Information transmission method and device, and storage medium

Abstract

The present disclosure provides an information transmission method and device, and a storage medium. The method comprises: determining feedback indication information; determining first information according to the feedback indication information and a target first AI model, the target first AI model being one of X1 first AI models included by a terminal; and sending the first information to a base station, the first information being used to determine input data of a target second AI model, the target second AI model being one of X2 second AI models included by the base station. The method reduces the storage overhead and maintenance complexity of the AI model.

Description

Technical Field

[0001] This disclosure relates to the field of communication technology, and in particular to an information transmission method, apparatus and storage medium. Background Technology

[0002] Wireless communication systems face numerous challenges, such as nonlinear problems, the time complexity of optimal solution computation, the difficulty in accurately characterizing some problems with formulas or models, the accumulation of errors between different modules in the wireless link preventing overall optimization, and the increased difficulty of optimization due to changes in non-ideal factors in practical applications. Currently, both academic research and 3GPP are exploring the use of Artificial Intelligence (AI) / Machine Learning (ML) to solve various problems in wireless communication systems. In NR Rel-17, the FS_NR_ENDC_data_collect project has already begun related research on AI / ML. In NR Rel-18, the potential of AI / ML at the air interface physical layer will be further investigated, such as improving performance indicators like throughput, accuracy, reliability, and robustness, and reducing resource overhead.

[0003] Currently, regardless of the actual application scenario, if AI is needed to solve problems in wireless communication systems, multiple corresponding AI models need to be deployed on the UE and gNB sides to meet the different information transmission needs between the UE and gNB, resulting in high storage overhead and maintenance complexity of the AI ​​models. Summary of the Invention

[0004] This disclosure provides an information transmission method, apparatus, and storage medium.

[0005] In a first aspect, embodiments of this disclosure provide an information transmission method applied to a terminal, the terminal including X1 first AI models, the method comprising:

[0006] Determine the feedback instruction information;

[0007] First information is determined based on the feedback indication information and the target first AI model, wherein the target first AI model is one of the X1 first AI models;

[0008] The first information is sent to the base station, and the first information is used to determine the input data of the target second AI model, wherein the target second AI model is one of the X2 second AI models included by the base station;

[0009] Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0010] In one implementation, determining the feedback indication information includes:

[0011] The feedback indication information is determined based on the first parameter.

[0012] In one implementation, the first parameter includes at least one of the following: Reference Signal Quality (RSRQ), Reference Signal Power (RSRP), Signal-to-Noise Ratio (SNR / SINR), Received Signal Strength Indicator (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Strategy (MCS).

[0013] In one implementation, determining the feedback indication information includes:

[0014] Obtain the second AI model;

[0015] The feedback indication information is determined based on the first AI model and the second AI model.

[0016] In one implementation, determining the feedback indication information includes:

[0017] The target first AI model is determined from the X1 first AI models;

[0018] The feedback indication information is determined based on the target first AI model.

[0019] In one implementation, determining the feedback indication information includes:

[0020] Receive the feedback indication information sent by the base station.

[0021] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0022] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0023] In one implementation, the number of bits of the output data of the target first AI model included in the first information corresponds to the feedback indication information.

[0024] In one implementation, X1 is greater than 1, and the number of bits in the output data of the X1 first AI models is different.

[0025] In one implementation, it further includes:

[0026] The feedback indication information is sent to the base station.

[0027] In one implementation, where X1 is greater than 1, after determining the feedback indication information, the method further includes:

[0028] The target first AI model is determined from the X1 first AI models based on the feedback indication information.

[0029] In one implementation, X2 is greater than 1, and the target second AI model corresponds to the feedback indication information.

[0030] Secondly, embodiments of this disclosure provide an information transmission method applied to a base station, wherein the base station includes X2 second AI models, and the method includes:

[0031] Determine the feedback instruction information;

[0032] The input data of the target second AI model is determined according to the feedback indication information and the first information sent by the terminal. The first information is determined by the terminal according to the feedback indication information and the target first AI model. The target second AI model is one of the X2 second AI models, and the target first AI model is one of the X1 first AI models in the terminal.

[0033] Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0034] In one embodiment, the method further includes:

[0035] The target second AI model is determined from the X2 second AI models based on the feedback indication information.

[0036] In one implementation, determining the feedback indication information includes:

[0037] The target second AI model is determined from the X2 second AI models;

[0038] The feedback indication information is determined based on the target second AI model.

[0039] In one implementation, determining the feedback indication information includes:

[0040] Receive the feedback indication information sent by the terminal.

[0041] In one implementation, it further includes:

[0042] The feedback instruction information is sent to the terminal.

[0043] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0044] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0045] In one implementation, the number of bits of the first information corresponds to the feedback indication information.

[0046] In one implementation, X2 is greater than 1, and the number of bits of input data for each of the X2 second AI models is different.

[0047] In one implementation, determining the input data for the target second AI model based on the feedback indication information and the first information sent by the terminal includes:

[0048] The number of bits of the first information is determined based on the feedback indication information;

[0049] The input data of the target second AI model is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information.

[0050] In one implementation, determining the input data of the target second AI model based on the number of bits of the input data of the target second AI model and the number of bits of the first information includes:

[0051] The number of bits M of the second information is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information, so as to determine the second information;

[0052] The input data for the target second AI model is determined based on the first information and the second information.

[0053] In one implementation, the second information is a bit string consisting of M zeros, where M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information.

[0054] Thirdly, this disclosure provides an information transmission device applied to a terminal, the terminal including X1 first AI models, and the device including a memory, a transceiver, and a processor:

[0055] The memory is used to store computer programs;

[0056] The transceiver is used to send and receive data under the control of the processor;

[0057] The processor is configured to read the computer program stored in the memory and perform the following operations:

[0058] Determine the feedback instruction information;

[0059] First information is determined based on the feedback indication information and the target first AI model, wherein the target first AI model is one of the X1 first AI models;

[0060] The first information is sent to the base station, and the first information is used to determine the input data of the target second AI model, wherein the target second AI model is one of the X2 second AI models included by the base station;

[0061] Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0062] In one implementation, the processor is configured to perform the following operations:

[0063] The feedback indication information is determined based on the first parameter.

[0064] In one implementation, the first parameter includes at least one of the following: Reference Signal Quality (RSRQ), Reference Signal Power (RSRP), Signal-to-Noise Ratio (SNR / SINR), Received Signal Strength Indicator (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Strategy (MCS).

[0065] In one implementation, the processor is configured to perform the following operations:

[0066] Obtain the second AI model;

[0067] The feedback indication information is determined based on the first AI model and the second AI model.

[0068] In one implementation, the processor is configured to perform the following operations:

[0069] The target first AI model is determined from the X1 first AI models;

[0070] The feedback indication information is determined based on the target first AI model.

[0071] In one implementation, the processor is configured to perform the following operations:

[0072] Receive the feedback indication information sent by the base station.

[0073] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0074] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0075] In one implementation, the number of bits of the output data of the target first AI model included in the first information corresponds to the feedback indication information.

[0076] In one implementation, X1 is greater than 1, and the number of bits in the output data of the X1 first AI models is different.

[0077] In one implementation, the processor is configured to perform the following operations:

[0078] The feedback indication information is sent to the base station.

[0079] In one implementation, where X1 is greater than 1, the processor is configured to perform the following operations:

[0080] The target first AI model is determined from the X1 first AI models based on the feedback indication information.

[0081] In one implementation, X2 is greater than 1, and the target second AI model corresponds to the feedback indication information.

[0082] Fourthly, this disclosure provides an information transmission device applied to a base station, wherein the base station includes X2 second AI models, and the device includes a memory, a transceiver, and a processor.

[0083] The memory is used to store computer programs;

[0084] The transceiver is used to send and receive data under the control of the processor;

[0085] The processor is configured to read the computer program stored in the memory and perform the following operations:

[0086] Determine the feedback instruction information;

[0087] The input data of the target second AI model is determined according to the feedback indication information and the first information sent by the terminal. The first information is determined by the terminal according to the feedback indication information and the target first AI model. The target second AI model is one of the X2 second AI models, and the target first AI model is one of the X1 first AI models in the terminal.

