the first aspect, in some embodiments, the method further comprises:

[0115] sending the performance of the first model to the network device.

[0116] In the above embodiments, the terminal device can report the performance of the monitored first model to the network device.

[0117] In some embodiments of the first aspect, in some embodiments, the first time is determined by:

[0118] agreement of a protocol;

[0119] indication of the network device;

[0120] indication of the terminal device.

[0121] In the above embodiments, the first time can be indicated in various ways, which is more flexible.

[0122] In a second aspect, the embodiments of the present disclosure provide a model performance monitoring method, executed by a network device, comprising:

[0123] obtaining first information, the first information comprising second information and / or third information; the second information being used to indicate feedback performance of a codebook, the third information comprising fourth information and fifth information, the fourth information being channel state information (CSI) at a first time, and the fifth information being CSI at the first time recovered by a second part model of a first model; the first model comprising a first part model and the second part model, the first part model being used to compress CSI, and the second part model being used to recover CSI according to the compressed CSI of the first part model;

[0124] determining performance of the first model according to the first information.

[0125] In some embodiments of the second aspect, in some embodiments, the first part model is deployed on the terminal device, and the second part model is deployed on the network device.

[0126] In some embodiments of the second aspect, in some embodiments, the method further comprises:

[0127] receiving seventh information sent by the terminal device, the seventh information being determined by the terminal device according to sixth information by using the first part model, the sixth information being CSI at a second time, and the second time being before the first time;

[0128] input the seventh information into the second part model to obtain fifteenth information, the fifteenth information comprising recovered CSI of the fourth time, the fourth time comprising at least the first time.

[0129] In the above embodiments, the network device can recover the CSI of the fourth time and send all or part of the recovered CSI to the terminal device.

[0130] In combination with some embodiments of the second aspect, in some embodiments, the method further comprises:

[0131] sending a first resource to the terminal device, the first resource being used by the terminal device to measure to obtain the sixth information.

[0132] In combination with some embodiments of the second aspect, in some embodiments, the method further comprises:

[0133] sending the fifth information to the terminal device according to the fifteenth information.

[0134] In combination with some embodiments of the second aspect, in some embodiments, the sending of the fifth information to the terminal device comprises at least one of:

[0135] sending all of the fifth information of the first time to the terminal device at the same time;

[0136] sending the fifth information of each time in the first time to the terminal device in sequence according to a first order, the first order being a protocol agreement or being indicated by the network device.

[0137] In combination with some embodiments of the second aspect, in some embodiments, the sending mode of the fifth information comprises at least one of: a floating-point quantization mode, a codebook mode.

[0138] In combination with some embodiments of the second aspect, in some embodiments, the obtaining of the fourth information comprises:

[0139] receiving the fourth information sent by the terminal device.

[0140] In combination with some embodiments of the second aspect, in some embodiments, the receiving of the fourth information sent by the terminal device comprises at least one of:

[0141] receiving all of the fourth information of the first time sent by the terminal device at the same time;

[0142] receiving the fourth information of each time in the first time sent by the terminal device in sequence according to a second order, the second order being a protocol agreement or being indicated by the terminal device.

[0143] In some embodiments of the second aspect, in some embodiments, the receiving the fourth information sent by the terminal device comprises at least one of:

[0144] receiving the fourth information sent by the terminal device in a floating-point quantization manner;

[0145] receiving the fourth information sent by the terminal device in a codebook manner;

[0146] receiving thirteenth information sent by the terminal device, the thirteenth information being obtained by compressing the fourth information by a third partial model in a third model, the third model comprising the third partial model and a fourth partial model, the third partial model being configured to compress the CSI, and the fourth partial model being configured to recover the compressed CSI.

[0147] In some embodiments of the second aspect, in some embodiments, the method further comprises:

[0148] recovering the fourth information by recovering the thirteenth information by a fourth partial model in the third model.

[0149] In some embodiments of the second aspect, in some embodiments, the method further comprises:

[0150] sending a second resource to the terminal device at the first time, the second resource being configured to be used by the terminal device to measure the fourth information.

[0151] In some embodiments of the second aspect, in some embodiments, the method further comprises:

[0152] sending a third resource to the terminal device, the third resource being configured to be used by the terminal device to measure eighth information, predict twelfth information according to the eighth information, and determine the fourth information from the twelfth information according to first indication information, the first indication information being configured to indicate the first time.

[0153] In some embodiments of the second aspect, in some embodiments, the first indication information is protocol agreement or indication of the terminal device.

[0154] In some embodiments of the second aspect, in some embodiments, the method further comprises:

[0155] sending second indication information to the terminal device, the second indication information being configured to indicate the terminal device to send the fourth information.

[0156] In some embodiments combined with the second aspect, in some embodiments, the determining the performance of the first model according to the first information comprises:

[0157] determining fourteenth information according to the first information, the fourteenth information being used to indicate the performance of the first model; wherein the fourteenth information comprises at least one of square of cosine similarity SGCS, normalized mean square error NMSE, block error rate BLER, and throughput.

[0158] In some embodiments combined with the second aspect, in some embodiments, the method further comprises:

[0159] receiving the performance of the first model sent by the terminal device.

[0160] In some embodiments combined with the second aspect, in some embodiments, the first time point is determined by:

[0161] agreement of protocol;

[0162] indication of the network device;

[0163] indication of the terminal device.

[0164] In a third aspect, the embodiments of the present disclosure provide a terminal device, which can comprise at least one of a transceiver module and a processing module; wherein the terminal device can be configured to perform the optional implementation manners of the first aspect.

[0165] In a fourth aspect, the embodiments of the present disclosure provide a network device, which can comprise at least one of a transceiver module and a processing module; wherein the network device can be configured to perform the optional implementation manners of the second aspect.

[0166] In a fifth aspect, the embodiments of the present disclosure provide a terminal device, which can comprise one or more processors; wherein the terminal device can be configured to perform the optional implementation manners of the first aspect.

[0167] In a sixth aspect, the embodiments of the present disclosure provide a network device, which can comprise one or more processors; wherein the network device can be configured to perform the optional implementation manners of the second aspect.

[0168] In a seventh aspect, the embodiments of the present disclosure provide a communication system, which can comprise a terminal device and a network device; wherein the terminal device is configured to perform the method described in the optional implementation manners of the first aspect, and the network device is configured to perform the method described in the optional implementation manners of the second aspect.

[0169] In an eighth aspect, the embodiments of the present disclosure provide a storage medium, which stores instructions. When the instructions are executed on a communication device, the communication device performs the method described in the optional implementation of the first aspect or the second aspect.

[0170] In a ninth aspect, the embodiments of the present disclosure provide a program product, which, when executed by a communication device, causes the communication device to perform the method described in the optional implementation of the first aspect or the second aspect.

[0171] In a tenth aspect, the embodiments of the present disclosure provide a computer program, which, when executed on a computer, causes the computer to perform the method described in the optional implementation of the first aspect or the second aspect.

[0172] In an eleventh aspect, the embodiments of the present disclosure provide a chip or chip system. The chip or chip system includes processing circuitry configured to perform the method described in the optional implementation of the first aspect or the second aspect.

[0173] It can be understood that the terminal device, the network device, the communication device, the communication system, the storage medium, the program product, the computer program, the chip or the chip system can be used to execute the method proposed in the embodiments of the present disclosure. Therefore, the beneficial effects that can be achieved thereby can refer to the beneficial effects in the corresponding method, which will not be described here again.

[0174] The embodiments of the present disclosure propose a model performance monitoring method, device and storage medium. In some embodiments, the terms of the model performance monitoring method and the information processing method, the communication method can be replaced with each other; the terms of the model performance monitoring device and the information processing device, the communication device, the communication equipment can be replaced with each other; the terms of the model performance monitoring system and the communication system can be replaced with each other.

[0175] The embodiments of the present disclosure are not exhaustive, but only illustrate some embodiments, and are not specific limitations on the protection scope of the present disclosure. In the case of no contradiction, each step in an embodiment can be implemented as an independent embodiment, and the steps can be combined arbitrarily, for example, the scheme after removing part of the steps in an embodiment can also be implemented as an independent embodiment, and the order of the steps in an embodiment can be exchanged arbitrarily, in addition, the optional implementation in an embodiment can be combined arbitrarily; in addition, the embodiments can be combined arbitrarily, for example, part or all steps of different embodiments can be combined arbitrarily, an embodiment can be combined with the optional implementation of other embodiments.

[0176] In the embodiments of the present disclosure, the terms and / or descriptions among the embodiments are consistent and can be referred to each other if there is no special description and logical conflict, and the technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0177] The terms used in the embodiments of the present disclosure are only for the purpose of describing particular embodiments and are not used as limitations of the present disclosure.

[0178] In the embodiments of the present disclosure, unless otherwise specified and logically conflicted, the elements expressed in singular form, such as "one", "one kind", "the", "the above", "the above", "the above", "this" and the like, can represent "one and only one", and can also represent "one or more", "at least one" and the like. For example, in the case of using articles such as "a", "an", "the" and the like in English, the noun after the article can be understood as singular expression, and can also be understood as plural expression.

[0179] In some embodiments, "plurality" can refer to two or more.

[0180] In some embodiments, the terms "at least one of", "one or more of", "a plurality of", "multiple" and the like can be replaced with each other.

[0181] In some embodiments, the writing manner of "at least one of A, B", "A and / or B", "A in one case and B in another case", "A in response to one case and B in response to another case" and the like can include the following technical solutions according to the case: A in some embodiments (A is executed regardless of B); B in some embodiments (B is executed regardless of A); A and B are selectively executed in some embodiments (A and B are selected from A and B); A and B are executed in some embodiments (A and B are executed). When there are more branches of A, B, C and the like, it is similar to the above.

[0182] In some embodiments, the writing manner of "A or B" and the like can include the following technical solutions according to the case: A in some embodiments (A is executed regardless of B); B in some embodiments (B is executed regardless of A); A and B are selectively executed in some embodiments (A and B are selected from A and B). When there are more branches of A, B, C and the like, it is similar to the above.

[0183] The prefix words of "first", "second" and the like in the embodiments of the present disclosure are merely used to distinguish different description objects, and do not constitute limitation on the position, order, priority, quantity or content of the description objects. The description objects are described in the claims or embodiments in the context of the description, and should not be construed as redundant limitation because of the use of the prefix words. For example, the description object is "field", and the ordinal words before "field" in "first field" and "second field" do not limit the position or order between "fields", and "first" and "second" do not limit whether the "fields" modified by them are in the same message or not, nor limit the order of "first field" and "second field". For another example, the description object is "level", and the ordinal words before "level" in "first level" and "second level" do not limit the priority between "levels". For another example, the quantity of the description object is not limited by the ordinal words, and can be one or more. For example, "first device", wherein the quantity of "device" can be one or more. In addition, the objects modified by different prefix words can be the same or different, for example, the description object is "device", and "first device" and "second device" can be the same device or different devices, and their types can be the same or different; for another example, the description object is "information", and "first information" and "second information" can be the same information or different information, and their contents can be the same or different.

[0184] In some embodiments, "including A", "containing A", "for indicating A", "carrying A" can be interpreted as directly carrying A, or indirectly indicating A.

[0185] In some embodiments, the terms of "in response to", "in response to determining", "in the case of", "when", "when", "if", and the like can be replaced with each other.

[0186] In some embodiments, the terms of "greater than", "greater than or equal to", "not less than", "more than", "more than or equal to", "not less than", "higher than", "higher than or equal to", "not lower than", "above", and the like can be replaced with each other, and the terms of "less than", "less than or equal to", "not greater than", "less than", "less than or equal to", "not more than", "lower than", "lower than or equal to", "not higher than", "below", and the like can be replaced with each other.

[0187] In some embodiments, an apparatus or the like can be interpreted as an entity, and can also be interpreted as virtual, and the name thereof is not limited to the name described in the embodiments. The terms "apparatus", "equipment", "device", "circuit", "network element", "node", "function", "unit", "section", "system", "network", "chip", "chip system", "entity", "subject" and the like can be replaced with each other.

[0188] In some embodiments, a "network" can be interpreted as an apparatus (for example, an access network device, a core network device, and the like) included in the network.

