Communication method and communication device

Abstract

The invention provides a communication method and a communication device applied to the field of wireless communication. In the technical scheme provided by the invention, the receiving end can perform matched training on the AI sequence and the AI detection model, and indicates the generation mode of the AI sequence to the sending end after the AI sequence training is completed, so that the sending end can generate the AI sequence, and the receiving end does not need to transmit all AI sequences generated by training, thereby reducing the signaling overhead. Wherein the AI sequence sent by the sending end and the AI detection model of the receiving end are trained in a matched manner, so that the detection performance of the receiving end on the AI sequence can be improved.

Description

Technical Field

[0001] This application relates to the field of wireless communication, and more particularly to a communication method and a communication device. Background Technology

[0002] In the field of communications, reference signals can be generated based on reference sequences, enabling the two communicating parties to agree on the transmitted signals. This allows the receiving end to detect the received signals based on the agreed-upon signals after receiving them from the sending end, achieving purposes such as access, synchronization, demodulation, measurement, and estimation.

[0003] However, this method suffers from low detection performance. Summary of the Invention

[0004] This application provides a communication method and a communication device, applicable to the field of wireless communication. The technical solution provided by this application allows network devices to instruct terminals to autonomously generate artificial intelligence (AI) sequences, thereby improving detection performance while reducing signaling overhead.

[0005] In a first aspect, embodiments of this application provide a communication method that can be applied to the transmitting end of an AI sequence. Taking the transmitting end as a terminal as an example, the method can be applied to the terminal side, such as the terminal or the communication module in the terminal, or the circuit or chip responsible for communication functions in the terminal (such as a modem chip, also known as a baseband chip, or a system-on-chip (SoC) chip or system-in-package (SIP) chip containing a modem module). Taking the application of this method to a terminal as an example, the method includes: receiving first information, the first information being used to indicate a vector, the vector satisfying a first relationship with the identifier of an artificial intelligence (AI) sequence, the vector being used as input to an AI model, the AI ​​model being used to generate the AI ​​sequence; and transmitting a first AI sequence, the first AI sequence being generated based on the first information.

[0006] In this technical solution, the network device can instruct the terminal on the input of the AI ​​model, such as a vector. The terminal can input the vector into the AI ​​model to generate an AI sequence, and then send the generated AI sequence to the network device. The AI ​​model can reside within the terminal, or in other words, the terminal can invoke the AI ​​model. The first AI sequence can be understood as the AI ​​sequence determined and sent by the terminal. The AI ​​model can be predefined through a protocol.

[0007] In this technical solution, the AI ​​sequences on the terminal side and the AI ​​detection model on the network device side can be trained together by the network device. Therefore, when the terminal sends AI sequences to the network device for purposes such as access, synchronization, demodulation, measurement, and estimation, the network device achieves high detection accuracy for the AI ​​sequences sent by the terminal. Furthermore, by instructing the terminal on the input to the AI ​​model, the terminal can autonomously generate AI sequences based on the model. This eliminates the need for the network device to send all trained AI sequences to the terminal, reducing signaling overhead. It should be understood that the output of the AI ​​model can include some or all of the AI ​​sequences in the set of AI sequences generated by the network device during training.

[0008] In conjunction with the first aspect, in some implementations of the first aspect, the first information is used to indicate a vector, including: the first information is used to indicate the first relationship.

[0009] In this implementation, the indicator vector can be indicated by a first relationship between the indicator vector and the identifier of the AI ​​sequence, so that the terminal can map the identifier of the first AI sequence to a first vector based on the first relationship, and input the first vector into the AI ​​model to generate the first AI sequence.

[0010] In conjunction with the first aspect, in some implementations of the first aspect, the first information includes at least one of the following: a first formula satisfied between the identifier of the AI ​​sequence and the vector, parameters in the first formula, the length of the vector, or the processing method of the vector.

[0011] As an example, the first relationship can be a parameterized mapping. For instance, the identifiers and vectors of an AI sequence can satisfy a first formula, which can contain parameters. Optionally, the network device can simultaneously indicate the first formula and the parameters within it to improve communication efficiency. Alternatively, the first formula can be predefined by a protocol or standard, and the parameters within the first formula can be indicated by the network device to enhance the flexibility of using the first formula.

[0012] As an example, the first relationship can be a parameterless mapping. For instance, the identifiers and vectors of an AI sequence can satisfy multi-base conversion. The specific conversion mechanism can be predefined by the protocol or configured by the network device, and is not limited here.

[0013] Optionally, the network device can indicate the length of the vector to the terminal to ensure that the vectors input to the AI ​​model have a uniform length, thereby improving the performance of the AI ​​model. The length of a vector can also be referred to as the dimension of the vector.