[0088] Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0089] In one implementation, the processor is configured to perform the following operations:

[0090] The target second AI model is determined from the X2 second AI models based on the feedback indication information.

[0091] In one implementation, the processor is configured to perform the following operations:

[0092] The target second AI model is determined from the X2 second AI models;

[0093] The feedback indication information is determined based on the target second AI model.

[0094] In one implementation, the processor is configured to perform the following operations:

[0095] Receive the feedback indication information sent by the terminal.

[0096] In one implementation, the processor is configured to perform the following operations:

[0097] The feedback instruction information is sent to the terminal.

[0098] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0099] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0100] In one implementation, the number of bits of the first information corresponds to the feedback indication information.

[0101] In one implementation, X2 is greater than 1, and the number of bits of input data for each of the X2 second AI models is different.

[0102] In one implementation, the processor is configured to perform the following operations:

[0103] The number of bits of the first information is determined based on the feedback indication information;

[0104] The input data of the target second AI model is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information.

[0105] In one implementation, the processor is configured to perform the following operations:

[0106] The number of bits M of the second information is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information, so as to determine the second information;

[0107] The input data for the target second AI model is determined based on the first information and the second information.

[0108] In one implementation, the second information is a bit string consisting of M zeros, where M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information.

[0109] Fifthly, this disclosure provides an information transmission device applied to a terminal, the terminal including X1 first AI models, the device including:

[0110] The first determining unit is used to determine feedback indication information;

[0111] The second determining unit is used to determine first information based on the feedback indication information and the target first AI model, wherein the target first AI model is one of the X1 first AI models;

[0112] A sending unit is configured to send the first information to a base station, the first information being used to determine the input data of a target second AI model, wherein the target second AI model is one of the X2 second AI models included in the base station;

[0113] Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0114] In one implementation, the first determining unit is configured to:

[0115] The feedback indication information is determined based on the first parameter.

[0116] In one implementation, the first parameter includes at least one of the following: Reference Signal Quality (RSRQ), Reference Signal Power (RSRP), Signal-to-Noise Ratio (SNR / SINR), Received Signal Strength Indicator (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Strategy (MCS).

[0117] In one implementation, the first determining unit is configured to:

[0118] Obtain the second AI model;

[0119] The feedback indication information is determined based on the first AI model and the second AI model.

[0120] In one implementation, the first determining unit is configured to:

[0121] The target first AI model is determined from the X1 first AI models;

[0122] The feedback indication information is determined based on the target first AI model.

[0123] In one implementation, the first determining unit is configured to:

[0124] Receive the feedback indication information sent by the base station.

[0125] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0126] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0127] In one implementation, the number of bits of the output data of the target first AI model included in the first information corresponds to the feedback indication information.

[0128] In one implementation, X1 is greater than 1, and the number of bits in the output data of the X1 first AI models is different.

[0129] In one embodiment, the transmitting unit is further configured to:

[0130] The feedback indication information is sent to the base station.

[0131] In one embodiment, where X1 is greater than 1, the device further includes:

[0132] The third determining unit is used to determine the target first AI model from the X1 first AI models based on the feedback indication information.

[0133] In one implementation, X2 is greater than 1, and the target second AI model corresponds to the feedback indication information.

[0134] Sixthly, this disclosure provides an information transmission device applied to a base station, wherein the base station includes X2 second AI models, and the device includes:

[0135] The first determining unit is used to determine feedback indication information;

[0136] The second determining unit is used to determine the input data of the target second AI model based on the feedback indication information and the first information sent by the terminal. The first information is determined by the terminal based on the feedback indication information and the target first AI model. The target second AI model is one of the X2 second AI models, and the target first AI model is one of the X1 first AI models in the terminal.

[0137] Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0138] In one embodiment, the apparatus further includes: a third determining unit, configured to:

[0139] The target second AI model is determined from the X2 second AI models based on the feedback indication information.

[0140] In one implementation, the first determining unit is configured to:

[0141] The target second AI model is determined from the X2 second AI models;

[0142] The feedback indication information is determined based on the target second AI model.

[0143] In one implementation, the first determining unit is configured to:

[0144] Receive the feedback indication information sent by the terminal.

[0145] In one embodiment, the device further includes:

[0146] A sending unit is used to send the feedback indication information to the terminal.

[0147] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0148] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0149] In one implementation, the number of bits of the first information corresponds to the feedback indication information.

[0150] In one implementation, X2 is greater than 1, and the number of bits of input data for each of the X2 second AI models is different.

[0151] In one embodiment, the second determining unit is used to:

[0152] The number of bits of the first information is determined based on the feedback indication information;

[0153] The input data of the target second AI model is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information.

[0154] In one embodiment, the second determining unit is used to:

[0155] The number of bits M of the second information is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information, so as to determine the second information;

[0156] The input data for the target second AI model is determined based on the first information and the second information.

[0157] In one implementation, the second information is a bit string consisting of M zeros, where M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information.

[0158] In a seventh aspect, this disclosure provides a computer-readable storage medium storing a computer program for causing a computer to perform the method described in the first aspect.

[0159] This disclosure provides an information transmission method, apparatus, and storage medium. In this method, a terminal determines the information to be transmitted to a base station based on feedback indication information and an AI model on the terminal side. The base station determines the input data of an AI model on the base station side based on the feedback indication information and the information sent by the terminal. The AI ​​models on the terminal side and the base station side do not need to maintain a one-to-one correspondence, which can reduce the number of AI models deployed on the terminal side or the base station side, and reduce the storage overhead and maintenance complexity of AI models.

[0160] It should be understood that the description in the foregoing summary section is not intended to limit the key or essential features of the embodiments of the present invention, nor is it intended to restrict the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0161] To more clearly illustrate the technical solutions in this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0162] Figure 1 This is a schematic diagram of a CSI feedback scenario based on an AI model.

[0163] Figure 2 A flowchart illustrating the information transmission method provided in this embodiment of the disclosure;

[0164] Figure 3 This is a schematic diagram of an AI model-based CSI feedback scenario provided in an embodiment of this disclosure. Figure 1 ;

[0165] Figure 4 This is a schematic diagram of an AI model-based CSI feedback scenario provided in an embodiment of this disclosure. Figure 2 ;

[0166] Figure 5 This disclosure provides a schematic diagram of model training and application in the embodiments of the present disclosure. Figure 1 ;

[0167] Figure 6 This is a schematic diagram of an AI model-based CSI feedback scenario provided in an embodiment of this disclosure. Figure 3 ;

[0168] Figure 7 This disclosure provides a schematic diagram of model training and application in the embodiments of the present disclosure. Figure 2 ;

[0169] Figure 8 Schematic diagram of the structure of the information transmission device provided in the embodiments of this disclosure Figure 1 ;

[0170] Figure 9 Schematic diagram of the channel transmission device provided in the embodiments of this disclosure Figure 2 ;

[0171] Figure 10 Schematic diagram of the structure of the information transmission device provided in the embodiments of this disclosure Figure 3 ;

[0172] Figure 11 Schematic diagram of the structure of the information transmission device provided in the embodiments of this disclosure Figure 4 . Detailed Implementation

[0173] In this disclosure, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. In the embodiments of this disclosure, the term "multiple" refers to two or more, and other quantifiers are similar.

[0174] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this disclosure.

[0175] This disclosure provides an information transmission method and apparatus, which reduces model storage overhead and maintenance complexity. The method and apparatus are based on the same concept, and since the principles by which they solve problems are similar, embodiments of the apparatus and method can be referred to interchangeably; repeated details will not be repeated.