[0189] In some embodiments, the terms "Access Network Device (AN Device)", "Radio Access Network Device (RAN Device)", "Base Station (BS)", "Radio Base Station", "Fixed Station", "Node", "Access Point", "Transmission Point (TP)", "Reception Point (RP)", "Transmission / Reception Point (TRP)", "Panel", "Antenna Panel", "Antenna Array", "Cell", "Macro Cell", "Small Cell", "Femto Cell", "Pico Cell", "Sector", "Cell Group", "Serving Cell", "Carrier", "Component Carrier", "Bandwidth Part (BWP)" and the like can be replaced with each other.

[0190] In some embodiments, the terms "terminal," "terminal device," "user equipment" (UE), "user terminal," "mobile station" (MS), "mobile terminal" (MT), subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, and the like can be used interchangeably.

[0191] In some embodiments, an access network device, a core network device, or a network device can be replaced with a terminal. For example, the embodiments of the present disclosure can also be applied to a structure in which communication between an access network device, a core network device, or a network device and a terminal is replaced with communication between a plurality of terminals (e.g., device-to-device (D2D), vehicle-to-everything (V2X), or the like). In this case, the terminal can also be configured to have all or part of the functions of the access network device. In addition, the terms "uplink," "downlink," and the like can also be replaced with terms corresponding to inter-terminal communication (e.g., "side"). For example, an uplink channel, a downlink channel, and the like can be replaced with a side channel or a direct connection channel, and an uplink, a downlink, and the like can be replaced with a side link or a direct connection link.

[0192] In some embodiments, a terminal can be replaced with an access network device, a core network device, or a network device. In this case, the access network device, the core network device, or the network device can also be configured to have all or part of the functions of the terminal.

[0193] In some embodiments, the data, information, etc. can be obtained in compliance with the laws and regulations of the country where the location is situated.

[0194] In some embodiments, the data, information, etc. can be obtained after obtaining the consent of the user.

[0195] In addition, each element, each row, or each column in the table of the embodiments of the present disclosure can be implemented as an independent embodiment, and any combination of any element, any row, or any column can also be implemented as an independent embodiment.

[0196] FIG. 1A is a schematic diagram of an architecture of a communication system according to an embodiment of the present disclosure. As shown in FIG. 1A, the communication system 100 can include a terminal device 101 and a network device 102.

[0197] In some embodiments, the terminal device 101 can include at least one of a mobile phone, a wearable device, an Internet of Things (IoT) device, a communication-capable automobile, a smart automobile, a Pad, a computer with wireless transceiver function, a Virtual Reality (VR) terminal device, an Augmented Reality (AR) terminal device, a wireless terminal device in industrial control, a wireless terminal device in self-driving, a wireless terminal device in remote medical surgery, a wireless terminal device in smart grid, a wireless terminal device in transportation safety, a wireless terminal device in smart city, a wireless terminal device in smart home, and the like, but is not limited thereto.

[0198] In some embodiments, the network device 102 can include at least one of an access network device and a core network device.

[0199] In some embodiments, the access network device can be at least one of a node or a device that accesses a terminal device to a wireless network, and the access network device can include at least one of an evolved node B (eNB) in a 5G communication system, a next generation eNB (ng-eNB), a next generation node B (gNB), a node B (NB), a home node B (HNB), a home evolved node B (HeNB), a wireless backhaul device, a radio network controller (RNC), a base station controller (BSC), a base transceiver station (BTS), a base band unit (BBU), a mobile switching center, a base station in a 6G communication system, an open base station (Open RAN), a cloud base station (Cloud RAN), a base station in other communication systems, an access node in a Wi-Fi system, but is not limited thereto.

[0200] In some embodiments, the technical solutions of the present disclosure can be applied to an Open RAN architecture, at which time the interfaces between or within the access network devices involved in the embodiments of the present disclosure can become internal interfaces of the Open RAN, and the processes and information interactions between these internal interfaces can be realized through software or programs.

[0201] In some embodiments, the access network device can be composed of a central unit (CU) and a distributed unit (DU), wherein the CU can also be referred to as a control unit (Control Unit). The CU-DU structure can split the protocol layers of the access network device, and some of the protocol layers are controlled by the CU, and the rest or all of the protocol layers are distributed in the DU and controlled by the CU, but is not limited thereto.

[0202] In some embodiments, the core network device can be one device, or a plurality of devices or device groups. The core network can include at least one of an evolved packet core (EPC), a 5G core network (5GCN), and a next generation core (NGC).

[0203] It can be understood that the communication system described in the embodiments of the present disclosure is for more clearly illustrating the technical solutions of the embodiments of the present disclosure, and does not constitute a limitation on the technical solutions proposed by the embodiments of the present disclosure. Those skilled in the art can know that, with the evolution of system architecture and the appearance of new business scenarios, the technical solutions proposed by the embodiments of the present disclosure are also applicable to similar technical problems.

[0204] The following embodiments of the present disclosure can be applied to the communication system 100 shown in FIG. 1A or part of the subject, but are not limited thereto. The subjects shown in FIG. 1A are examples, and the communication system can include all or part of the subjects in FIG. 1A, or other subjects other than FIG. 1A. The number and form of each subject is arbitrary, each subject can be real or virtual, the connection relationship between each subject is an example, each subject can not be connected or can be connected, the connection can be in any way, can be direct connection or indirect connection, can be wired connection or wireless connection.

[0205] Embodiments of the present disclosure can be applied to Long Term Evolution (LTE), LTE-Advanced (LTE-A), LTE-Beyond (LTE-B), SUPER 3G, IMT-Advanced, 4th generation mobile communication system (4G), 5th generation mobile communication system (5G), 5G new radio (NR), Future Radio Access (FRA), New-Radio Access Technology (RAT), New Radio (NR), New Radio Access (NX), Future generation radio access (FX), Global System for Mobile communications (GSM (registered trademark)), CDMA2000, Ultra Mobile Broadband (UMB), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), IEEE 802.20, Ultra-WideBand (UWB), Bluetooth (Bluetooth (registered trademark)), Public Land Mobile Network (PLMN) network, Device-to-Device (D2D) system, Machine to Machine (M2M) system, Internet of Things (IoT) system, Vehicle-to-Everything (V2X), system using other communication methods, next-generation system expanded based thereon, and the like. Further, a plurality of systems can be applied in combination (for example, combination of LTE or LTE-A and 5G, and the like).

[0206] In some embodiments of the present disclosure, the reduction of feedback overhead or the improvement of CSI feedback accuracy of the terminal can be achieved by adopting AI technology, and the bilateral AI / Machine Learning (ML) model based on the terminal-side CSI generation part model and the network-side CSI recovery part model has been developed in the 3GPP standardization research to respectively realize the compressed feedback and recovery of CSI. FIG. 1B is a schematic diagram of realizing CSI compressed feedback and recovery based on a bilateral AI / ML model according to an embodiment of the present disclosure. As shown in FIG. 1B, the UE side compresses the downlink channel information H through a CSI generation part model (defined as an encoder) and sends it to the gNB through quantization as a binary bit stream s, and the gNB side recovers H' similar to the original downlink information through a CSI recovery part model (defined as a decoder).

[0207] In some embodiments, the channel time correlation can be utilized to improve system performance, and the specific implementation can include the following two methods:

[0208] Method 1: The historical CSI before the current time t is utilized when compressing the CSI at the current time t.

[0209] Method 2: The CSI at the future time is first predicted according to the estimated historical channel information, and then the predicted CSI is compressed and reported through an AI model, or the CSI at the future time is predicted and compressed through an AI model.

[0210] For method 2, the CSI prediction and the CSI compression can be respectively realized through a corresponding UE-side unilateral CSI prediction AI / ML model and a bilateral CSI compression model, or a bilateral model can be used to realize the joint CSI prediction and CSI compression. If the joint CSI prediction and CSI compression function is realized through a bilateral CSI prediction compression model, the bilateral CSI prediction compression model still includes a CSI prediction compression part model deployed on the UE side and a recovered predicted CSI part model deployed on the NW side. FIG. 1C is a schematic diagram of CSI compressed feedback based on a bilateral CSI prediction compression model according to an embodiment of the present disclosure. As shown in FIG. 1C, the input information of the AI / ML-based CSI prediction compression part model is the channel information at the historical time, and the output information is the compressed CSI or the binary bit stream obtained by quantizing the compressed CSI. The UE reports the binary bit stream obtained by quantizing the compressed CSI to the NW side, and the binary bit stream is used as the input of the recovered predicted CSI part model on the NW side. The recovered M>=1 channel information at the future time is obtained through the AI / ML-based recovered predicted CSI part model.

[0211] In some embodiments, the CSI generation part of the bilateral CSI compression AI / ML model is located at the UE side, and its input information is the channel information at the current measurement time. The CSI recovery part of the bilateral CSI compression AI / ML model is located at the NW side, and its output information is the recovered channel information at the current measurement time. Compared with the aforementioned bilateral CSI prediction compression model, the input information of the generation part model at the UE side is different, and the output information of the recovery part model at the NW side is also different. Therefore, the traditional method for monitoring the bilateral CSI compression AI / ML model is no longer applicable to the bilateral CSI compression AI / ML model, and a new bilateral model monitoring method needs to be designed.

[0212] In some embodiments, the bilateral AI / ML model performance monitoring can include the following ways:

[0213] Way 1, NW side monitoring model performance:

[0214] Based on the target CSI reported by the UE side (the input of the encoder in FIG. 1B), the target CSI can be indicated by the eType II codebook or a higher-precision eType II codebook.

[0215] Way 2, UE side monitoring model performance:

[0216] Monitoring model performance based on decoder output information deployed at the UE side;

[0217] Note that the decoder deployed at the UE side can be the same as or different from the decoder deployed at the NW side, or it can be a reference model provided by the NW or a proxy model developed by the UE side;

[0218] Monitoring model performance based on estimated intermediate KPI;

[0219] Monitoring model based on monitoring results other than intermediate KPI;

[0220] Based on the precoded reference signal such as precoded CSI-RS or DMRS sent by the NW side, the precoding is obtained according to the decoder output information at the NW side;

[0221] Monitoring based on the decoder output information at the NW side, the decoder output information is indicated to the UE by the NW through the eType II codebook or the high-precision eType II codebook.

[0222] In some embodiments, the current performance monitoring method is only applicable to the CSI compression feedback without CSI prediction function, and is not applicable to the AI / ML model monitoring of the joint CSI prediction compression described in the background art.

[0223] FIG. 2A is an interaction schematic diagram of a model performance monitoring method according to an embodiment of the present disclosure. The method can be performed by the communication system described above. As shown in FIG. 2A, the method can include the following steps.

[0224] In step S2101, the network device obtains fifth information.

[0225] In some embodiments, the fifth information can be CSI at the first time point recovered by the second part model of the first model.

[0226] In some embodiments, the first model can include a first part model and a second part model, the first part model can be used for compression processing of CSI, and the second part model can be used for CSI recovery according to the CSI compressed by the first part model.

[0227] In some embodiments, the first part model can be deployed in the terminal device, and the second part model can be deployed in the network device.

[0228] For example, the first model can be a bilateral CSI prediction compression model capable of realizing joint CSI prediction and CSI compression functions, the first part model can compress the measured CSI or CSI obtained by model prediction (inference) to obtain compressed CSI or a binary bit stream after quantization of the compressed CSI, and the second part model can recover and predict according to the binary bit stream to obtain M (M > 1) CSI at future time points.

[0229] In some embodiments, the first part model can also predict M (M > 1) CSI at future time points according to the measured CSI or CSI obtained by model prediction (inference), compress the M CSI at future time points to obtain compressed M CSI at future time points or a binary bit stream after quantization of the compressed M CSI at future time points, and the second part model can recover according to the binary bit stream to obtain M (M > 1) CSI at future time points.

[0230] In some embodiments, the network device can recover the fifth information by the second part model.

[0231] In some embodiments, the first time point can be determined by the following methods:

[0232] Protocol agreement;

[0233] Network device indication;

[0234] Terminal device indication.

[0235] In some embodiments, the terminal device can negotiate with the network device to define the first time, the network device can indicate the first time, and the terminal device can also indicate the first time.

[0236] In some embodiments, if the network device indicates the first time, the network device can send first indication information to the terminal device, where the first indication information is used to indicate the first time; and if the terminal device indicates the first time, the terminal device can send the first indication information to the network device.