[0014] Optionally, the network device can also indicate to the terminal how the vector is processed, such as whether to perform normalization, to improve the performance of the AI ​​model. In some embodiments, the network device can also indicate to the terminal the normalization interval.

[0015] In conjunction with the first aspect, in some implementations of the first aspect, the first information is used to indicate a vector, including: the first information is used to indicate the vector set in which the vector is located, multiple vectors in the vector set correspond one-to-one with multiple AI sequences, each of the multiple vectors satisfies the first relationship with the identifier of the corresponding AI sequence, and the multiple AI sequences belong to a subset of the sequence set in which the AI ​​sequence is located.

[0016] In this implementation, the network device can indicate vectors through a set of indicator vectors. Each vector in the vector set corresponds one-to-one with a sequence in the set of AI sequences generated by the network device during training. For example, when the first vector in the vector set corresponds to the first AI sequence in the AI ​​sequence set, a first relation can be satisfied between the identifiers of the first vector and the first AI sequence. It should be understood that each vector in the vector set can be used as input to the first AI model.

[0017] In this implementation, the network device can directly send the mapped vector set to the terminal without indicating the first relation. Since the data size of the vector set can be smaller than that of the AI ​​sequence set, transmitting the vector set can reduce signaling overhead compared to transmitting the AI ​​sequence set.

[0018] In conjunction with the first aspect, in some implementations of the first aspect, the first information includes at least one of the following: the plurality of vectors, or the processing method of each of the plurality of vectors.

[0019] As an example, a vector set can be represented as [M, k]. Here, M represents the number of vectors in the vector set, and k represents the length of each vector. The number of vectors in the vector set can be the same as the number of AI sequences in the AI ​​sequence set.

[0020] Optionally, the length of the vector can be less than the length of the AI ​​sequence, so that the data volume of the vector set is less than the data volume of the AI ​​sequence set. Transmitting the vector set can reduce transmission overhead compared to transmitting the AI ​​sequence set.

[0021] Optionally, the network device can also indicate to the terminal how each vector in the vector set should be processed, such as whether to normalize the vectors to improve the performance of the AI ​​model. In some embodiments, the network device can also indicate to the terminal the normalization interval.

[0022] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: receiving second information, the second information being used to instruct the AI ​​model.

[0023] In this implementation, the network device can also instruct the AI ​​model to the terminal.

[0024] Optionally, when the structure of the AI ​​model is predefined by a protocol or standard, or defined offline by network devices and terminals, the second information may only indicate the parameters of the AI ​​model, or the second information may indicate both the parameters of the AI ​​model and the identifier of the structure of the AI ​​model, in order to reduce the overhead and complexity of AI model configuration.

[0025] Optionally, when the structure and parameters of the AI ​​model are predefined by a protocol or standard, or are defined offline by network devices and terminals, the second information may only indicate the identifier of the AI ​​model.

[0026] Optionally, when the structure and / or parameters of the AI ​​model are configured by the network device, the second information can indicate the structure and / or parameters of the AI ​​model to improve the flexibility of AI model configuration.

[0027] Optionally, the AI ​​model structure can be a model structure supported by the Z4 mode with a fixed model structure in model transfer.

[0028] In some embodiments, the AI ​​model may be trained by a network device.

[0029] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: sending third information, the third information being used to indicate at least one of the following: supported AI model information, supported first relation, or the duration required to generate the AI ​​sequence.

[0030] In this implementation, the terminal can report its capabilities to the network device, such as whether it supports the ability to generate AI sequences through AI models. When the terminal supports the ability to generate AI sequences through AI models, it can report the AI ​​model information it supports (such as model structure and / or model parameters), and / or the first relation it supports, so that the network device can be aware of the terminal's capabilities and improve communication performance.

[0031] In some embodiments, when the terminal supports the ability to generate AI sequences through an AI model, the terminal can report the time required for the terminal to generate the AI ​​sequence to the network device, so that the network device can know the terminal's capabilities and improve communication performance.

[0032] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: receiving sixth information, the sixth information being used to indicate the identifier of the first AI sequence.

[0033] In this implementation, the network device can indicate the identifier of the AI ​​sequence to the terminal, enabling the terminal to determine the vector input into the AI ​​model based on the identifier indicated by the network device, and thus generate the AI ​​sequence corresponding to the identifier. This implementation, where the network device indicates the identifier of the AI ​​sequence, reduces interference between AI sequences generated by different terminals compared to the terminal randomly selecting the identifier.

[0034] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: performing any of the following operations within a first time period: sending a first sequence, or not sending a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0035] In this implementation, considering that the AI ​​sequence generation process will take a certain amount of time, or that there is a delay in the AI ​​sequence generation process, the behavior of the terminal during this period can be agreed upon, so that the terminal knows what behavior can be adopted during this period of time, thereby improving communication performance.