[0176] The technical solutions provided in this disclosure are applicable to a variety of systems, especially 5G systems. For example, applicable systems may include Global System for Mobile Communication (GSM), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA) General Packet Radio Service (GPRS), Long Term Evolution (LTE), LTE Frequency Division Duplex (FDD), LTE Time Division Duplex (TDD), Long Term Evolution Advanced (LTE-A), Universal Mobile Telecommunication System (UMTS), Worldwide Interoperability for Microwave Access (WiMAX), and 5G New Radio (NR). All of these systems include terminal equipment and network equipment. The systems may also include a core network component, such as Evolved Packet System (EPS) and 5G systems (5GS).

[0177] The terminal devices involved in the embodiments of this disclosure can be devices that provide voice and / or data connectivity to users, handheld devices with wireless connectivity, or other processing devices connected to a wireless modem. The names of the terminal devices may differ in different systems; for example, in a 5G system, a terminal device can be called User Equipment (UE). Wireless terminal devices can communicate with one or more core networks (CNs) via a Radio Access Network (RAN). Wireless terminal devices can be mobile terminal devices, such as mobile phones (or "cellular" phones) and computers with mobile terminal devices, for example, portable, pocket-sized, handheld, computer-embedded, or vehicle-mounted mobile devices that exchange voice and / or data with the RAN. Examples include Personal Communication Service (PCS) phones, cordless phones, Session Initiated Protocol (SIP) phones, Wireless Local Loop (WLL) stations, and Personal Digital Assistants (PDAs). Wireless terminal equipment can also be referred to as a system, subscriber unit, subscriber station, mobile station, mobile station, remote station, access point, remote terminal, access terminal, user terminal, user agent, or user device, but is not limited to these terms in the embodiments disclosed herein.

[0178] The base stations involved in this disclosure may include multiple cells providing services to terminals. Depending on the specific application, a base station may also be called an access point, or a device in an access network that communicates with wireless terminal devices via one or more sectors on the air interface, or other names. For example, the base stations involved in this disclosure may be evolved network devices (eNBs or e-NodeBs) in long term evolution (LTE) systems, 5G base stations (gNBs) in next generation systems, or home evolved Node Bs (HeNBs), relay nodes, femtos, picos, etc., and are not limited in this disclosure. In some network structures, a base station may include centralized unit (CU) nodes and distributed unit (DU) nodes, and the centralized and distributed units may be geographically separated.

[0179] Base stations and terminal devices can each use one or more antennas for multiple-input multiple-output (MIMO) transmission. MIMO transmission can be single-user MIMO (SU-MIMO) or multiple-user MIMO (MU-MIMO). Depending on the configuration and number of antenna combinations, MIMO transmission can be 2D-MIMO, 3D-MIMO, FD-MIMO, or massive-MIMO, and can also be diversity transmission, precoding transmission, or beamforming transmission, etc.

[0180] The following section introduces the application of AI models in wireless communication systems. The NR Rel-18 physical layer AI research primarily identified several use cases, such as Channel State Information (CSI) feedback, beam management, localization, and channel estimation. Taking CSI feedback as an example, the UE obtains CSI information by measuring the Channel State Information Reference Signal (CSI-RS) and feeds it back to the gNB. Utilizing AI models for CSI feedback can reduce CSI feedback overhead or improve channel recovery accuracy. For different feedback granularities of CSI feedback, multiple corresponding AI models need to be deployed on both the UE and gNB sides. In this case, channel recovery accuracy is high, but the storage overhead and maintenance complexity of the AI ​​models are relatively high from both the UE and gNB perspectives.

[0181] For example, such as Figure 1 The diagram illustrates a CSI feedback scenario based on an AI model. The AI ​​models in the diagram are exemplified by an encoder and a decoder, which can be considered as two separate AI models or two components of a single AI model. The UE can perform channel estimation by measuring CSI-RS and input this channel estimate into the AI ​​model CSI encoder to obtain the required CSI information or a portion of the CSI information (CSI information can include multiple components such as CQI, PMI, and RI), which is then quantized into binary bit data. The UE then feeds back the quantized CSI information to the gNB. Upon receiving the quantized binary bit data, the gNB dequantizes it back to the real number domain and inputs it into the AI ​​model CSI decoder to obtain the recovered channel estimate. Since CSI feedback has different feedback granularities—specifically, different numbers of bits quantized for CSI information—multiple corresponding AI models need to be deployed on both the UE and gNB sides. Each corresponding AI model corresponds to a specific feedback granularity, thus meeting the transmission requirements of various feedback granularities. Thus, from both the UE and gNB perspectives, the storage overhead and maintenance complexity of the AI ​​model are relatively high.

[0182] Similar to the CSI feedback described above, in other application scenarios, the output data of the terminal-side AI model will be used as the input data of the base station-side AI model. However, there are also multiple information transmission requirements when the UE and gNB transmit information. This requires the deployment of multiple one-to-one corresponding AI models on the UE and gNB sides. Each corresponding AI model satisfies a certain information transmission requirement, resulting in high storage overhead and maintenance complexity of the AI ​​model.

[0183] To address this, this disclosure proposes an information transmission method in which the UE determines the information to be transmitted to the gNB based on feedback indication information and an AI model on the UE side, and the gNB determines the input data of an AI model on the gNB side based on the feedback indication information and the information sent by the UE. The AI ​​models on the UE side and the gNB side do not need to maintain a one-to-one correspondence, which can reduce the number of AI models deployed on the UE side or the gNB side, and reduce the storage overhead and maintenance complexity of the AI ​​models.

[0184] The information transmission method provided in this disclosure will now be described in detail through specific embodiments. It is understood that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0185] Figure 2 This is a flowchart illustrating the information transmission method provided in an embodiment of this disclosure. Figure 2 As shown, the method includes:

[0186] S201, The terminal confirms the feedback instruction information.

[0187] Feedback indication information can be used to indicate the information transmission requirements when the terminal sends feedback information to the base station, so that the terminal can determine the information that needs to be sent back to the base station based on the feedback indication information. The feedback indication information can be determined by the terminal itself, or it can be determined by the base station. In the latter case, this step includes the terminal receiving the feedback indication information sent by the base station.

[0188] S202, The terminal determines the first information based on the feedback instruction information and the target first AI model, wherein the target first AI model is one of the X1 first AI models in the terminal.

[0189] The terminal determines the first information to be sent to the base station based on the feedback instruction information and the target first AI model in the terminal. The first information may include some or all of the output data of the target first AI model, or the first information may include data after processing some or all of the output data of the target first AI model.

[0190] S203. The terminal sends first information to the base station. The first information is used to determine the input data of the target second AI model. The target second AI model is one of the X2 second AI models included by the base station.

[0191] Where X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0192] It is understandable that, when the terminal determines the feedback indication information on its own, the terminal also sends the feedback indication information to the base station. Optionally, the terminal can send the first information and the feedback indication information separately through different resources.

[0193] S204, Base station confirms feedback indication information.

[0194] When the feedback indication information is determined by the terminal itself, this step refers to the base station receiving the feedback indication information sent by the terminal. Alternatively, when the feedback indication information is determined by the base station, this step is performed before S201, that is, the base station first determines the feedback indication information and then sends the feedback indication information to the terminal.

[0195] S205. The base station determines the input data for the target's second AI model based on the feedback instruction information and the first information sent by the terminal.

[0196] After determining the first information, the terminal sends it to the base station. The base station, based on the feedback indication information and the first information, determines the input data for the target second AI model on its side. The target first AI model is one of X1 first AI models in the terminal, and the target second AI model is one of X2 second AI models included in the base station. However, the target first AI model and the target second AI model do not need to be deployed in a one-to-one correspondence; there is no necessary correspondence between them. The first information transmitted between the terminal and the base station is determined by the terminal based on the feedback indication information and the target first AI model in the terminal, while the base station can determine the input data for the target second AI model on its side based on the feedback indication information and the first information. This allows for flexible AI model deployment on both the terminal and base station sides, reducing the number of AI models, lowering AI model storage overhead, and reducing maintenance complexity.