[0237] In some embodiments, the network device can send a first resource to the terminal device, where the first resource can be used by the terminal device to measure the sixth information.

[0238] In some embodiments, the network device can send the first resource to the terminal device at the second time.

[0239] In some embodiments, the first resource can be a pilot signal, such as a Channel Status Information-Reference Signal (CSI-RS) resource.

[0240] In some embodiments, the network device can send the first resource to the terminal device at the second time, the network device can also send the first resource to the terminal device periodically, the network device can also send a plurality of non-periodic first resources to the terminal device, for example, the network device can send a CSI burst containing 4 continuous non-periodic CSI-RS resources to the terminal device, the network device can also send a semi-persistent first resource to the terminal device, and the present disclosure does not limit this.

[0241] In some embodiments, the terminal device can receive the first resource sent by the network device at the second time, and measure the sixth information according to the first resource.

[0242] In some embodiments, after the terminal device measures the sixth information, the terminal device can input the sixth information into a first part model to obtain seventh information, and send the seventh information to the network device, where the seventh information is used by the network device to obtain the fifth information.

[0243] In some embodiments, after the terminal device inputs the sixth information into the first part model, the first part model can perform compression processing on the sixth information to obtain a binary bit stream, i.e., the seventh information, and send the seventh information to the network device.

[0244] In some embodiments, the network device can receive the seventh information sent by the terminal device, input the seventh information into the second part model, and obtain fifteenth information including recovered CSI of the fourth time, the fourth time including at least the first time.

[0245] In some embodiments, the first time can be any one or more of the fourth time.

[0246] In some embodiments, the network device can recover through the second part model to obtain CSI of the fourth time, the fourth time including at least the first time.

[0247] In some embodiments, after the network device inputs the seventh information into the second part model, the network device can perform recovery processing through the second part model to obtain CSI of the fourth time. For example, the fourth time includes M time points, the first time includes T time points, and T is less than or equal to M. The network device infers (recovered) CSI of the M time points through the second part model.

[0248] FIG. 2B is a schematic diagram of CSI transmission according to an embodiment of the present disclosure. As shown in FIG. 2B, for example, the network device can send a CSI-RS burst containing 4 consecutive aperiodic CSI-RS resources to the terminal device. The terminal device estimates historical full channel information corresponding to 4 time points according to the received CSI-RS burst, and takes the historical full channel information as input of the first part model at the terminal device side. After inference of the first part model, a binary bit stream with a size of 120 bits is obtained, which is reported to the network device by the terminal device at time n (corresponding to CSI reporting in FIG. 2B). The network device takes the received 120 bits binary bit stream as input of the second part model, and obtains recovered CSI of future time points M (for example, M = 4) through inference of the second part model.

[0249] In step S2102, the network device sends fifth information to the terminal device.

[0250] In some embodiments, the terminal device can receive the fifth information. For example, the terminal device can receive the fifth information sent by the network device. For another example, the terminal device can also receive the fifth information sent by other entities.

[0251] In some embodiments, if the network device recovers the fifteenth information, i.e., CSI of the fourth time, through the second part model, the network device can determine the fifth information from the fifteenth information, and send the fifth information to the terminal device.

[0252] In some embodiments, the network device can send the fifth information to the terminal device according to the fifteenth information.

[0253] For example, after inputting the seventh information into the second part model, the network device can perform recovery processing through the second part model to obtain the CSI at the fourth time. For example, the fourth time includes M time points, the first time includes T time points, and T is less than or equal to M. The network device infers (recovered) CSI at M time points through the second part model and sends (recovered) CSI at T time points to the terminal device. For example, as shown in FIG. 2B, the network device can send recovered CSI at t1 and t2 time points to the terminal device.

[0254] In some embodiments, the network device sending the fifth information to the terminal device can include at least one of the following:

[0255] sending all the fifth information at the first time to the terminal device at the same time;

[0256] sending the fifth information at each time in the first time to the terminal device in the first order.

[0257] In some embodiments, the first order can be a protocol agreement or an indication of the network device. For example, the first order can be from front to back, or from back to front.

[0258] In some embodiments, after obtaining the fifth information at each time in the first time, the network device can combine the plurality of fifth information in the first order and send the combined fifth information to the terminal device. For example, as shown in FIG. 2B, the first time can include T time points, corresponding to t1 and t2 time points in FIG. 2B, and t1 is before t2. The network device can combine the fifth information at t1 and t2 time points in the order of t1 first and t2 second, and send the fifth information at t1 and t2 time points to the terminal device. After receiving the fifth information sent by the network device, the terminal device can determine the fifth information at different times in the first order.

[0259] In some embodiments, the network device can also send the fifth information at each time in the first time in the first order. For example, as shown in FIG. 2B, the network device can first send the fifth information at t1 time point to the terminal device, and then send the fifth information at t2 time point to the terminal device.

[0260] In some embodiments, the network device can send the fifth information in at least one of the following ways: floating point quantization, codebook.

[0261] In some embodiments, the network device can send the fifth information to the terminal device through floating point quantization. For example, the network device can quantize the fifth information through float 32 and send it to the terminal device.

[0262] In some embodiments, the network device can send the fifth information to the terminal device in a codebook manner. For example, the network device can send the fifth information to the terminal device through a Rel-18 Type II Doppler codebook or an enhanced Rel-18 Type II Doppler codebook.

[0263] In some embodiments, the network device can send the fifth information to the terminal device in a floating-point quantization manner and a codebook manner. For example, the network device can send the fifth information at time t1 to the terminal device in a floating-point quantization manner and send the fifth information at time t2 to the terminal device in a codebook manner.

[0264] It should be noted that the above-mentioned manner of sending the fifth information is exemplary, and the embodiments of the present disclosure are not limited thereto.

[0265] Step S2103, the terminal device acquires the fourth information.

[0266] In some embodiments, the fourth information can be channel state information (CSI) at a first time.

[0267] In some embodiments, the first time can include multiple times, for example, the first time can include T times, and T is 2.

[0268] In some embodiments, the CSI can be channel information, or a feature vector of channel information, or other forms of information determined according to the channel information, and the embodiments of the present disclosure are not limited thereto.

[0269] In some embodiments, the terminal device can obtain the fourth information according to the resource measurement sent by the network device, or infer the fourth information according to the first indication information, or obtain the fourth information by combining measurement and inference. For example, the network device can send part of the resources at the first time, and the terminal device can measure part of the CSI at the first time, and infer the CSI at all the first time according to part of the CSI at the first time and / or the CSI at the historical time.

[0270] In some embodiments, acquiring the fourth information includes at least one of the following:

[0271] measuring the fourth information;

[0272] predicting the fourth information through a second model, which can be used to predict the CSI;

[0273] measuring the eighth information, and predicting the fourth information through the second model according to the eighth information.

[0274] In some embodiments, the terminal device can obtain the fourth information according to resource measurement of the resource sent by the network device.

[0275] In some embodiments, the terminal device can receive a second resource sent by the network device at the first time, and obtain the fourth information according to measurement of the second resource.

[0276] In some embodiments, after the network device restores the fifth information of the first time through the second part of the model, the network device can send the second resource to the terminal device at the first time. For example, as shown in FIG. 2B, after the network device infers the restored CSI of the M time through the second part of the model, the network device can send the second resource to the terminal device at the t1 time and the t2 time. The terminal device can measure the CSI of the t1 time according to the second resource at the t1 time, and measure the CSI of the t2 time according to the second resource at the t2 time.

[0277] In some embodiments, the sending mode of the second resource can refer to the sending mode of the first resource, which will not be described here.

[0278] In some embodiments, the terminal device can obtain the ninth information, input the tenth information into the second model to obtain the eleventh information, and determine the fourth information from the eleventh information according to the first indication information.

[0279] In some embodiments, the ninth information can be the CSI of the historical time.

[0280] In some embodiments, the terminal device can obtain the CSI measured at the historical time to obtain the ninth information.

[0281] In some embodiments, the second model can be a model for predicting the CSI.

[0282] In some embodiments, the tenth information can include the CSI of at least one historical time and / or the CSI of at least one first time predicted by the second model.

[0283] For example, if the first time includes 4 time points, the ninth information includes the CSI of 4 historical time points, after obtaining the ninth information, the terminal device can input the CSI of the 4 historical time points into the second model, obtain the CSI of the first time point by prediction of the second model, input the CSI of 3 historical time points and the predicted CSI of the first time point into the second model, obtain the CSI of the second time point by prediction of the second model, input the CSI of 2 historical time points, the predicted CSI of the first time point and the second time point into the second model, obtain the CSI of the third time point by prediction of the second model, input the CSI of 1 historical time point, the predicted CSI of the first time point, the predicted CSI of the second time point and the predicted CSI of the third time point into the second model, and obtain the CSI of the fourth time point by prediction of the second model.

[0284] It should be noted that the prediction process of the second model is exemplary, and the CSI input into the second model is not limited in the embodiments of the present disclosure. For example, the CSI of 4 first time points can be predicted by the CSI of 4 historical time points.

[0285] In some embodiments, the first indication information can be used to indicate the first time point.

[0286] In some embodiments, after the terminal device obtains the CSI of multiple time points by prediction of the second model, the CSI of the first time point can be determined according to the first indication information, so as to correspond to the time point of sending the CSI by the network device.

[0287] In some embodiments, the first indication information can be protocol agreement or network device indication.

[0288] In some embodiments, the first indication information can also be protocol agreement or terminal device indication. After the terminal device determines the fourth information according to the first indication information, the terminal device can inform the network device of the first time point corresponding to the fourth information through the first indication information.

[0289] In some embodiments, the terminal device can receive a third resource sent by the network device at a third time point, measure the eighth information according to the third resource, input the eighth information into the second model to obtain the twelfth information, and determine the fourth information from the twelfth information according to the first indication information.

[0290] In some embodiments, the third time point can include the first N time points in the first time point and / or historical time points before the first time point.

[0291] In some embodiments, the sending mode of the third resource can refer to the sending mode of the first resource, which will not be described here.

[0292] In some embodiments, if the first time period includes T time periods, T is less than or equal to M, the network device can send the third resource to the terminal device at M' (M' < M) time periods, the terminal device can obtain the CSI of the M' time periods according to the third resource measurement of the M' time periods, input the CSI of the M' time periods and / or the CSI of the historical time periods into the second model, and obtain the CSI of the remaining M-M' time periods through the second model. Then, the terminal device can determine the CSI of the T time periods from the measured CSI of the M' time periods and the predicted CSI of the M-M' time periods according to the first indication information, and obtain the fourth information.

[0293] It should be noted that the information input into the second model can only include the CSI of the M' time periods or the CSI of the historical time periods, and the embodiments of the present disclosure do not limit this.

[0294] In step S2104, the terminal device determines the performance of the first model according to the fourth information and the fifth information.

[0295] In some embodiments, the terminal device can determine the fourteenth information according to the fourth information and the fifth information, and the fourteenth information can be used to indicate the performance of the first model.

[0296] In some embodiments, the fourteenth information can include at least one of the following: square generalized cosine similarity (SGCS), normalized mean square error (NMSE).

[0297] In some embodiments, the terminal device can calculate the SGCS according to the fourth information and the fifth information through the following formula:

[0298] Wherein, K1 is the SGCS, i represents the i th frequency domain unit, such as a sub-band, N3 represents the number of frequency domain units, and T is the number of time periods included in the first time period. For example, T = 2 in FIG. 2B, and v i,t represents the fourth information of the i th frequency domain unit at the t th time period, e i,t represents the fifth information of the i th frequency domain unit at the t th time period.

[0299] In step S2105, the terminal device sends the performance of the first model to the network device.

[0300] In some embodiments, the network device can receive the performance of the first model. For example, the network device can receive the performance of the first model sent by the terminal device. For another example, the network device can also receive the performance of the first model sent by other entities.

[0301] In some embodiments, the terminal device can also send fourteenth information to the network device, and the network device can determine the performance of the first model according to the fourteenth information.

[0302] By using the above method, the terminal device can obtain the CSI at the first time, receive the CSI at the first time recovered by the network device through the second part of the model, and monitor the performance of the first model according to the obtained CSI at the first time and the recovered CSI at the first time, thereby reducing the signaling overhead during model performance monitoring.