[0036] Secondly, embodiments of this application provide a communication method that can be applied to the transmitting end of an AI sequence. Taking the transmitting end as a terminal as an example, the method can be applied to the terminal side, such as the terminal or the communication module in the terminal, or the circuit or chip responsible for communication functions in the terminal (such as a modem chip, also known as a baseband chip, or a system-on-chip (SoC) chip or system-in-package (SIP) chip containing a modem module). Taking the application of this method to a terminal as an example, the method includes: receiving first information, the first information being used to indicate a plurality of sub-sequence sets, the plurality of sub-sequence sets satisfying a second relationship with an AI sequence set; transmitting a first AI sequence, the first AI sequence being generated based on the first information, the first AI sequence being included in the AI ​​sequence set; wherein, the product of the number of sub-AI sequences included in each of the plurality of sub-sequence sets is equal to the number of AI sequences included in the AI ​​sequence set, and the sum of the number of sub-AI sequences included in each of the sub-sequence sets is less than the number of AI sequences included in the AI ​​sequence set.

[0037] In this technical solution, after training and generating an AI sequence set, the network device can divide the AI ​​sequence set into multiple sub-sequence sets according to the hierarchical mapping method to be adopted, and send them to the terminal. Alternatively, the network device can directly train multiple sub-sequence sets during the AI ​​sequence training process, based on the hierarchical mapping method to be adopted, and send them to the terminal. It should be noted that no restrictions are placed on the hierarchical mapping method used or its specific implementation.

[0038] Taking the quotient-remainder mapping method as an example, suppose the AI ​​sequence set is represented as M×N, meaning the AI ​​sequence set includes M AI sequences, and each AI sequence has a length of N. If the number of subsequence sets is 2, then the two subsequence sets can be represented as M1×N and M2×N respectively. Here, M1×M2=M, and M1+M2<M, thus allowing the total data size of the subsequence sets to be less than the data size of the AI ​​sequence set, reducing transmission overhead. Here, M, N, M1, and M2 are all positive integers.

[0039] Optionally, the fourth piece of information can be predefined through the protocol.

[0040] In this technical solution, the terminal can determine the AI ​​sequence set based on multiple sub-sequence sets and a second relationship, and then randomly select an AI sequence from the AI ​​sequence set to send to the network device. Since the total data volume of the sub-sequence set is less than the data volume of the AI ​​sequence set, the transmission overhead when the network device sends the sub-sequence set to the terminal is low.

[0041] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: receiving fourth information, the fourth information being used to indicate the second relationship.

[0042] In this implementation, the second relationship can be configured by the network device for the terminal. For example, the network device can determine the second relationship based on real-time network conditions, load, etc., and send it to the terminal to improve communication performance.

[0043] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: receiving fifth information, the fifth information being used to indicate a third relationship, the third relationship being the relationship between the identifier of the AI ​​sequence and multiple sub-identifiers, the multiple sub-identifiers corresponding to a set of multiple sub-sequences, each of the multiple sub-identifiers being used to determine a sub-AI sequence in the corresponding set of sub-sequences, and the AI ​​sequence being generated based on the multiple sub-AI sequences determined by the multiple sub-identifiers.

[0044] In this implementation, a sub-identifier can be understood as the identifier of a sub-AI sequence within a set of sub-sequences. The terminal can map the identifier of an AI sequence to multiple sub-identifiers based on a third relation, and determine the corresponding sub-AI sequence from the set of sub-sequences based on these multiple sub-identifiers, thus determining the AI ​​sequence. The AI ​​sequence can be the product of multiple sub-AI sequences.

[0045] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: receiving sixth information, the sixth information being used to indicate the identifier of the first AI sequence.

[0046] In this implementation, the network device can indicate the identifier of the AI ​​sequence to the terminal, enabling the terminal to determine the vector input into the AI ​​model based on the identifier indicated by the network device, and thus generate the AI ​​sequence corresponding to the identifier. This implementation, where the network device indicates the identifier of the AI ​​sequence, reduces interference between AI sequences generated by different terminals compared to the terminal randomly selecting the identifier.

[0047] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: performing any of the following operations during a first time period: sending a first sequence, or not sending a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0048] In this implementation, considering that the AI ​​sequence generation process will take a certain amount of time, or that there is a delay in the AI ​​sequence generation process, the behavior of the terminal during this period can be agreed upon, so that the terminal knows what behavior can be adopted during this period of time, thereby improving communication performance.