[0197] The terminal involved in this embodiment includes X1 first AI models, and the base station includes X2 second AI models. Optionally, the first AI models can be used to generate information that the terminal needs to transmit to the base station. For example, the first AI models can be used to encode input data to generate information that needs to be transmitted to the base station, while the second AI models can be used to process the information transmitted by the terminal to the base station, such as decoding the information transmitted by the terminal to the base station.

[0198] For example, in Figure 1 In the scenario of CSI feedback based on the AI ​​model shown, the input of the first AI model is the channel estimate, and the output of the first AI model is CSI information or a portion of CSI information (CSI information may include multiple components such as CQI, PMI, and RI). Therefore, the first information is CSI information or a portion of CSI information. The input of the second AI model is CSI information or a portion of CSI information, and the output of the second AI model is the recovered channel estimate.

[0199] Based on the above embodiments, further explanation is provided on how to determine feedback indication information.

[0200] Optionally, the terminal determines feedback indication information based on a first parameter. Optionally, the first parameter includes at least one of the following: Reference Signal Receiving Quality (RSRQ), Reference Signal Receiving Power (RSRP), Signal-to-Noise Ratio (SNR) or Signal-to-Interference Plus Noise Ratio (SINR), Received Signal Strength Indication (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Scheme (MCS). The terminal determines the feedback indication information based on the above parameters, and then determines the first information to be fed back to the base station based on the feedback indication information. This allows the first information to be fed back to the base station to better match the current channel state, signal quality, etc., thereby improving the accuracy of data recovery on the base station side.

[0201] Optionally, the terminal acquires a second AI model and determines feedback indication information based on the first and second AI models. The first AI model may be part or all of the X1 first AI models included in the terminal, and the second AI model may be part or all of the X2 second AI models included in the base station. The terminal can perform multiple joint inferences on the first and second AI models to determine the model with the highest accuracy, and then determine the corresponding feedback indication information based on the model with the highest accuracy. By performing joint inferences based on the first and second AI models and ultimately determining the feedback indication information based on the model with the highest accuracy, the terminal can determine the first information that needs to be fed back to the base station, thereby maximizing the accuracy of the base station's data recovery.

[0202] Optionally, the terminal determines a target first AI model from X1 first AI models, and determines feedback indication information based on the target first AI model. That is, there is a correspondence between the first AI models and the feedback indication information. The terminal first determines the target first AI model, and then determines the feedback indication information based on the correspondence between the first AI models and the feedback indication information. It is understood that the terminal may also first determine the feedback indication information using the aforementioned method or receive the feedback indication information sent by the base station, and then determine the target first AI model from X1 first AI models based on the feedback indication information.

[0203] Optionally, the feedback indication information can be determined by the base station, that is, the base station determines the target second AI model from X2 second AI models; and determines the feedback indication information based on the target second AI model. In other words, there is a correspondence between the second AI models and the feedback indication information; the base station first determines the target second AI model, and then determines the feedback indication information based on the correspondence between the second AI models and the feedback indication information.

[0204] Optionally, the feedback indication information may be at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0205] The feedback indication information characterizes the number of bits, application scenario, or encoding method of the first information fed back by the terminal to the base station. The first AI model or the second AI model can correspond to the aforementioned number of bits, application scenario, or encoding method. When determining the feedback indication information, the terminal and the base station can directly determine the information related to the first information, the first AI model, or the second AI model through the feedback indication information, without directly indicating the number of bits, application scenario, or encoding method of the first information. Specifically, when the feedback indication information is determined by the terminal, it can be the model level, model identifier, parameters, or target first AI model of the target first AI model. When the feedback indication information is determined by the base station, it can be the model level, model identifier, parameters, or target second AI model of the target second AI model. When the feedback indication information is different, the target first AI model can be a different model among the X1 first AI models included by the terminal. When the feedback indication information is different, the target second AI model can be a different model among the X2 second AI models included by the base station.

[0206] The first information includes the output data of the target first AI model. Optionally, the number of bits of the output data of the target first AI model included in the first information corresponds to the feedback indication information. That is, the number of bits of the first information determined by the terminal is different under different feedback indication information, and the number of bits of the first information corresponds to the feedback indication information.

[0207] Optionally, if the number of first AI models X1 included in the terminal is greater than 1, the number of bits of the output data of the X1 first AI models is different. Thus, when the terminal determines the target first AI model corresponding to the feedback indication information, it determines the number of bits of the first information.

[0208] Optionally, the target second AI model corresponds to the feedback indication information. That is, the base station determines the target second AI model from X2 second AI models according to the feedback indication information, and then determines the input data of the target second AI model based on the first information.

[0209] Optionally, when X2 is greater than 1, the number of bits in the input data of each of the X2 second AI models is different. When the base station determines the input data of the target second AI model based on the feedback indication information and the first information, the base station determines the number of bits in the first information based on the feedback indication information; and determines the input data of the target second AI model based on the number of bits in the input data of the target second AI model and the number of bits in the first information.

[0210] Optionally, the base station determines the number of bits M of the second information based on the number of bits in the input data of the target second AI model and the number of bits in the first information, thereby determining the second information; it then determines the input data of the target second AI model based on the first and second information, wherein the second information is a bit string consisting of M zeros, and M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information. Optionally, the second information can also be a bit string consisting of M ones, or several other real numbers quantized into a bit string of length M, where each bit in the bit string of length M is either 0 or 1.

[0211] Based on the above embodiments, the information transmission method of this disclosure will be described in conjunction with specific examples.

[0212] Example 1

[0213] Taking X1 as 1 and X2 as greater than 1, meaning that the terminal includes one first AI model, which is also the target first AI model, and the base station includes multiple second AI models as an example, this will be explained.

[0214] Step 1: The terminal confirms the feedback instruction information.

[0215] The terminal can determine the feedback indication information either by determining it itself or by receiving feedback indication information sent by the base station.

[0216] Optionally, the terminal can determine the feedback indication information itself, as mentioned above, by determining the feedback indication information based on the first parameter. The classification of feedback indication information is related to the threshold of the first parameter. For example, as shown in Table 1, there are X2 types of feedback indication information. The threshold of the first parameter corresponding to each type of feedback indication information can be specified by the protocol or configured by the base station through system messages or Radio Resource Control (RRC) messages. Of course, the types of feedback indication information may not be equal to X2.

[0217] Table 1

[0218] First parameter < threshold 1 Feedback Instruction Message 1 Threshold 1 < First parameter < Threshold 2 Feedback Instruction Message 2 … … Threshold X2-1 < First parameter Feedback Instruction Information X2

[0219] Optionally, the terminal can determine the feedback indication information based on the first AI model and the second AI model. For example, the terminal can obtain part or all of the second AI models from the base station after acquiring the first AI model, such as by downloading the second AI models from the base station. Assume the terminal acquires all of X2 second AI models, denoted as second AI model_1, second AI model_2, ..., second AI model_X2. The terminal performs joint inference using the first AI model and each of the multiple second AI models. The input data of the first AI model is the same in different joint inferences; for example, the input data of the first AI model is CSI-RS channel estimation. The first information determined by the first AI model is different in different joint inferences, as shown in Table 2 below. There is a one-to-one correspondence between the second AI models and the feedback indication information. The terminal determines the model with the highest accuracy based on multiple joint inferences and determines the feedback indication information corresponding to that model.

[0220] Table 2

[0221]

[0222]

[0223] For example, if the model with the highest accuracy is the first AI model + the second AI model_2, then the terminal determines the feedback indication information as feedback indication information 2. The terminal can determine the feedback indication information through the second AI model in the model with the highest accuracy, and the base station can also determine the target second AI model through the feedback indication information.

[0224] Optionally, the terminal receives feedback indication information from the base station via higher-layer signaling, or the terminal receives feedback indication information from the base station via downlink control information (DCI) or media access control element (MAC CE).

[0225] Step 2: The terminal determines the first information based on the feedback instruction information and the first AI model, and sends the first information to the base station.