[0303] The method related to the embodiments of the present disclosure can include at least one of the above steps S2101-S2105. For example, step S2101 can be implemented as an independent embodiment, step S2102 can be implemented as an independent embodiment, step S2103 can be implemented as an independent embodiment, step S2104 can be implemented as an independent embodiment, step S2105 can be implemented as an independent embodiment, step S2101+step S2102 can be implemented as an independent embodiment, step S2102+step S2103 can be implemented as an independent embodiment, step S2104+step S2105 can be implemented as an independent embodiment, step S2102+step S2103+step S2104 can be implemented as an independent embodiment, step S2101+step S2102+step S2103+step S2104 can be implemented as an independent embodiment, but not limited thereto.

[0304] In some embodiments, the order of any two steps in steps S2101-S2105 can be exchanged or executed simultaneously. For example, step S2102 and step S2103 can be exchanged or executed simultaneously.

[0305] In some embodiments, steps S2101-S2105 are optional, and one or more of these steps can be omitted or replaced in different embodiments. For example, step S2105 can be omitted.

[0306] In some embodiments, other optional implementations described before or after the description corresponding to FIG. 2A can be referred to.

[0307] FIG. 2C is an interaction schematic diagram of a model performance monitoring method according to an embodiment of the present disclosure. The method can be performed by the above communication system. As shown in FIG. 2C, the method can include:

[0308] Step S2301, the network device obtains fifth information.

[0309] The optional implementation of step S2301 can refer to the optional implementation of step S2101 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0310] In step S2302, the network device sends second indication information to the terminal device.

[0311] In some embodiments, the second indication information is used to instruct the terminal device to send fourth information.

[0312] In some embodiments, step S2302 can be omitted. For example, the network device can send, to the terminal device at the first time, a second resource used by the terminal device to measure the CSI of the first time to obtain the fourth information. In this way, the network device informs the terminal device to report the CSI of the first time by sending the second resource at the first time. For another example, the content and period of the fourth information reported by the terminal device can be agreed through a protocol.

[0313] In step S2303, the terminal device obtains the fourth information.

[0314] The optional implementation of step S2303 can refer to the optional implementation of step S2103 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0315] In some embodiments, if the terminal device receives the second indication information sent by the network device, the terminal device can obtain the fourth information.

[0316] In some embodiments, if the network device does not send the second indication information, and the terminal device receives the second resource sent by the network device at the first time, the terminal device can measure the CSI corresponding to the first time according to the second resource, that is, the fourth information.

[0317] In step S2304, the terminal device sends the fourth information to the network device.

[0318] In some embodiments, the network device can receive the fourth information. For example, the network device can receive the fourth information sent by the terminal device. For another example, the network device can also receive the fourth information sent by another entity.

[0319] In some embodiments, after the terminal device obtains the fourth information, the terminal device can send the fourth information to the network device.

[0320] In some embodiments, the terminal device sending the fourth information to the network device can include at least one of the following:

[0321] Sending all the fourth information of the first time to the network device at the same time;

[0322] The fourth information of each time point in the first time points is sequentially sent to the network device according to a second order.

[0323] In some embodiments, the second order can be a protocol convention or a network device indication. For example, the second order can be from front to back or from back to front.

[0324] In some embodiments, after obtaining the fourth information of each time point in the first time points, the terminal device can combine all the fourth information according to the second order, and send the combined fourth information of all time points to the network device at the same time. For example, after measuring the fourth information of t1 and t2, the terminal device can combine the two fourth information according to the order of t1 fourth information in front and t2 fourth information behind, and send the combined fourth information of t1 and t2 to the network device at the same time.

[0325] In some embodiments, the terminal device can also send the fourth information of each first time point to the network device after obtaining it. For example, after obtaining the fourth information of t1, the terminal device can send the fourth information of t1 to the network device, and after obtaining the fourth information of t2, the terminal device can send the fourth information of t2 to the network device.

[0326] In some embodiments, the terminal device sending the fourth information to the network device can include at least one of the following:

[0327] Sending the fourth information to the network device by floating point quantization;

[0328] Sending the fourth information to the network device by codebook;

[0329] Compressing the fourth information by a third part of a third model to obtain thirteenth information, and sending the thirteenth information to the network device.

[0330] In some embodiments, the third model can be a bilateral AI model, and the third model can include a third part model and a fourth part model, the third part model is used for compressing the CSI, and the fourth part model is used for recovering the compressed CSI.

[0331] In some embodiments, the third part model can be deployed on the terminal device, and the fourth part model can be deployed on the network device.

[0332] For example, the third model can be a bilateral CSI compression model capable of realizing joint CSI compression feedback and recovery functions, the third part model can compress the measured CSI or CSI obtained through model prediction (inference) to obtain compressed CSI or a binary bit stream after quantization of the compressed CSI, and the fourth part model can recover according to the binary bit stream to obtain CSI similar to the measured CSI or the inferred CSI. For example, the CSI measured by the terminal device is H, the third part model is used to compress and quantize H into a binary bit stream, the binary bit stream is sent to the network device, the network device inputs the binary bit stream into the fourth part model, and the fourth part model is used for recovery to obtain H' similar to H.

[0333] It should be noted that the third model can be a model trained in the prior art.

[0334] In some embodiments, the terminal device can send all fourth information of the first time to the network device at the same time through a floating point quantization manner.

[0335] In some embodiments, the terminal device can send all fourth information of the first time to the network device at the same time through a codebook manner. For example, the terminal device can send all fourth information of the first time to the network device at the same time through a Rel-18 Type II Doppler codebook or a high-precision Rel-18 Type II Doppler codebook.

[0336] In some embodiments, the terminal device can perform joint compression processing on all fourth information of the first time through a third part model in the third model according to the second order to obtain thirteenth information, and send the thirteenth information to the network device. For example, the terminal device can perform compression processing on the fourth information of t1 and the fourth information of t2 in the order from front to back through the third part model, quantize into a binary bit stream, and send the binary bit stream to the network device.

[0337] In some embodiments, the terminal device can send the fourth information to the network device through a floating point quantization manner and a codebook manner. For example, the terminal device can send the fourth information of t1 to the network device through a floating point quantization manner, and send the fourth information of t2 to the network device through a codebook manner.

[0338] In some embodiments, the terminal device can also sequentially compress each fourth information at the first time according to the second order through a third partial model in the third model to obtain thirteenth information, and send the thirteenth information to the network device. For example, the terminal device can compress the fourth information at t1 through the third partial model to obtain the thirteenth information at t1, and send the thirteenth information at t1 to the network device. Then, the terminal device can compress the fourth information at t2 through the third partial model to obtain the thirteenth information at t2, and send the thirteenth information at t2 to the network device.

[0339] In some embodiments, after receiving the thirteenth information sent by the terminal device, the network device can recover the thirteenth information through a fourth partial model of the third model to obtain the fourth information.

[0340] It should be noted that the fourth information recovered by the network device is approximately the same as the fourth information reported by the terminal device. The higher the approximation degree, the more accurate the performance of the determined first model. The approximation degree depends on the performance of the third model. The better the performance of the third model, the more accurate the fourth information recovered by the network device.

[0341] In step S2305, the network device determines the performance of the first model according to the fourth information and the fifth information.

[0342] In some embodiments, the network device can determine a fourteenth information according to the fourth information and the fifth information, the fourteenth information being used to indicate the performance of the first model; wherein the fourteenth information includes at least one of the following: square of cosine similarity SGCS, normalized mean square error NMSE, block error rate BLER, and throughput.

[0343] It should be noted that the specific manner in which the network device determines the fourteenth information can refer to the implementation manner of step S2104, which will not be described here.

[0344] By using the above method, the network device can receive the CSI at the first time sent by the terminal device, and recover the CSI at the first time through the second partial model. The performance of the first model is monitored according to the received CSI at the first time and the recovered CSI at the first time, which reduces the signaling overhead during model performance monitoring.

[0345] The method related to the embodiments of the present disclosure can include at least one of the steps S2301-S2305. For example, the step S2301 can be implemented as an independent embodiment, the step S2302 can be implemented as an independent embodiment, the step S2303 can be implemented as an independent embodiment, the step S2304 can be implemented as an independent embodiment, the step S2305 can be implemented as an independent embodiment, the step S2301+the step S2302 can be implemented as an independent embodiment, the step S2302+the step S2303 can be implemented as an independent embodiment, the step S2303+the step S2304 can be implemented as an independent embodiment, the step S2301+the step S2304+the step S2305 can be implemented as an independent embodiment, but not limited thereto.

[0346] In some embodiments, the order between any two of the steps S2301-S2305 can be exchanged or simultaneously performed.

[0347] In some embodiments, the steps S2301-S2305 are optional, and one or more of the steps can be omitted or replaced in different embodiments. For example, the step S2302 can be omitted.

[0348] In some embodiments, the names of information and the like are not limited to the names described in the embodiments, and the terms of “information”, “message”, “signal”, “signaling”, “report”, “configuration”, “indication”, “instruction”, “command”, “channel”, “parameter”, “domain”, “field”, “symbol”, “symbol”, “codebook”, “codeword”, “codepoint”, “bit”, “data”, “program”, “chip”, and the like can be replaced with each other.

[0349] In some embodiments, “acquire”, “obtain”, “get”, “receive”, “transmit”, “bidirectional transmission”, “send and / or receive” can be replaced with each other, which can be interpreted as receiving from other subjects, acquiring from protocols, acquiring from higher layers, obtaining by self-processing, autonomously implementing, and the like.

[0350] In some embodiments, the terms “sending”, “transmitting”, “reporting”, “issuing”, “transferring”, “bidirectional transferring”, “sending and / or receiving” and the like can be replaced by each other.

[0351] In some embodiments, the terms “certain”, “preseted”, “preset”, “set”, “indicated”, “certain”, “arbitrary”, “first” and the like can be replaced by each other, “certain A”, “preset A”, “preset A”, “set A”, “indicated A”, “certain A”, “arbitrary A”, “first A” can be interpreted as A specified in advance in a protocol or the like, A obtained by setting, configuration, or indication, or A specific, certain, arbitrary, or first A, but not limited thereto.

[0352] FIG. 3A is a flow diagram illustrating a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 3A, the present embodiment relates to a model performance monitoring method, which can be executed by a terminal device. The method can include:

[0353] Step S3101, receiving fifth information.

[0354] The optional implementation of step S3101 can refer to the optional implementation of step S2102 of FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0355] Step S3102, receiving a second resource.

[0356] The optional implementation of step S3102 can refer to the optional implementation of step S2103 of FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0357] Step S3103, measuring fourth information according to the second resource.

[0358] The optional implementation of step S3103 can refer to the optional implementation of step S2103 of FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0359] Step S3104, determining the performance of the first model according to the fourth information and the fifth information.

[0360] The optional implementation of step S3104 can refer to the optional implementation of step S2104 of FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0361] Step S3105, sending the performance of the first model.

[0362] The optional implementation of step S3105 can refer to the optional implementation of step S2105 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0363] The method involved in the embodiments of the present disclosure can include at least one of steps S3101-S3105. For example, step S3101 can be implemented as an independent embodiment, step S3102 can be implemented as an independent embodiment, step S3103 can be implemented as an independent embodiment, step S3104 can be implemented as an independent embodiment, step S3102+step S3103 can be implemented as an independent embodiment, step S3101+step S3102+step S3103+step S3104 can be implemented as an independent embodiment, but is not limited thereto.

[0364] In some embodiments, the order of any two of steps S3101-S3105 can be exchanged or executed simultaneously. For example, step S3101 and step S3102 can be exchanged or executed simultaneously.

[0365] In some embodiments, steps S3101-S3105 are optional, and one or more of these steps can be omitted or replaced in different embodiments. For example, step S3102 and step S3105 can be omitted.

[0366] FIG. 3B is a flow diagram illustrating a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 3B, the embodiments of the present disclosure involve a model performance monitoring method, which can be executed by a terminal device. The method can include:

[0367] Step S3201, receiving fifth information.

[0368] The optional implementation of step S3201 can refer to the optional implementation of step S2102 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0369] Step S3202, obtaining fourth information.

[0370] The optional implementation of step S3202 can refer to the optional implementation of step S2103 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0371] Step S3203, determining the performance of the first model according to the fourth information and the fifth information.

[0372] The optional implementation of step S3203 can refer to the optional implementation of step S2104 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0373] Step S3204: sending the performance of the first model.