[0226] like Figure 3 As shown, the first AI model is illustrated using a CSI encoder, with only one CSI encoder model on the terminal side. The second AI model is illustrated using a CSI decoder, with multiple CSI decoder models on the base station side. The terminal determines the number of bits, application scenario, or encoding method of the output data of the first AI model included in the first information based on the feedback indication information. Figure 3 Taking the example of different numbers of bits in the first information corresponding to different feedback indication information, assuming that the output of the first AI model includes N information blocks or bit strings, the terminal determines that the first information includes one or more of the N information blocks based on the feedback indication information. For example, for feedback indication information 1, the first information includes the first bit string in the N information blocks; for feedback indication information 2, the first information includes the first and second bit strings in the N information blocks, and so on. The terminal sends the first information determined according to the feedback level and the first AI model to the base station.

[0227] When the feedback indication information is determined by the terminal, the terminal will also send the feedback indication information to the base station, such as... Figure 4As shown, the first information and feedback indication information can be transmitted through different resources. Optionally, the terminal receives first configuration information from the base station for the first resource used to carry the feedback indication information, and the terminal determines the first resource to carry the feedback indication information based on the first configuration information. The terminal transmits the feedback indication information on the first resource, which can be a Physical Uplink Control Channel (PUCCH) or a Physical Uplink Shared Channel (PUSCH). Alternatively, feedback level information can be carried through signals on the first resource, such as a Demodulation Reference Signal (DMRS), a Sounding Reference Signal (SRS), or a Physical Random Access Channel (PRACH) on the first resource. Different feedback levels correspond to different signal patterns or different generation sequences.

[0228] Furthermore, it should be noted that when the feedback indication information is the model level of the first AI model, the model identifier of the first AI model, the parameters of the first AI model, or the first AI model itself, this step involves the terminal determining the first information based on the feedback indication information and sending the first information to the base station; or, this step involves the terminal determining the first information based on the first AI model and sending the first information to the base station. Alternatively, when the feedback indication information is the model level of the target first AI model, the model identifier of the target first AI model, the parameters of the target first AI model, or the target first AI model itself, this step involves the terminal determining the first information based on the feedback indication information and sending the first information to the base station; or, this step involves the terminal determining the first information based on the target first AI model and sending the first information to the base station.

[0229] Step 3: The base station determines the input data for the target's second AI model based on the feedback indication information and the first information.

[0230] Optionally, when the feedback indication information is determined by the terminal, the base station receives the feedback indication information sent by the terminal and determines the target second AI model among the X2 second AI models based on the feedback indication information. For example, the feedback indication information is feedback indication information 2, the target second AI model is second AI model_2, and the base station uses the first information as the input data of the second AI model_2.

[0231] Optionally, if the feedback indication information is determined by the base station, the base station determines the feedback indication information and the corresponding target second AI model before step 1, and sends the feedback indication information to the terminal. After receiving the first information, the base station determines the input data of the target second AI model based on the first information.

[0232] Optionally, the number of bits in the input data of different second AI models can be different, and there can be a one-to-one correspondence between the feedback indication information and the second AI model. Optionally, the number of bits in the input data of different second AI models can be the same. In this case, for different feedback indication information, the number of bits in the first information output by the first AI model on the terminal side is the same, while the second AI models corresponding to different feedback indication information are applied to different application scenarios or encoding methods, etc. For example, multiple second AI models can be applied to scenarios such as Ultra-reliable and Low Latency Communications (URLLC), Reduced Capability (REDCAP), and Enhanced Mobile Broadband (eMBB), respectively.

[0233] Combination Figure 5 The training and application process of the AI ​​model in this embodiment will be explained.

[0234] During model training, for each piece of training data in the training dataset, the forward propagation process of the neural network is as follows:

[0235] Step 1: Obtain the output layer or output data of length P {x1,x2,…,xP} from the first AI model.

[0236] Step 2: Based on the feedback indication information and the output data {x1,x2,…,xP} of length P, obtain the first information {x1,x2,…,xQ} of length Q, where Q is less than or equal to P.

[0237] Step 3: Determine the i-th second AI model as the target second AI model based on the feedback instructions.

[0238] Step 4: Use the first information of length Q as the input data of the i-th second AI model, and obtain the output of the i-th second AI model based on the input data.

[0239] The backpropagation process is as follows:

[0240] Step 5: Determine the error of the training data based on the output data of the i-th second AI model and the labels of the training data.

[0241] Step 6: Update the weights of the i-th second AI model based on the error in Step 5 and calculate the input layer error of the i-th second AI model.

[0242] Step 7: PQ zeros are added to the input layer error of the i-th second AI model, and the error of the output layer of the first AI model is determined.

[0243] Step 8: Update the weights of the first AI model based on the error of the first AI model's output layer.

[0244] As in step 2 above, the output data {x1,x2,…,xP} of length P can include N information blocks, and the first information {x1,x2,…,xQ} of length Q is composed of one or more of the N information blocks. Each of the N information blocks can be composed of data output by consecutive neurons in the output layer of the first AI model, such as {x1,x2,x3,x4,…}; or each of the N information blocks can be composed of data output by non-consecutive neurons in the output layer of the first AI model, such as {x1,x3,x5,x7,…}.

[0245] As can be seen from the training process, each piece of training data in the training dataset is input into the first AI model, but different training data will be input into different second AI models according to the feedback instructions.

[0246] During model application, the forward propagation process for each data point is as follows:

[0247] Step 1: Obtain the output layer or output data of length P {x1,x2,…,xP} from the first AI model.

[0248] Step 2: Based on the feedback indication information and the output data {x1,x2,…,xP} of length P, obtain the first information {x1,x2,…,xQ} of length Q, where Q is less than or equal to P.

[0249] Step 3: Determine the i-th second AI model as the target second AI model based on the feedback instructions.

[0250] Step 4: Use the first information of length Q as the input data of the i-th second AI model, and obtain the output of the i-th second AI model based on the input data.

[0251] As can be seen from the information transmission method in this embodiment, only one AI model needs to be deployed on the terminal side to meet the different information transmission requirements between the terminal and the base station, thereby reducing the storage overhead and maintenance complexity of the AI ​​model.

[0252] Example 2

[0253] The following example illustrates the concept of X2 being greater than 1 or X2 being 1, meaning that the terminal includes multiple first AI models and the base station includes one second AI model, which is also the target second AI model.

[0254] Step 1: The terminal confirms the feedback instruction information.

[0255] The terminal can determine the feedback indication information either by determining it itself or by receiving feedback indication information sent by the base station.

[0256] Similar to the previous examples, optionally, the terminal can determine the feedback indication information itself by determining the feedback indication information based on the first parameter, and the division of the feedback indication information is related to the threshold of the first parameter.

[0257] Optionally, the terminal can determine the feedback indication information based on the first AI model and the second AI model. For example, the terminal can obtain the second AI model from the base station based on the first AI model, such as by downloading the second AI model from the base station. Assume the terminal has X1 first AI models, denoted as first AI model_1, first AI model_2, ..., first AI model_X1. The terminal performs joint inference using each of the multiple first AI models and the second AI model. The input data of the first AI models is the same in different joint inferences; for example, the input data of the first AI model is CSI-RS channel estimation. The first information determined by the first AI model is different in different joint inferences, as shown in Table 3 below. There is a one-to-one correspondence between the first AI models and the feedback indication information. The terminal determines the model with the highest accuracy based on multiple joint inferences and determines the feedback indication information corresponding to that model.

[0258] Table 3

[0259]

[0260]

[0261] For example, if the model with the highest accuracy is the first AI model_2 + the second AI model, then the terminal determines the feedback indication information as feedback indication information 2. The terminal can determine the feedback indication information through the first AI model in the model with the highest accuracy, and the base station can also determine the target second AI model through the feedback indication information.

[0262] Optionally, the terminal receives feedback indication information from the base station via higher-layer signaling, or the terminal receives feedback indication information from the base station via DCI or MAC CE.

[0263] Step 2: The terminal determines the first information based on the feedback instruction information and the target first AI model, and sends the first information to the base station.

[0264] like Figure 6 As shown, the first AI model is illustrated using a CSI encoder, and there are multiple CSI encoder models on the terminal side. The second AI model is illustrated using a CSI decoder, and there is only one CSI decoder model on the base station side. The terminal determines the target first AI model among the multiple first AI models based on feedback indication information and outputs first information. For example, the terminal determines the target first AI model as the first AI model_1 among the multiple first AI models based on feedback level 1. When the feedback indication information is determined by the terminal, the terminal also sends the feedback indication information to the base station.

[0265] Optionally, the terminal may first determine the target first AI model among X1 first AI models, and then determine the feedback indication information based on the target first AI model. For example, the terminal determines that the target first AI model is the first AI model_1 among multiple first AI models, and determines the feedback indication information 1 based on the target first AI model.

[0266] Step 3: The base station determines the input data for the second AI model based on the feedback indication information and the first information.

[0267] Assume the input data of the second AI model contains A bits, and the first information contains B bits, where A is greater than or equal to B. When A is greater than B, the input data of the second AI model consists of the first information and the second information, with the second information consisting of M zeros, where M = AB. That is, the base station determines the length of the first information to be B bits based on the feedback indication information, and adds M zeros to the first information of length B to form the input data of the second AI model.

[0268] Optionally, different first AI models can output different numbers of bits for the first information. In this case, there can be a one-to-one correspondence between the feedback level and the first AI model. Optionally, different first AI models can output the same number of bits for the first information. In this case, the different feedback indications corresponding to different first AI models are model feedback indications applied to different application scenarios. For example, multiple first AI models can be applied to URLLC, REDCAP, and eMBB scenarios respectively.

[0269] Combination Figure 7 The training and application process of the AI ​​model in this embodiment will be explained.

[0270] During model training, for each piece of training data in the training dataset, the forward propagation process of the neural network is as follows:

[0271] Step 1: Determine the i-th first AI model as the target first AI model based on the feedback indication information.

[0272] Step 2: Obtain the first information {x1,x2,…,xP1} of length P1 based on the i-th first AI model.

[0273] Step 3: Based on the feedback indication information and the first information {x1,x2,…,xP1} of length P1, obtain the input data {x1,x2,…,xQ1} of the second AI model of length Q1, where Q1 is greater than or equal to P1.

[0274] Step 4: Use {x1,x2,…,xQ1} of length Q1 as input data for the second AI model, and obtain the output of the second AI model based on the input data.

[0275] The backpropagation process is as follows:

[0276] Step 5: Determine the error of the training data based on the output data of the second AI model and the labels of the training data.

[0277] Step 6: Update the weights of the second AI model based on the error in Step 5 and calculate the input layer error of the second AI model.

[0278] Step 7: Extract a portion of length P1 from the input layer error of the second AI model to determine the output layer error of the first AI model.

[0279] Step 8: Update the weights of the first AI model based on the error of the first AI model's output layer.

[0280] As in step 3 above, the input data {x1,x2,…,xQ1} of the second AI model with length Q1 is obtained through feedback indication information and first information {x1,x2,…,xP1} of length P1. The position of the first information in the input data of the second AI model can be continuous or discontinuous.

[0281] As can be seen from the training process, each piece of training data in the training dataset is input into the second AI model, but different training data will be input into different first AI models according to the feedback instructions.

[0282] In the model application process, for each piece of data to be applied, the forward propagation process is as follows:

[0283] Step 1: Determine the i-th first AI model as the target first AI model based on the feedback indication information.

[0284] Step 2: Obtain the first information {x1,x2,…,xP1} of length P1 based on the i-th first AI model.

[0285] Step 3: Based on the feedback indication information and the first information {x1,x2,…,xP1} of length P1, obtain the input data {x1,x2,…,xQ1} of the second AI model of length Q1, where Q1 is greater than or equal to P1.

[0286] Step 4: Use {x1,x2,…,xQ1} of length Q1 as input data for the second AI model, and obtain the output of the second AI model based on the input data.

[0287] As can be seen from the information transmission method in this embodiment, only one AI model needs to be deployed on the base station side to meet the different information transmission needs between the terminal and the base station, thereby reducing the storage overhead and maintenance complexity of the AI ​​model.

[0288] Figure 8 Schematic diagram of the structure of the information transmission device provided in the embodiments of this disclosure Figure 1 .like Figure 8 As shown, the device includes a memory 801, a transceiver 802, and a processor 803.

[0289] Memory 801 is used to store computer programs;

[0290] Transceiver 802 is used to send and receive data under the control of processor 803;

[0291] Processor 803 is used to read the computer program stored in memory 801 and perform the following operations:

[0292] Determine the feedback instruction information;

[0293] The first information is determined based on the feedback instructions and the target first AI model, where the target first AI model is one of X1 first AI models;

[0294] Send first information to the base station. The first information is used to determine the input data of the target second AI model. The target second AI model is one of the X2 second AI models included by the base station.

[0295] Where X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0296] In one implementation, processor 803 is configured to perform the following operations:

[0297] The feedback indication information is determined based on the first parameter.

[0298] In one implementation, the first parameter includes at least one of the following: Reference Signal Quality (RSRQ), Reference Signal Power (RSRP), Signal-to-Noise Ratio (SNR / SINR), Received Signal Strength Indicator (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Strategy (MCS).

[0299] In one implementation, processor 803 is configured to perform the following operations:

[0300] Obtain the second AI model;

[0301] The feedback instruction information is determined based on the first AI model and the second AI model.

[0302] In one implementation, processor 803 is configured to perform the following operations:

[0303] Determine the target first AI model from X1 first AI models.

[0304] Based on the target AI model, determine the feedback instruction information.

[0305] In one implementation, processor 803 is configured to perform the following operations:

[0306] Receive feedback indication information sent by the base station.

[0307] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0308] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0309] In one implementation, the number of bits of the output data of the target first AI model included in the first information corresponds to the feedback indication information.

[0310] In one implementation, X1 is greater than 1, and the number of bits in the output data of the X1 first AI models is different.

[0311] In one implementation, processor 803 is configured to perform the following operations:

[0312] Send feedback instruction information to the base station.

[0313] In one implementation, X1 is greater than 1, and the processor 803 is configured to perform the following operations:

[0314] The target first AI model is determined from X1 first AI models based on the feedback instructions.

[0315] In one implementation, X2 is greater than 1, and the target second AI model corresponds to the feedback indication information.

[0316] It should be noted that the device provided in this disclosure can implement all the method steps implemented by the terminal in the above method embodiment and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiment will not be described in detail here.

[0317] Figure 9 Schematic diagram of the structure of the information transmission device provided in the embodiments of this disclosure Figure 2 .like Figure 9 As shown, the device includes a memory 901, a transceiver 902, and a processor 903.

[0318] Memory 901 is used to store computer programs;

[0319] Transceiver 902 is used to send and receive data under the control of processor 903;

[0320] Processor 903 is used to read the computer program stored in memory 901 and perform the following operations:

[0321] Determine the feedback instruction information;

[0322] The input data of the target second AI model is determined based on the feedback instruction information and the first information sent by the terminal. The first information is determined by the terminal based on the feedback instruction information and the target first AI model. The target second AI model is one of X2 second AI models, and the target first AI model is one of X1 first AI models in the terminal.

[0323] Where X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0324] In one implementation, processor 903 is configured to perform the following operations:

[0325] The target second AI model is determined from X2 second AI models based on the feedback instructions.

[0326] In one implementation, processor 903 is configured to perform the following operations:

[0327] Determine the target second AI model from X2 second AI models;

[0328] The feedback instruction information is determined based on the target second AI model.

[0329] In one implementation, processor 903 is configured to perform the following operations:

[0330] Receive feedback instructions sent by the terminal.

[0331] In one implementation, processor 903 is configured to perform the following operations:

[0332] Send feedback instructions to the terminal.