[0374] The optional implementation of step S3204 can refer to the optional implementation of step S2105 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0375] The method involved in the embodiments of the present disclosure can include at least one of the above steps S3201 to step S3204. For example, step S3201 can be implemented as an independent embodiment, step S3202 can be implemented as an independent embodiment, step S3203 can be implemented as an independent embodiment, step S3204 can be implemented as an independent embodiment, step S3201+step S3202+step S3203 can be implemented as an independent embodiment, but not limited thereto.

[0376] In some embodiments, the order of any two steps among steps S3201 to step S3204 can be exchanged or executed simultaneously. For example, step S3201 and step S3202 can exchange the order or be executed simultaneously.

[0377] In some embodiments, steps S3201 to step S3204 are optional, and one or more of these steps can be omitted or replaced in different embodiments. For example, step S3204 can be omitted.

[0378] FIG. 3C is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 3C, the embodiments of the present disclosure involve a model performance monitoring method, which can be executed by a terminal device. The method can include:

[0379] Step S3301: obtaining fourth information.

[0380] The optional implementation of step S3301 can refer to the optional implementation of step S2103 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0381] Step S3302: obtaining fifth information by using the fourth model.

[0382] In some embodiments, the terminal device side can also deploy a fourth model, which is used for recovery processing of the compressed CSI.

[0383] In some embodiments, the fourth model can be a model identical in function to the second partial model. The fourth model can be identical to the second partial model or different from the second partial model, which is not limited in the embodiments of the present disclosure.

[0384] In some embodiments, the terminal device can acquire the fifth information through the fourth model.

[0385] It should be noted that the specific implementation manner of the terminal device to restore the fifth information through the fourth model can refer to the manner of the network device to acquire the fifth information in step S2101, which will not be described here.

[0386] In step S3303, the performance of the first model is determined according to the fourth information and the fifth information.

[0387] The optional implementation manner of step S3303 can refer to the optional implementation manner of step S2104 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be described here.

[0388] In step S3304, the performance of the first model is sent.

[0389] The optional implementation manner of step S3304 can refer to the optional implementation manner of step S2105 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be described here.

[0390] The method involved in the embodiments of the present disclosure can include at least one of steps S3301 to S3304. For example, step S3301 can be implemented as an independent embodiment, step S3302 can be implemented as an independent embodiment, step S3303 can be implemented as an independent embodiment, step S3304 can be implemented as an independent embodiment, step S3301+step S3302+step S3303 can be implemented as an independent embodiment, but is not limited thereto.

[0391] In some embodiments, the order of any two steps among steps S3301 to S3304 can be exchanged or executed simultaneously. For example, the order of step S3301 and step S3302 can be exchanged or executed simultaneously.

[0392] In some embodiments, steps S3301 to S3304 are optional, and one or more of the steps can be omitted or replaced in different embodiments. For example, step S3304 can be omitted.

[0393] FIG. 3D is a flow diagram illustrating a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 3D, the method of monitoring the performance of the model can be performed by the terminal device. The method can include:

[0394] Step S3401, receiving fifth information.

[0395] The optional implementation of step S3401 can refer to the optional implementation of step S2102 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0396] Step S3402, obtaining fourth information.

[0397] The optional implementation of step S3402 can refer to the optional implementation of step S2103 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0398] Step S3403, determining the performance of the first model according to the fourth information and the fifth information.

[0399] The optional implementation of step S3403 can refer to the optional implementation of step S2104 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0400] In some embodiments, the above steps are optional steps.

[0401] FIG. 3E is a flow diagram of a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 3E, the embodiments of the present disclosure relate to a model performance monitoring method, which can be executed by a terminal device. The method can include:

[0402] Step S3501, obtaining third information.

[0403] The optional implementation of step S3501 can refer to steps S2102-S2103 in FIG. 2A, the optional implementation of step S3302 in FIG. 3C, and other associated parts in the embodiments involved in FIG. 2A and FIG. 3C, which will not be repeated here.

[0404] In some embodiments, the third information can include fourth information and fifth information.

[0405] Step S3502, determining the performance of the first model according to the third information.

[0406] The optional implementation of step S3502 can refer to the optional implementation of step S2104 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0407] In some embodiments, the above steps are optional steps.

[0408] FIG. 3F is a flow diagram illustrating a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 3F, the embodiments of the present disclosure relate to a model performance monitoring method, which can be performed by a terminal device. The method can include the following steps S3601-S3603.

[0409] Step S3601: Obtain second information.

[0410] In some embodiments, the second information can be used to indicate the feedback performance of the codebook.

[0411] In some embodiments, the terminal device can obtain the feedback performance of the codebook in a manner of an existing protocol, which will not be described herein again.

[0412] Step S3602: Determine the performance of the first model according to the second information.

[0413] In some embodiments, the terminal device can determine the performance of the first model according to the second information.

[0414] For example, the terminal device can determine the block error rate (BLER) according to the second information, so as to reflect the performance of the first model through the BLER.

[0415] Step S3603: Transmit the performance of the first model.

[0416] The optional implementation of step S3603 can refer to the optional implementation of step S2105 in FIG. 2A and other associated parts in the embodiments related to FIG. 2A, which will not be described herein again.

[0417] In some embodiments, the above steps are optional steps.

[0418] FIG. 3G is a flow diagram illustrating a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 3G, the embodiments of the present disclosure relate to a model performance monitoring method, which can be performed by a terminal device. The method can include the following steps S3701-S3703.

[0419] Step S3701: Obtain first information.

[0420] The optional implementation of step S3701 can refer to the optional implementation of steps S2102-S2103 in FIG. 2A, step S3302 in FIG. 3C, step S3601 in FIG. 3F, and other associated parts in the embodiments related to FIG. 2A, FIG. 3C and FIG. 3F, which will not be described herein again.

[0421] In some embodiments, the third information can include fourth information and fifth information.

[0422] Step S3702: Determine the performance of the first model according to the first information.

[0423] The optional implementation of step S3702 can refer to the optional implementation of step S2104 in FIG.2A, step S3602 in FIG.3F, and other associated parts in the embodiments related to FIG.2A and FIG.3F, which are not described herein again.

[0424] In some embodiments, the first part model is deployed on the terminal device, and the second part model is deployed on a network device.

[0425] In some embodiments, the obtaining the fifth information includes:

[0426] The network device sends the fifth information.

[0427] In some embodiments, the receiving the fifth information sent by the network device includes at least one of:

[0428] Receiving all the fifth information of the first time sent by the network device at the same time;

[0429] Receiving the fifth information of each time in the first time sent by the network device in turn according to a first order, and the first order is protocol agreement or indication of the network device.

[0430] In some embodiments, the sending mode of the fifth information includes at least one of: floating point quantization mode, codebook mode.

[0431] In some embodiments, the method further includes:

[0432] Obtaining sixth information, the sixth information being CSI of a second time, and the second time being before the first time;

[0433] Inputting the sixth information into the first part model to obtain seventh information;

[0434] Sending the seventh information to the network device, and the seventh information being used by the network device to obtain the fifth information.

[0435] In some embodiments, the obtaining the sixth information includes:

[0436] Receiving a first resource sent by the network device at the second time;

[0437] Obtaining the sixth information according to the first resource.

[0438] In some embodiments, the obtaining the fourth information includes at least one of:

[0439] Obtaining the fourth information by measurement;

[0440] The fourth information is predicted by a second model, the second model being used for predicting CSI.

[0441] The eighth information is measured, and the fourth information is predicted by the second model according to the eighth information.

[0442] In some embodiments, the measuring of the fourth information comprises:

[0443] A second resource sent by the network device at the first time point is received.

[0444] The fourth information is measured according to the second resource.

[0445] In some embodiments, the predicting of the fourth information by the second model comprises:

[0446] Ninth information is obtained, the ninth information being CSI at a historical time point.

[0447] Tenth information is input into the second model to obtain eleventh information, the tenth information comprising at least one CSI at a historical time point and / or at least one CSI at the first time point predicted by the second model.

[0448] The fourth information is determined from the eleventh information according to first indication information, the first indication information being used for indicating the first time point.

[0449] In some embodiments, the measuring of the eighth information and the predicting of the fourth information by the second model according to the eighth information comprise:

[0450] A third resource sent by the network device at a third time point is received, the third time point comprising the first N time points in the first time point and / or historical time points before the first time point.

[0451] The eighth information is measured according to the third resource.

[0452] The eighth information is input into the second model to obtain twelfth information.

[0453] The fourth information is determined from the twelfth information according to first indication information, the first indication information being used for indicating the first time point.

[0454] In some embodiments, the first indication information is protocol agreement or network device indication.

[0455] In some embodiments, the method further comprises:

[0456] The fourth information is sent to the network device.

[0457] In some embodiments, the sending the fourth information to the network device comprises at least one of:

[0458] sending all the fourth information of the first time to the network device at the same time;

[0459] sending the fourth information of each time in the first time to the network device in turn according to a second order, the second order being a protocol agreement or an indication of the terminal device.

[0460] In some embodiments, the sending the fourth information to the network device comprises at least one of:

[0461] sending the fourth information to the network device by a floating-point quantization method;

[0462] sending the fourth information to the network device by a codebook method;

[0463] performing compression processing on the fourth information by a third partial model in a third model to obtain thirteenth information, and sending the thirteenth information to the network device, the third model comprising the third partial model and a fourth partial model, the third partial model being used for compression processing on CSI, and the fourth partial model being used for recovery processing on the CSI after compression processing.

[0464] In some embodiments, the method further comprises:

[0465] receiving second indication information sent by the network device, the second indication information being used for indicating the terminal device to send the fourth information.

[0466] In some embodiments, a fourth model is further deployed on the terminal device, the fourth model being used for recovery processing on the CSI after compression processing.

[0467] In some embodiments, the obtaining the fifth information comprises:

[0468] obtaining the fifth information by the fourth model.

[0469] In some embodiments, the determining the performance of the first model according to the first information comprises:

[0470] determining fourteenth information according to the first information, the fourteenth information being used for indicating the performance of the first model; wherein the fourteenth information comprises at least one of: square of cosine similarity SGCS, normalized mean square error NMSE, block error rate BLER, and throughput.

[0471] In some embodiments, the method further comprises:

[0472] sending the performance of the first model to a network device.

[0473] In some embodiments, the first time point is determined by:

[0474] a protocol agreement;

[0475] a network device indication;

[0476] a terminal device indication.

[0477] FIG. 4A is a flow diagram illustrating a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 4A, the present embodiment relates to a model performance monitoring method, which can be performed by a network device. The method can include:

[0478] Step S4101, obtaining fifth information.

[0479] Optional implementation of step S4101 can refer to optional implementation of step S2301 in FIG. 2C and other associated parts in the embodiments involved in FIG. 2C, which will not be repeated here.

[0480] Step S4102, sending second indication information.

[0481] Optional implementation of step S4102 can refer to optional implementation of step S2302 in FIG. 2C and other associated parts in the embodiments involved in FIG. 2C, which will not be repeated here.

[0482] Step S4103, receiving fourth information.

[0483] Optional implementation of step S4103 can refer to optional implementation of step S2304 in FIG. 2C and other associated parts in the embodiments involved in FIG. 2C, which will not be repeated here.

[0484] Step S4104, determining the performance of the first model according to the fourth information and the fifth information.

[0485] Optional implementation of step S4104 can refer to optional implementation of step S2305 in FIG. 2C and other associated parts in the embodiments involved in FIG. 2C, which will not be repeated here.

[0486] The method related to the embodiments of the present disclosure can include at least one of the steps S4101-S4104. For example, the step S4101 can be implemented as an independent embodiment, the step S4102 can be implemented as an independent embodiment, the step S4103 can be implemented as an independent embodiment, the step S4104 can be implemented as an independent embodiment, the step S4102+the step S4103 can be implemented as an independent embodiment, but the present disclosure is not limited thereto.

[0487] In some embodiments, any two steps among the steps S4101-S4104 can be exchanged in order or executed simultaneously. For example, the step S4101 and the step S4102 can be exchanged in order or executed simultaneously.

[0488] In some embodiments, the steps S4101-S4104 are optional, and one or more of the steps can be omitted or replaced in different embodiments. For example, the step S4102 can be omitted.

[0489] FIG. 4B is a flow diagram illustrating a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 4B, the embodiments of the present disclosure relate to a model performance monitoring method, which can be executed by a network device. The method can include the following steps:

[0490] Step S4201, obtaining fifth information.

[0491] Optional implementation of the step S4201 can be referred to optional implementation of step S2301 in FIG. 2C and other associated parts in the embodiments related to FIG. 2C, which will not be described here.

[0492] Step S4202, receiving fourth information.

[0493] Optional implementation of the step S4202 can be referred to optional implementation of step S2304 in FIG. 2C and other associated parts in the embodiments related to FIG. 2C, which will not be described here.

[0494] Step S4203, determining the performance of the first model according to the fourth information and the fifth information.

[0495] Optional implementation of the step S4203 can be referred to optional implementation of step S2305 in FIG. 2C and other associated parts in the embodiments related to FIG. 2C, which will not be described here.

[0496] In some embodiments, the above steps are optional steps.

[0497] FIG. 4C is a flow diagram illustrating a method for monitoring model performance, according to an embodiment of the present disclosure. As shown in FIG. 4C, the embodiments of the present disclosure relate to a method for monitoring model performance, which can be performed by a network device. The method can include:

[0498] At step S4301, second information is obtained.

[0499] Optional implementation of step S4301 can refer to optional implementation of step S3601 in FIG. 3F and other associated parts in embodiments related to FIG. 3F. Details are not described herein again.

[0500] At step S4302, performance of the first model is determined according to the second information.

[0501] Optional implementation of step S4302 can refer to optional implementation of step S3602 in FIG. 3F and other associated parts in embodiments related to FIG. 3F. Details are not described herein again.

[0502] In some embodiments, the above steps are optional steps.

[0503] FIG. 4D is a flow diagram illustrating a method for monitoring model performance, according to an embodiment of the present disclosure. As shown in FIG. 4D, the embodiments of the present disclosure relate to a method for monitoring model performance, which can be performed by a network device. The method can include:

[0504] At step S4401, first information is obtained.

[0505] Optional implementation of step S4401 can refer to steps S2301-S2304 in FIG. 2C, optional implementation of step S4301 in FIG. 4C, and other associated parts in embodiments related to FIG. 2C and FIG. 4C. Details are not described herein again.

[0506] At step S4402, performance of the first model is determined according to the first information.

[0507] Optional implementation of step S4402 can refer to step S2305 in FIG. 2C, optional implementation of step S4302 in FIG. 4C, and other associated parts in embodiments related to FIG. 2C and FIG. 4C. Details are not described herein again.

[0508] In some embodiments, the first part of the model is deployed on the terminal device, and the second part of the model is deployed on the network device.

[0509] In some embodiments, the method further includes:

[0510] receive seventh information sent by the terminal device, the seventh information being determined by the terminal device according to sixth information by using the first part model, the sixth information being CSI at a second time point, the second time point being before the first time point;

[0511] input the seventh information into the second part model to obtain fifteenth information, the fifteenth information including recovered CSI at a fourth time point, the fourth time point including at least the first time point.

[0512] In some embodiments, the method further includes:

[0513] send, to the terminal device, first resources, the first resources being used by the terminal device to measure to obtain the sixth information.

[0514] In some embodiments, the method further includes:

[0515] send, to the terminal device, the fifth information according to the fifteenth information.

[0516] In some embodiments, the sending, to the terminal device, of the fifth information includes at least one of:

[0517] send, to the terminal device, all fifth information of the first time point at the same time point;

[0518] send, to the terminal device, the fifth information of each time point in the first time point in a first order, the first order being agreed by a protocol or indicated by the network device.

[0519] In some embodiments, the sending mode of the fifth information includes at least one of: a floating-point quantization mode and a codebook mode.

[0520] In some embodiments, the obtaining of the fourth information includes:

[0521] receive the fourth information sent by the terminal device.

[0522] In some embodiments, the receiving of the fourth information sent by the terminal device includes at least one of:

[0523] receive all fourth information of the first time point sent by the terminal device at the same time point;

[0524] receive the fourth information of each time point in the first time point sent by the terminal device in a second order, the second order being agreed by a protocol or indicated by the terminal device.

[0525] In some embodiments, the receiving of the fourth information sent by the terminal device includes at least one of:

[0526] receiving the fourth information sent by the terminal device in a floating-point quantization manner;

[0527] receiving the fourth information sent by the terminal device in a codebook manner;

[0528] receiving thirteenth information sent by the terminal device, the thirteenth information being obtained by compressing the fourth information by a third partial model in a third model, the third model comprising the third partial model and a fourth partial model, the third partial model being configured to compress the CSI, and the fourth partial model being configured to recover the compressed CSI.

[0529] In some embodiments, the method further comprises:

[0530] recovering the fourth information by the fourth partial model in the third model from the thirteenth information.

[0531] In some embodiments, the method further comprises:

[0532] sending a second resource to the terminal device at the first time, the second resource being configured to be used by the terminal device to measure the fourth information.

[0533] In some embodiments, the method further comprises:

[0534] sending a third resource to the terminal device, the third resource being configured to be used by the terminal device to measure eighth information, predict twelfth information from the eighth information, and determine the fourth information from the twelfth information according to first indication information, the first indication information being configured to indicate the first time.

[0535] In some embodiments, the first indication information is protocol agreement or indication of the terminal device.

[0536] In some embodiments, the method further comprises:

[0537] sending second indication information to the terminal device, the second indication information being configured to indicate the terminal device to send the fourth information.

[0538] In some embodiments, the determining the performance of the first model according to the first information comprises:

[0539] determining fourteenth information according to the first information, the fourteenth information being configured to indicate the performance of the first model; wherein the fourteenth information comprises at least one of square of cosine similarity SGCS, normalized mean square error NMSE, block error rate BLER, and throughput.

[0540] In some embodiments, the method further comprises:

[0541] receiving the performance of the first model sent by the terminal device.

[0542] In some embodiments, the first time point is determined by:

[0543] an agreement of protocol;

[0544] an indication of the network device;

[0545] an indication of the terminal device.

[0546] FIG. 5A is an interaction schematic diagram of a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 5A, the embodiment of the present disclosure relates to a model performance monitoring method, which can be executed by a communication system. The method can comprise:

[0547] S5101, the network device sends fifth information to the terminal device.

[0548] The optional implementation of step S5101 can refer to the optional implementation of step S2102 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0549] S5102, the terminal device acquires fourth information.

[0550] The optional implementation of step S5102 can refer to the optional implementation of step S2103 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0551] S5103, the terminal device determines the performance of the first model according to the fourth information and the fifth information.

[0552] The optional implementation of step S5103 can refer to the optional implementation of step S2104 in FIG. 2A and other associated parts in the embodiments involved in FIG. 2A, which will not be repeated here.

[0553] FIG. 5B is an interaction schematic diagram of a model performance monitoring method according to an embodiment of the present disclosure. As shown in FIG. 5B, the embodiment of the present disclosure relates to a model performance monitoring method, which can be executed by a communication system. The method can comprise:

[0554] S5201, the network device acquires fifth information.

[0555] The optional implementation of step S5201 can refer to the optional implementation of step S2301 in FIG. 2C and other associated parts in the embodiments involved in FIG. 2C, details are not repeated here.

[0556] In step S5202, the terminal device sends fourth information to the network device.

[0557] The optional implementation of step S5202 can refer to the optional implementation of step S2304 in FIG. 2C and other associated parts in the embodiments involved in FIG. 2C, details are not repeated here.

[0558] In step S5203, the network device determines the performance of the first model according to the fourth information and the fifth information.

[0559] The optional implementation of step S5203 can refer to the optional implementation of step S2305 in FIG. 2C and other associated parts in the embodiments involved in FIG. 2C, details are not repeated here.

[0560] In some embodiments, the above method can include the method described in the embodiments of the above communication system, terminal device, network device, etc., details are not repeated here.

[0561] In some embodiments, the embodiments of the present disclosure can obtain the CSI through the traditional codebook or enhanced codebook feedback mode between the UE and the NW based on the downlink pilot measurement or the AI / ML model prediction, and realize the monitoring of the bilateral AI / ML model at the UE side or the NW side.

[0562] In some embodiments, the NW side can implement performance monitoring, and the NW side monitors the performance of the bilateral CSI prediction compression model based on the CSI measurement of the future M time instants and the future M time instants of the CSI recovered by the prediction CSI model inference at the NW side. The criterion for monitoring the bilateral model can be SGCS or NMSE.

[0563] In some embodiments, the method for obtaining the CSI of the future M time instants can include:

[0564] Method 1: the NW sends downlink pilot signals at the future M time instants, and the UE obtains the CSI of the future M time instants according to the received pilot signals.

[0565] Method 2: the CSI of the future M time instants is obtained through the CSI prediction model inference at the UE side AI / ML. The input of the AI / ML CSI prediction model can be the CSI of the N historical time instants, or the CSI of part or all of the future M time instants obtained by the inference of the AI / ML CSI prediction model.

[0566] Way 3, a combination of way 1 and way 2, i.e., the NW sends downlink pilot signals to the UE at the first M' (M' < M) of the future M time instants, and then the UE infers the CSI corresponding to the remaining M-M' time instants through the AI / ML CSI prediction model.

[0567] In some embodiments, the reporting mode of the estimated CSI (M time instants) can include:

[0568] Way 1, through a Rel-18 Type II Doppler codebook or a high-precision Rel-18 Type II Doppler codebook to realize the CSI reporting of the future M time instants.

[0569] Way 2, through the CSI generation part of a two-sided AI / ML model for CSI compression feedback to realize the CSI reporting of the future M time instants, which can be separately compressed for each time instant corresponding to the future CSI, or jointly compressed for the CSI of the future M time instants, and then the UE quantizes the CSI generation part of the model and reports it to the NW side. The NW side infers the CSI of the future M time instants through the CSI recovery part of the CSI compression feedback two-sided AI / ML model.

[0570] In some embodiments, the UE side can implement performance monitoring, and the UE side monitors the performance of the two-sided CSI prediction compression model based on the measured CSI of the future T time instants and the CSI of the future T time instants inferred by the recovery prediction CSI model sent by the NW side. The criterion for monitoring the two-sided model can be SGCS or NMSE.

[0571] Way 1, the NW sends downlink pilot signals and recovery prediction CSI corresponding to the T (T < M) time instants to the UE, and the recovery prediction CSI of the T time instants is obtained by the recovery prediction CSI model in the two-sided CSI prediction compression model on the NW side. The recovery prediction CSI corresponding to the T time instants can be sent to the UE through floating point quantization or traditional codebook or enhanced codebook (such as Rel-18 Type II Doppler codebook or higher-precision Rel-18 Type II Doppler codebook). The UE determines the performance of the AI / ML model based on the estimated CSI from the received downlink pilot and the received recovery prediction CSI.

[0572] Way 2, the NW sends the recovery prediction CSI corresponding to the future T time instants to the UE. The UE infers the CSI corresponding to the future T time instants through the AI / ML CSI prediction model, and then the UE monitors the performance of the two-sided CSI prediction compression model based on the inferred CSI of the T time instants and the received recovery prediction CSI.

[0573] Way 3, UE side deploys the same or different recovery prediction CSI model as NW side, UE side monitors the performance of the bilateral AI / ML model based on the measured CSI at T time instants and the recovered CSI at future T time instants inferred by the UE side recovery prediction CSI model, or sends the monitored results to the NW side.

[0574] In some embodiments, the NW determines the performance of the bilateral model based on the calculated results at M time instants or more than M future time instants. Or the NW determines the performance of the bilateral model based on the calculation results reported by the UE at T time instants or more than T future time instants. The UE reported can be the average of SGCS or NMMSE corresponding to T or more than T future time instants.

[0575] In some embodiments, the feedback performance based on the traditional Rel-18 Type II Doppler codebook or the high-precision Rel-18 Type II Doppler codebook is used as the reference performance of the bilateral model. The performance can be the calculation result of SGCS / NMSE, or the calculation result of determined BLER or throughput.

[0576] In some embodiments, if the recovery prediction CSI part model can infer the CSI at future M time instants, the NW side can send downlink pilots at T (T≤M) time instants, and then the NW sends the inferred recovery CSI at future T time instants corresponding to the T time instants to the UE. Alternatively, the UE sends the CSI at T time instants to the NW side. The T time instants can be pre-defined through UE and NW negotiation, or reported by the UE to the NW, or indicated by the NW to the UE through one or more of RRC / MAC-CE or DCI signaling.