[0333] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0334] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0335] In one implementation, the number of bits of the first information corresponds to the feedback indication information.

[0336] In one implementation, X2 is greater than 1, and the number of bits of input data for each of the X2 second AI models is different.

[0337] In one implementation, processor 903 is configured to perform the following operations:

[0338] Determine the number of bits in the first information based on the feedback instruction information;

[0339] The input data of the target second AI model is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information.

[0340] In one implementation, processor 903 is configured to perform the following operations:

[0341] The number of bits M of the second information is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information, so as to determine the second information;

[0342] The input data for the target second AI model is determined based on the first and second information.

[0343] In one implementation, the second information is a bit string consisting of M zeros, where M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information.

[0344] It should be noted that the apparatus provided in this disclosure can implement all the method steps implemented by the base station in the above method embodiments and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiments will not be described in detail here.

[0345] Figure 10 Schematic diagram of the channel transmission device provided in the embodiments of this disclosure Figure 3 .like Figure 10 As shown, the device includes:

[0346] The first determining unit 1001 is used to determine feedback indication information;

[0347] The second determining unit 1002 is used to determine first information based on feedback instruction information and target first AI model, wherein the target first AI model is one of X1 first AI models;

[0348] The sending unit 1003 is used to send first information to the base station. The first information is used to determine the input data of the target second AI model. The target second AI model is one of the X2 second AI models included by the base station.

[0349] Where X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0350] In one embodiment, the first determining unit 1001 is configured to:

[0351] The feedback indication information is determined based on the first parameter.

[0352] In one implementation, the first parameter includes at least one of the following: Reference Signal Quality (RSRQ), Reference Signal Power (RSRP), Signal-to-Noise Ratio (SNR / SINR), Received Signal Strength Indicator (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Strategy (MCS).

[0353] In one embodiment, the first determining unit 1001 is configured to:

[0354] Obtain the second AI model;

[0355] The feedback instruction information is determined based on the first AI model and the second AI model.

[0356] In one embodiment, the first determining unit 1001 is configured to:

[0357] Determine the target first AI model from X1 first AI models.

[0358] Based on the target AI model, determine the feedback instruction information.

[0359] In one embodiment, the first determining unit 1001 is configured to:

[0360] Receive feedback indication information sent by the base station.

[0361] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0362] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0363] In one implementation, the number of bits of the output data of the target first AI model included in the first information corresponds to the feedback indication information.

[0364] In one implementation, X1 is greater than 1, and the number of bits in the output data of the X1 first AI models is different.

[0365] In one embodiment, the transmitting unit 1003 is further configured to:

[0366] Send feedback instruction information to the base station.

[0367] In one embodiment, X1 is greater than 1, and the device further includes:

[0368] The third determining unit is used to determine the target first AI model from X1 first AI models based on feedback indication information.

[0369] In one implementation, X2 is greater than 1, and the target second AI model corresponds to the feedback indication information.

[0370] It should be noted that the device provided in this disclosure can implement all the method steps implemented by the terminal in the above method embodiment and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiment will not be described in detail here.

[0371] Figure 11 Schematic diagram of the channel transmission device provided in the embodiments of this disclosure Figure 4 .like Figure 11 As shown, the device includes:

[0372] The first determining unit 1101 is used to determine feedback indication information;

[0373] The second determining unit 1102 is used to determine the input data of the target second AI model according to the feedback instruction information and the first information sent by the terminal. The first information is determined by the terminal according to the feedback instruction information and the target first AI model. The target second AI model is one of X2 second AI models, and the target first AI model is one of X1 first AI models in the terminal.

[0374] Where X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

[0375] In one embodiment, the apparatus further includes: a third determining unit, configured to:

[0376] The target second AI model is determined from X2 second AI models based on the feedback instructions.

[0377] In one embodiment, the first determining unit 1101 is configured to:

[0378] Determine the target second AI model from X2 second AI models;

[0379] The feedback instruction information is determined based on the target second AI model.

[0380] In one embodiment, the first determining unit 1101 is configured to:

[0381] Receive feedback instructions sent by the terminal.

[0382] In one embodiment, the apparatus further includes:

[0383] The sending unit is used to send feedback indication information to the terminal.

[0384] In one implementation, the feedback indication information is at least one of the following: the number of bits of the first information, the level corresponding to the number of bits of the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

[0385] In one implementation, the first information includes part or all of the output data of the target first AI model.

[0386] In one implementation, the number of bits of the first information corresponds to the feedback indication information.

[0387] In one implementation, X2 is greater than 1, and the number of bits of input data for each of the X2 second AI models is different.

[0388] In one embodiment, the second determining unit 1102 is used to:

[0389] Determine the number of bits in the first information based on the feedback instruction information;

[0390] The input data of the target second AI model is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information.

[0391] In one embodiment, the second determining unit 1102 is used to:

[0392] The number of bits M of the second information is determined based on the number of bits of the input data of the target second AI model and the number of bits of the first information, so as to determine the second information;

[0393] The input data for the target second AI model is determined based on the first and second information.

[0394] In one implementation, the second information is a bit string consisting of M zeros, where M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information.

[0395] It should be noted that the apparatus provided in this disclosure can implement all the method steps implemented by the base station in the above method embodiments and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiments will not be described in detail here.

[0396] It should be noted that the division of units in the embodiments of this disclosure is illustrative and only represents one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units.

[0397] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0398] This disclosure also provides a computer-readable storage medium storing a computer program for causing a computer to perform the methods executed by a terminal or base station in the above method embodiments.

[0399] Computer-readable storage media can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs), etc.), optical storage (such as CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (such as ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0400] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the method executed by a terminal or base station in the above method embodiments.

[0401] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0402] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-executable instructions. These computer-executable instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0403] These processor-executable instructions may also be stored in a processor-readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the processor-readable memory produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0404] These processors can execute instructions that can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0405] Obviously, those skilled in the art can make various modifications and variations to this disclosure without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims of this disclosure and their equivalents, this disclosure is also intended to include such modifications and variations.

Claims

1. An information transmission method applied to a terminal, characterized in that, The terminal includes X1 first AI models, and the method includes: Determine the feedback instruction information; First information is determined based on the feedback indication information and the target first AI model, wherein the target first AI model is one of the X1 first AI models; The first information is sent to the base station. This first information is used to determine the input data of the target second AI model. The input data of the target second AI model is determined by the second information and the first information. The second information is determined based on the number of bits A of the input data of the target second AI model and the number of bits B of the first information. When A is greater than B, the number of bits M of the second information is the difference between the number of bits of the input data of the target second AI model and the number of bits of the first information. The input data of the target second AI model is composed of the first information and the second information. The target second AI model is one of X2 second AI models included in the base station. Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

2. The method of claim 1, wherein, The determined feedback indication information includes: The feedback indication information is determined based on the first parameter.

3. The method of claim 2, wherein, The first parameter includes at least one of the following: Reference Signal Quality (RSRQ), Reference Signal Power (RSRP), Signal-to-Noise Ratio (SNR / SINR), Received Signal Strength Indication (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Strategy (MCS).

4. The method according to claim 1, characterized in that, The determined feedback indication information includes: Obtain the second AI model; The feedback indication information is determined based on the first AI model and the second AI model.

5. The method according to claim 1, characterized in that, The determined feedback indication information includes: The target first AI model is determined from the X1 first AI models; The feedback indication information is determined based on the target first AI model.

6. The method according to claim 1, characterized in that, The determined feedback indication information includes: Receive the feedback indication information sent by the base station.

7. The method according to any one of claims 1-6, characterized in that, The feedback indication information includes at least one of the following: the number of bits in the first information, the level corresponding to the number of bits in the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

8. The method according to any one of claims 1-6, characterized in that, The first information includes part or all of the output data of the target first AI model.

9. The method according to claim 8, characterized in that, The number of bits of the output data of the target first AI model included in the first information corresponds to the feedback indication information.

10. The method according to claim 8, characterized in that, X1 is greater than 1, and the number of bits in the output data of the X1 first AI models is different.