[0577] In some embodiments, the above-mentioned CSI can be full channel information, feature vectors corresponding to full channel information, or information pre-processed from full channel information such as DFT inverse transform.

[0578] Embodiment 1 (NW side monitors model performance)

[0579] Assume that the CSI prediction compression part model and the recovered predicted CSI part model are deployed at the UE side and the NW side respectively. The NW side sends a CSI burst containing 4 consecutive aperiodic CSI-RS resources to the UE, and the UE estimates the historical full channel information corresponding to 4 time points according to the received CSI burst, and takes it as the input of the UE-side CSI prediction compression part model. After the inference of the part model, a binary bit stream with a size of 120 bits is obtained, which is reported to the NW by the UE at time n, as shown in FIG. 2B. The NW takes the received 120 bits binary bit stream as the input based on the recovered predicted CSI part model, and obtains the recovered CSI of the future time point M = 4 time points through the inference of the part model. In order to monitor the bilateral model, the NW sends CSI-RS to the UE at future T = 2 time points, as shown in FIG. 2B. The UE estimates the downlink CSI information through the CSI-RS received at the two time points.

[0580] Assume that the CSI information is the channel feature vector corresponding to the full channel information, and the UE can report the channel feature vectors at T = 2 time points through float 32 quantization, or report the CSI corresponding to the two time points to the NW through the Rel-18 Type II Doppler codebook or enhanced Rel-18 Type II Doppler codebook parameters. Assume that the CSI at the T time points is determined according to the predefinition, and the NW side calculates the SGCS according to the CSI (defined as v i,t ) received at the two time points and the CSI (defined as e i,t ) obtained by the NW-side recovered predicted CSI part model, through formula (1).

[0581] In some embodiments, the NW side determines the performance of the bilateral model according to the average value of the calculated SGCS at the two time points.

[0582] Embodiment 2 (UE side monitoring model performance)

[0583] Assume that the CSI prediction compression part model and the recovered predicted CSI part model are deployed at the UE side and the NW side respectively. The NW side sends a CSI burst containing 4 consecutive aperiodic CSI-RS resources to the UE, and the UE estimates the historical full channel information corresponding to 4 time points according to the received CSI burst, and takes it as the input of the UE-side CSI prediction compression part model. After the inference of the part model, a binary bit stream with a size of 120 bits is obtained, which is reported to the NW by the UE at time n, as shown in FIG. 2B. The NW takes the received 120 bits binary bit stream as the input based on the recovered predicted CSI part model, and obtains the recovered CSI of the future time point M = 4 time points through the inference of the part model.

[0584] To monitor the double-sided model, the NW transmits CSI-RS to the UE at two future time instances T = 2, as shown in two time instances t1 and t2 in FIG. 2B. The UE estimates the downlink CSI information by the CSI-RS received at the two time instances. In addition, the NW also transmits the recovered predicted CSI corresponding to the two time instances to the UE, which can be quantized by the NW through float 32 or sent to the UE by the Rel-18 Type II Doppler codebook or enhanced Rel-18 Type II Doppler codebook parameter mode. Similar to Embodiment 1, the UE calculates the SGCS result by formula (1) according to the estimated CSI and the received recovered CSI. The result can be reported to the NW by the UE, and then the performance of the double-sided model is determined by the NW.

[0585] In some embodiments of the present disclosure, a communication system can include a terminal device and a network device, wherein the terminal device can perform the model performance monitoring method performed by the terminal device in the foregoing embodiments of the present disclosure; and the network device can perform the model performance monitoring method performed by the network device in the foregoing embodiments of the present disclosure.

[0586] Embodiments of the present disclosure also propose an apparatus for implementing any of the above methods, for example, an apparatus including units or modules for implementing each step performed by the terminal in any of the above methods. For another example, another apparatus is also proposed, including units or modules for implementing each step performed by the network device (such as an access network device, a core network function node, a core network device, etc.) in any of the above methods.

[0587] It should be understood that the division of each unit or module in the above apparatus is only a logical function division, and all or part of them can be integrated into a physical entity or physically separated in actual implementation. In addition, the units or modules in the apparatus can be implemented in the form of processor calling software: for example, the apparatus includes a processor connected with a memory, the memory stores instructions, and the processor calls the instructions stored in the memory to implement any of the above methods or realize the functions of the units or modules of the above apparatus, wherein the processor is a general processor such as a central processing unit (CPU) or a microprocessor, and the memory is a memory in the apparatus or a memory outside the apparatus. Alternatively, the units or modules in the apparatus can be implemented in the form of hardware circuit, and the functions of part or all of the units or modules can be realized by the design of hardware circuit. The above hardware circuit can be understood as one or more processors; for example, in one implementation, the above hardware circuit is an application-specific integrated circuit (ASIC), and the functions of part or all of the units or modules are realized by the design of the logical relationship of elements in the circuit; for another example, in another implementation, the above hardware circuit is a programmable logic device (PLD), and a field programmable gate array (FPGA) is taken as an example, which can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by a configuration file, so as to realize the functions of part or all of the units or modules. All units or modules of the above apparatus can be implemented in the form of processor calling software, or all units or modules can be implemented in the form of hardware circuit, or part of the units or modules are implemented in the form of processor calling software, and the remaining part is implemented in the form of hardware circuit.

[0588] In the embodiments of the present disclosure, the processor is a circuit with signal processing capability. In one implementation, the processor can be a circuit with instruction reading and running capability, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a microprocessor), a digital signal processor (DSP), or the like. In another implementation, the processor can implement certain functions through a logical relationship of a hardware circuit, and the logical relationship of the hardware circuit is fixed or reconfigurable. For example, the processor is a hardware circuit implemented by an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as an FPGA. In the reconfigurable hardware circuit, the processor loads a configuration document to implement the hardware circuit configuration. It can be understood that the processor loads instructions to implement the functions of the above part or all units or modules. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), a tensor processing unit (TPU), a deep learning processing unit (DPU), and the like.

[0589] FIG. 6A is a schematic diagram of a structure of a terminal device according to an embodiment of the present disclosure. As shown in FIG. 6A, the terminal device 101 can include at least one of a transceiver module 6101, a processing module 6102, and the like. In some embodiments, the transceiver module 6101 is configured to acquire first information, the first information including second information and / or third information, the second information being used to indicate feedback performance of a codebook, the third information including fourth information and fifth information, the fourth information being channel state information (CSI) at a first time, and the fifth information being the CSI at the first time recovered by a second part model of a first model, the first model including a first part model and the second part model, the first part model being used to perform compression processing on the CSI, and the second part model being used to perform CSI recovery according to the CSI compressed by the first part model, and the processing module 6102 is configured to determine performance of the first model according to the first information. Optionally, the transceiver module 6101 can be used to perform at least one of the communication steps (for example, steps S2102 and S2105, but not limited thereto) of the sending and / or receiving performed by the terminal device 101 in any of the above methods, which will not be described herein. Optionally, the processing module 6102 can be used to perform at least one of the other steps (for example, steps S2103 and S2104, but not limited thereto) performed by the terminal device 101 in any of the above methods, which will not be described herein.

[0590] In some embodiments, the transceiver module can include a sending module and / or a receiving module, and the sending module and the receiving module can be separate or integrated together. Optionally, the transceiver module can be mutually replaced with a transceiver.

[0591] In some embodiments, the processing module can be one module or can include multiple sub-modules. Optionally, the multiple sub-modules perform all or part of the steps required to be performed by the processing module. Optionally, the processing module can be mutually replaced with a processor.

[0592] FIG. 6B is a structural schematic diagram of a network device according to an embodiment of the present disclosure. As shown in FIG. 6B, the network device 102 can include at least one of a transceiver module 6201, a processing module 6202, and the like. In some embodiments, the transceiver module 6201 is configured to obtain first information, wherein the first information includes second information and / or third information, the second information is used to indicate feedback performance of a codebook, and the third information includes fourth information and fifth information, the fourth information is channel state information (CSI) at a first time, and the fifth information is the CSI at the first time recovered by a second part model of a first model, the first model includes a first part model and the second part model, the first part model is used for compression processing of the CSI, and the second part model is used for CSI recovery according to the CSI after the compression processing of the first part model; and the processing module 6202 is configured to determine performance of the first model according to the first information. Optionally, the transceiver module 6201 can be used to perform at least one of the communication steps (for example, step S2302, but not limited thereto) of the sending and / or receiving performed by the network device 102 in any of the above methods, and details are not described herein again. Optionally, the processing module 6202 can be used to perform at least one of the other steps (for example, step S2305, but not limited thereto) performed by the network device 102 in any of the above methods, and details are not described herein again.

[0593] In some embodiments, the transceiver module can include a sending module and / or a receiving module, and the sending module and the receiving module can be separate or integrated together. Optionally, the transceiver module can be mutually replaced with a transceiver.

[0594] In some embodiments, the processing module can be one module, or can include multiple sub-modules. Optionally, the multiple sub-modules perform all or part of the steps required to be performed by the processing module. Optionally, the processing module can be mutually replaced with a processor.

[0595] FIG. 7A is a structural schematic diagram of a communication device 7100 according to an embodiment of the present disclosure. The communication device 7100 can be a network device (for example, an access network device, a core network device, and the like), a terminal (for example, a user equipment, and the like), a chip, a chip system, or a processor supporting the network device to implement any of the above methods, or a chip, a chip system, or a processor supporting the terminal to implement any of the above methods. The communication device 7100 can be used to implement the methods described in the above method embodiments, and details can be referred to the descriptions in the above method embodiments.

[0596] As shown in FIG. 7A, the communication device 7100 includes one or more processors 7101. The processor 7101 can be a general processor or a special-purpose processor, etc., for example, a baseband processor or a central processing unit. The baseband processor can be used to process communication protocols and communication data, the central processing unit can be used to control the communication device (e.g., a base station, a baseband chip, an Internet of Things device, an Internet of Things device chip, a DU or a CU, etc.), execute programs, and process data of the programs. The communication device 7100 is configured to perform any of the above methods.

[0597] In some embodiments, the communication device 7100 further includes one or more memories 7102 for storing instructions. Alternatively, all or part of the memory 7102 can also be outside the communication device 7100.

[0598] In some embodiments, the communication device 7100 further includes one or more transceivers 7103. When the communication device 7100 includes one or more transceivers 7103, the transceiver 7103 performs at least one of the communication steps (e.g., step S2101, step S2105, but not limited to) in the above methods, and the processor 7101 performs at least one of the other steps (e.g., step S2103, but not limited to).

[0599] In some embodiments, the transceiver can include a receiver and / or a transmitter, which can be separate or integrated together. Alternatively, the terms transceiver, transceiving unit, transceiver, transceiving circuit, etc. can be replaced by each other, the terms transmitter, transmitting unit, transmitter, transmitting circuit, etc. can be replaced by each other, and the terms receiver, receiving unit, receiver, receiving circuit, etc. can be replaced by each other.

[0600] In some embodiments, the communication device 7100 can include one or more interface circuits. Alternatively, the interface circuit is connected with the memory 7102, and the interface circuit can be used to receive signals from the memory 7102 or other devices, and can be used to send signals to the memory 7102 or other devices. For example, the interface circuit can read the instructions stored in the memory 7102 and send the instructions to the processor 7101.

[0601] The communication device 7100 described in the above embodiments can be the first device or the IoT device, but the scope of the communication device 7100 described in the present disclosure is not limited thereto, and the structure of the communication device 7100 can not be limited by FIG. 7A. The communication device can be a standalone device or can be part of a larger device. For example, the communication device can be: 1) a standalone integrated circuit (IC), or a chip, or a chip system or subsystem; (2) a set of one or more ICs, which can optionally also include storage components for storing data, programs; (3) an ASIC, such as a modem; (4) a module that can be embedded in other devices; (5) a receiver, an IoT device, a smart IoT device, a cellular phone, a wireless device, a handset, a mobile unit, a car device, a first device, a cloud device, an artificial intelligence device, and the like; (6) other devices, and the like.

[0602] FIG. 7B is a structural diagram of a chip 7200 according to an embodiment of the present disclosure. For the case where the communication device 7100 is a chip or a chip system, the structural diagram of the chip 7200 shown in FIG. 7B can be referred to, but is not limited thereto.

[0603] The chip 7200 includes one or more processors 7201, and the chip 7200 is configured to execute any of the above methods.

[0604] In some embodiments, the chip 7200 further includes one or more interface circuits 7203. Optionally, the interface circuit 7203 is connected to the memory 7202, and the interface circuit 7203 can be configured to receive signals from the memory 7202 or other devices, and the interface circuit 7203 can be configured to send signals to the memory 7202 or other devices. For example, the interface circuit 7203 can read instructions stored in the memory 7202 and send the instructions to the processor 7201.

[0605] In some embodiments, the interface circuit 7203 performs at least one of the communication steps (such as step S2101, step S2105, but not limited thereto) in the above methods, and the processor 7201 performs at least one of the other steps (such as step S2103, but not limited thereto).

[0606] In some embodiments, the terms interface circuit, interface, transceiver pin, and transceiver can be replaced with each other.

[0607] In some embodiments, the chip 7200 further includes one or more memories 7202 for storing instructions. Optionally, all or part of the memory 7202 can be outside the chip 7200.

[0608] The embodiments of the present disclosure further provide a storage medium having stored instructions, which, when executed on the communication device 7100, cause the communication device 7100 to perform any of the above methods. Alternatively, the storage medium is an electronic storage medium. Alternatively, the storage medium is a computer-readable storage medium, but is not limited to this, and can also be a storage medium readable by other devices. Alternatively, the storage medium can be a non-transitory storage medium, but is not limited to this, and can also be a transitory storage medium.

[0609] The embodiments of the present disclosure further provide a program product, which, when executed by the communication device 7100, causes the communication device 7100 to perform any of the above methods. Alternatively, the program product can be a computer program product.

[0610] The embodiments of the present disclosure further provide a computer program, which, when executed on a computer, causes the computer to perform any of the above methods.

Claims

1. A method for monitoring model performance, characterized in that, The method, executed by a terminal device, includes: The system acquires first information, which includes second and / or third information; the second information is used to indicate the feedback performance of the codebook, and the third information includes fourth and fifth information, wherein the fourth information is the channel state information (CSI) at a first time step, and the fifth information is the CSI at the first time step recovered through the second part of the first model; the first model includes a first part model and a second part model, wherein the first part model is used to compress the CSI, and the second part model is used to recover the CSI based on the CSI compressed by the first part model. The performance of the first model is determined based on the first information.

2. The method according to claim 1, characterized in that, The first part of the model is deployed on the terminal device, and the second part of the model is deployed on the network device.

3. The method according to claim 2, characterized in that, Obtaining the fifth piece of information includes: Receive the fifth information sent by the network device.

4. The method according to claim 3, characterized in that, The fifth information received from the network device includes at least one of the following: Receive all the fifth information of the first moment sent by the network device at the same time; The network device receives the fifth information at each moment of the first time according to a first order, wherein the first order is agreed upon by the protocol or indicated by the network device.

5. The method according to claim 3 or 4, characterized in that, The method of sending the fifth information includes at least one of the following: floating-point quantization method and codebook method.

6. The method according to any one of claims 3-5, characterized in that, The method further includes: Obtain the sixth piece of information, which is the CSI at the second time point, which is before the first time point; Input the sixth piece of information into the first part of the model to obtain the seventh piece of information; The seventh information is sent to the network device, and the seventh information is used by the network device to obtain the fifth information.

7. The method according to claim 6, characterized in that, The acquisition of the sixth information includes: Receive the first resource sent by the network device at the second time; The sixth information is obtained based on the first resource measurement.

8. The method according to any one of claims 1-7, characterized in that, Obtaining fourth information includes at least one of the following: The fourth information was obtained through measurement; The fourth information is obtained by predicting using the second model, which is used to predict CSI. The eighth piece of information is obtained by measurement, and the fourth piece of information is predicted by the second model based on the eighth piece of information.

9. The method according to claim 8, characterized in that, The measurement yields the fourth information, including: Receive the second resource sent by the network device at the first moment; The fourth information is obtained based on the second resource measurement.

10. The method according to claim 8 or 9, characterized in that, The fourth information obtained through prediction by the second model includes: Obtain the ninth piece of information, which is the CSI at a historical moment; The tenth information is input into the second model to obtain the eleventh information, wherein the tenth information includes at least one historical CSI and / or at least one first-time CSI predicted by the second model; The fourth information is determined from the eleventh information based on the first indication information, wherein the first indication information is used to indicate the first moment.

11. The method according to any one of claims 8-10, characterized in that, The measurement yields the eighth piece of information, and based on the eighth piece of information, the fourth piece of information is predicted using the second model, including: Receive a third resource sent by a network device at a third time, wherein the third time includes the N times preceding the first time and / or historical times preceding the first time; The eighth information is obtained based on the third resource measurement; Input the eighth piece of information into the second model to obtain the twelfth piece of information; The fourth information is determined from the twelfth information based on the first indication information, wherein the first indication information is used to indicate the first moment.

12. The method according to any one of claims 1-11, characterized in that, The method further includes: Send the fourth information to the network device.

13. The method according to claim 12, characterized in that, Sending the fourth information to the network device includes at least one of the following: At the same time, all the fourth information from the first moment is sent to the network device; The fourth information at each moment in the first moment is sent to the network device in a second order, wherein the second order is agreed upon by the protocol or indicated by the terminal device.

14. The method according to claim 12 or 13, characterized in that, Sending the fourth information to the network device includes at least one of the following: The fourth information is sent to the network device using floating-point quantization. The fourth information is sent to the network device via a codebook method; The fourth information is compressed using the third part of the third model to obtain the thirteenth information, which is then sent to the network device. The third model includes the third part model and the fourth part model. The third part model is used to compress the CSI, and the fourth part model is used to restore the compressed CSI.

15. The method according to any one of claims 12-14, characterized in that, The method further includes: The terminal device receives a second instruction message sent by the network device, the second instruction message being used to instruct the terminal device to send the fourth message.

16. The method according to any one of claims 1-15, characterized in that, The terminal device is also equipped with a fourth model, which is used to restore the compressed CSI.

17. The method according to claim 16, characterized in that, Obtaining the fifth piece of information includes: The fifth information is obtained through the fourth model.

18. The method according to any one of claims 1-17, characterized in that, Determining the performance of the first model based on the first information includes: Based on the first information, a fourteenth piece of information is determined, which is used to indicate the performance of the first model; wherein the fourteenth piece of information includes at least one of the following: squared cosine similarity (SGCS), normalized mean square error (NMSE), block error rate (BLER), and throughput.

19. The method according to any one of claims 1-18, characterized in that, The method further includes: Send the performance data of the first model to the network device.

20. The method according to any one of claims 1-19, characterized in that, The first moment is determined in the following way: The agreement stipulates; Network device indication; The terminal device indicates.

21. A method for monitoring model performance, characterized in that, Performed by a network device, the method includes: The system acquires first information, which includes second and / or third information; the second information is used to indicate the feedback performance of the codebook, and the third information includes fourth and fifth information, wherein the fourth information is the channel state information (CSI) at a first time step, and the fifth information is the CSI at the first time step recovered through the second part of the first model; the first model includes a first part model and a second part model, wherein the first part model is used to compress the CSI, and the second part model is used to recover the CSI based on the CSI compressed by the first part model. The performance of the first model is determined based on the first information.

22. The method according to claim 21, characterized in that, The first part of the model is deployed on the terminal device, and the second part of the model is deployed on the network device.

23. The method according to claim 22, characterized in that, The method further includes: The terminal device receives a seventh piece of information, which is determined by the terminal device based on the sixth piece of information and the first part of the model. The sixth piece of information is the CSI at a second time, which is before the first time. The seventh information is input into the second part of the model to obtain the fifteenth information, which includes the recovered CSI at the fourth time point, and the fourth time point includes at least the first time point.

24. The method according to claim 23, characterized in that, The method further includes: Send a first resource to the terminal device, the first resource being used by the terminal device to measure and obtain the sixth information.

25. The method according to claim 23 or 24, characterized in that, The method further includes: According to the fifteenth information, the fifth information is sent to the terminal device.

26. The method according to claim 25, characterized in that, Sending the fifth information to the terminal device includes at least one of the following: At the same time, all the fifth information from the first moment is sent to the terminal device; The fifth information at each moment in the first time is sent to the terminal device in a first order, wherein the first order is agreed upon by the protocol or indicated by the network device.

27. The method according to claim 25 or 26, characterized in that, The method of sending the fifth information includes at least one of the following: floating-point quantization method and codebook method.

28. The method according to any one of claims 21-27, characterized in that, Obtaining the fourth information includes: The fourth information sent by the receiving terminal device.

29. The method according to claim 28, characterized in that, The receipt of the fourth information sent by the terminal device includes at least one of the following: Receive all the fourth information of the first moment sent by the terminal device at the same time; The terminal device receives the fourth information at each moment of the first moment in a second order, wherein the second order is agreed upon by the protocol or indicated by the terminal device.

30. The method according to claim 28 or 29, characterized in that, The fourth information sent by the receiving terminal device includes at least one of the following: Receive the fourth information sent by the terminal device in a floating-point quantization manner; Receive the fourth information sent by the terminal device via a codebook; The terminal device receives a thirteenth message, which is obtained by the terminal device through compression processing of the fourth message using the third part of the third model. The third model includes the third part model and the fourth part model. The third part model is used to compress the CSI, and the fourth part model is used to restore the compressed CSI.

31. The method according to claim 30, characterized in that, The method further includes: The thirteenth information is recovered by using the fourth part of the third model to obtain the fourth information.

32. The method according to any one of claims 28-31, characterized in that, The method further includes: At the first moment, a second resource is sent to the terminal device, the second resource being used by the terminal device to measure and obtain the fourth information.

33. The method according to any one of claims 28-32, characterized in that, The method further includes: A third resource is sent to the terminal device. The third resource is used by the terminal device to measure and obtain the eighth information, predict the twelfth information based on the eighth information, and determine the fourth information from the twelfth information based on the first indication information. The first indication information is used to indicate the first time.

34. The method according to any one of claims 28-33, characterized in that, The method further includes: Send a second instruction message to the terminal device, the second instruction message being used to instruct the terminal device to send the fourth message.

35. The method according to any one of claims 21-34, characterized in that, Determining the performance of the first model based on the first information includes: Based on the first information, a fourteenth piece of information is determined, which is used to indicate the performance of the first model; wherein the fourteenth piece of information includes at least one of the following: squared cosine similarity (SGCS), normalized mean square error (NMSE), block error rate (BLER), and throughput.

36. The method according to any one of claims 21-35, characterized in that, The method further includes: The performance of the first model is received from the terminal device.

37. The method according to any one of claims 21-36, characterized in that, The first moment is determined in the following way: The agreement stipulates; The network device indication; Terminal device indication.

38. A terminal device, characterized in that, include: The transceiver module is configured to acquire first information, which includes second information and / or third information; The second information is used to indicate the feedback performance of the codebook. The third information includes the fourth and fifth information. The fourth information is the channel state information (CSI) at the first moment, and the fifth information is the CSI at the first moment recovered by the second part of the model of the first model. The first model includes a first part model and a second part model. The first part model is used to compress CSI, and the second part model is used to recover CSI based on the CSI compressed by the first part model. The processing module is configured to determine the performance of the first model based on the first information.

39. A network device, characterized in that, include: The transceiver module is configured to acquire first information, which includes second information and / or third information; The second information is used to indicate the feedback performance of the codebook. The third information includes the fourth and fifth information. The fourth information is the channel state information (CSI) at the first moment, and the fifth information is the CSI at the first moment recovered by the second part of the model of the first model. The first model includes a first part model and a second part model. The first part model is used to compress CSI, and the second part model is used to recover CSI based on the CSI compressed by the first part model. The processing module is configured to determine the performance of the first model based on the first information.

40. A communication device, characterized in that, Its features include: One or more processors; The communication device is used to execute the model performance monitoring method according to any one of claims 1 to 20 or claims 21 to 37.

41. A storage medium storing instructions, characterized in that, When the instruction is executed on the communication device, the communication device performs the model performance monitoring method as described in any one of claims 1 to 20 or claims 21 to 37.

Images