11. The method according to any one of claims 1-5, characterized in that, Also includes: The feedback indication information is sent to the base station.

12. The method according to any one of claims 1-6, characterized in that, Where X1 is greater than 1, after determining the feedback indication information, the method further includes: The target first AI model is determined from the X1 first AI models based on the feedback indication information.

13. The method according to any one of claims 1-6, characterized in that, Where X2 is greater than 1, the target second AI model corresponds to the feedback indication information.

14. An information transmission method applied to a base station, characterized in that, The base station includes X2 second AI models, and the method includes: Determine the feedback instruction information; The number of bits of the first information is determined based on the feedback indication information; The number of bits M of the second information is determined based on the number of bits A of the input data of the target second AI model and the number of bits B of the first information, so as to determine the second information; The input data of the target second AI model is determined based on the second information and the first information sent by the terminal. When A is greater than B, M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information. The input data of the target second AI model is composed of the first information and the second information. The first information is determined by the terminal based on the feedback indication information and the target first AI model. The first information is used to determine the input data of the target second AI model. The target second AI model is one of the X2 second AI models, and the target first AI model is one of the X1 first AI models in the terminal. Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

15. The method according to claim 14, characterized in that, The method further includes: The target second AI model is determined from the X2 second AI models based on the feedback indication information.

16. The method according to claim 14, characterized in that, The determined feedback indication information includes: The target second AI model is determined from the X2 second AI models; The feedback indication information is determined based on the target second AI model.

17. The method according to claim 14, characterized in that, The determined feedback indication information includes: Receive the feedback indication information sent by the terminal.

18. The method according to any one of claims 14-16, characterized in that, Also includes: The feedback instruction information is sent to the terminal.

19. The method according to any one of claims 14-17, characterized in that, The feedback indication information includes at least one of the following: the number of bits in the first information, the level corresponding to the number of bits in the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

20. The method according to any one of claims 14-17, characterized in that, The first information includes part or all of the output data of the target first AI model.

21. The method according to claim 20, characterized in that, The number of bits in the first information corresponds to the feedback indication information.

22. The method according to any one of claims 14-17, characterized in that, X2 is greater than 1, and the number of bits in the input data of each of the X2 second AI models is different.

23. The method according to claim 14, characterized in that, The second information is a bit string consisting of M zeros.

24. An information transmission device, applied to a terminal, characterized in that, The terminal includes X1 first AI models, and the device includes a memory, a transceiver, and a processor. The memory is used to store computer programs; The transceiver is used to send and receive data under the control of the processor; The processor is configured to read the computer program stored in the memory and perform the following operations: Determine the feedback instruction information; First information is determined based on the feedback indication information and the target first AI model, wherein the target first AI model is one of the X1 first AI models; The first information is sent to the base station. This first information is used to determine the input data of the target second AI model. The input data of the target second AI model is determined by the second information and the first information. The second information is determined based on the number of bits A of the input data of the target second AI model and the number of bits B of the first information. When A is greater than B, the number of bits M of the second information is the difference between the number of bits of the input data of the target second AI model and the number of bits of the first information. The input data of the target second AI model is composed of the first information and the second information. The target second AI model is one of X2 second AI models included in the base station. Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

25. The apparatus according to claim 24, characterized in that, The processor is used to perform the following operations: The feedback indication information is determined based on the first parameter.

26. The apparatus according to claim 25, characterized in that, The first parameter includes at least one of the following: Reference Signal Quality (RSRQ), Reference Signal Power (RSRP), Signal-to-Noise Ratio (SNR / SINR), Received Signal Strength Indication (RSSI), Bit Error Rate (BER), Block Error Rate (BLER), and Modulation and Coding Strategy (MCS).

27. The apparatus according to claim 25, characterized in that, The processor is used to perform the following operations: Obtain the second AI model; The feedback indication information is determined based on the first AI model and the second AI model.

28. The apparatus according to claim 25, characterized in that, The processor is used to perform the following operations: The target first AI model is determined from the X1 first AI models; The feedback indication information is determined based on the target first AI model.

29. The apparatus according to claim 25, characterized in that, The processor is used to perform the following operations: Receive the feedback indication information sent by the base station.

30. The apparatus according to any one of claims 24-29, characterized in that, The feedback indication information includes at least one of the following: the number of bits in the first information, the level corresponding to the number of bits in the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

31. An information transmission device applied to a base station, characterized in that, The base station includes X2 second AI models, and the device includes a memory, a transceiver, and a processor. The memory is used to store computer programs; The transceiver is used to send and receive data under the control of the processor; The processor is configured to read the computer program stored in the memory and perform the following operations: Determine the feedback instruction information; The number of bits of the first information is determined based on the feedback indication information; The number of bits M of the second information is determined based on the number of bits A of the input data of the target second AI model and the number of bits B of the first information, so as to determine the second information; The input data of the target second AI model is determined based on the second information and the first information sent by the terminal. When A is greater than B, M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information. The input data of the target second AI model is composed of the first information and the second information. The first information is determined by the terminal based on the feedback indication information and the target first AI model. The first information is used to determine the input data of the target second AI model. The target second AI model is one of the X2 second AI models, and the target first AI model is one of the X1 first AI models in the terminal. Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

32. The apparatus according to claim 31, characterized in that, The processor is used to perform the following operations: The target second AI model is determined from the X2 second AI models based on the feedback indication information.

33. The apparatus according to claim 31, characterized in that, The processor is used to perform the following operations: The target second AI model is determined from the X2 second AI models; The feedback indication information is determined based on the target second AI model.

34. The apparatus according to claim 31, characterized in that, The processor is used to perform the following operations: Receive the feedback indication information sent by the terminal.

35. The apparatus according to any one of claims 31-34, characterized in that, The feedback indication information includes at least one of the following: the number of bits in the first information, the level corresponding to the number of bits in the first information, the application scenario corresponding to the first information or the target first AI model, the identifier of the application scenario corresponding to the first information or the target first AI model, the encoding method corresponding to the first information or the target first AI model, the identifier of the encoding method corresponding to the first information or the target first AI model, the model level of the target first AI model or the target second AI model, the model identifier of the target first AI model or the target second AI model, the parameters of the target first AI model or the target second AI model.

36. An information transmission device, applied to a terminal, characterized in that, The terminal includes X1 first AI models, and the device includes: The first determining unit is used to determine feedback indication information; The second determining unit is used to determine first information based on the feedback indication information and the target first AI model, wherein the target first AI model is one of the X1 first AI models; A transmitting unit is configured to transmit the first information to a base station. The first information is used to determine the input data of a target second AI model. The input data of the target second AI model is determined by the second information and the first information. The second information is determined based on the number of bits A of the input data of the target second AI model and the number of bits B of the first information. When A is greater than B, the number of bits M of the second information is the difference between the number of bits of the input data of the target second AI model and the number of bits of the first information. The input data of the target second AI model is composed of the first information and the second information. The target second AI model is one of the X2 second AI models included in the base station. Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

37. An information transmission device applied to a base station, characterized in that, The base station includes X2 second AI models, and the device includes: The first determining unit is used to determine feedback indication information; The second determining unit is used to determine the number of bits of the first information based on the feedback indication information; The number of bits M of the second information is determined based on the number of bits A of the input data of the target second AI model and the number of bits B of the first information, so as to determine the second information; The input data of the target second AI model is determined based on the second information and the first information sent by the terminal. When A is greater than B, M is the difference between the number of bits in the input data of the target second AI model and the number of bits in the first information. The input data of the target second AI model is composed of the first information and the second information. The first information is determined by the terminal based on the feedback indication information and the target first AI model. The first information is used to determine the input data of the target second AI model. The target second AI model is one of the X2 second AI models, and the target first AI model is one of the X1 first AI models in the terminal. Wherein, X1 is greater than or equal to 1, X2 is greater than or equal to 1, and X1 is not equal to X2.

38. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that causes a computer to perform the method as described in any one of claims 1-23.

Images