[0049] Thirdly, embodiments of this application provide a communication method that can be applied to either a detection end or a receiving end of an AI sequence. Taking a network device as an example, the method can be applied to the network side, such as an access network device, a module (e.g., a circuit, chip, or chip system) within the access network device, or a logical node, logical module, or software capable of implementing all or part of the functions of the access network device. Taking the application of this method to a network device (such as an access network device) as an example, the method includes: sending first information, wherein the first information is used to indicate a vector, the vector and the identifier of the AI ​​sequence satisfy a first relationship, the vector is used as input to an AI model, and the AI ​​model is used to generate the AI ​​sequence; and receiving a first AI sequence, wherein the first AI sequence is generated based on the first information.

[0050] In this technical solution, the network device can instruct the terminal on the input of the AI ​​model, such as a vector. The terminal can then input the vector into the AI ​​model to generate an AI sequence, which is then sent to the network device. The AI ​​model can reside within the terminal, or in other words, the terminal can invoke the AI ​​model. The first AI sequence can be understood as the AI ​​sequence determined and sent by the terminal. The AI ​​model can be predefined through a protocol.

[0051] In this technical solution, the AI ​​sequences on the terminal side and the AI ​​detection model on the network device side can be trained together by the network device. Therefore, when the terminal sends AI sequences to the network device for purposes such as access, synchronization, demodulation, measurement, and estimation, the network device achieves high detection accuracy for the AI ​​sequences sent by the terminal. Furthermore, by instructing the terminal on the input to the AI ​​model, the terminal can autonomously generate AI sequences based on the model. This eliminates the need for the network device to send all trained AI sequences to the terminal, reducing signaling overhead. It should be understood that the output of the AI ​​model can include some or all of the AI ​​sequences in the set of AI sequences generated by the network device during training.

[0052] In conjunction with the third aspect, in some implementations of the third aspect, the first information is used to indicate a vector, including: the first information is used to indicate the first relationship.

[0053] In this implementation, the indicator vector can be indicated by a first relationship between the indicator vector and the identifier of the AI ​​sequence, so that the terminal can map the identifier of the first AI sequence to a first vector based on the first relationship, and input the first vector into the AI ​​model to generate the first AI sequence.

[0054] In conjunction with the third aspect, in some implementations of the third aspect, the first information includes at least one of the following: a first formula satisfied between the identifier of the AI ​​sequence and the vector, parameters in the first formula, the length of the vector, or the processing method of the vector.

[0055] As an example, the first relationship can be a parameterized mapping. For instance, the identifiers and vectors of an AI sequence can satisfy a first formula, which can contain parameters. Optionally, the network device can simultaneously indicate the first formula and the parameters within it to improve communication efficiency. Alternatively, the first formula can be predefined by a protocol or standard, and the parameters within the first formula can be indicated by the network device to enhance the flexibility of using the first formula.

[0056] As an example, the first relationship can be a parameterless mapping. For instance, the identifiers and vectors of an AI sequence can satisfy multi-base conversion. The specific conversion mechanism can be predefined by the protocol or configured by the network device, and is not limited here.

[0057] Optionally, the network device can indicate the length k of the vector to the terminal so that the vectors input to the AI ​​model remain uniform in length, thereby improving the performance of the AI ​​model.

[0058] Optionally, the network device can also indicate to the terminal how the vector is processed, such as whether to perform normalization, to improve the performance of the AI ​​model. In some embodiments, the network device can also indicate to the terminal the normalization interval.

[0059] In conjunction with the third aspect, in some implementations of the third aspect, the first information is used to indicate a vector, including: the first information is used to indicate the vector set in which the vector is located, multiple vectors in the vector set correspond one-to-one with multiple AI sequences, each of the multiple vectors satisfies the first relationship with the identifier of the corresponding AI sequence, and the multiple AI sequences belong to a subset of the sequence set in which the AI ​​sequence is located.

[0060] In this implementation, the network device can indicate vectors through a set of indicator vectors. Each vector in the vector set corresponds one-to-one with a sequence in the set of AI sequences generated by the network device during training. For example, when the first vector in the vector set corresponds to the first AI sequence in the AI ​​sequence set, a first relation can be satisfied between the identifiers of the first vector and the first AI sequence. It should be understood that each vector in the vector set can be used as input to the first AI model.

[0061] In this implementation, the network device can directly send the mapped vector set to the terminal without indicating the first relation. Since the data size of the vector set can be smaller than that of the AI ​​sequence set, transmitting the vector set can reduce signaling overhead compared to transmitting the AI ​​sequence set.

[0062] In conjunction with the third aspect, in some implementations of the third aspect, the first information includes at least one of the following: the plurality of vectors, or the processing method of each of the plurality of vectors.

[0063] As an example, a vector set can be represented as [M, k]. Here, M represents the number of vectors in the vector set, and k represents the length of each vector. The number of vectors in the vector set can be the same as the number of AI sequences in the AI ​​sequence set.

[0064] Optionally, the length of the vector can be less than the length of the AI ​​sequence, so that the data volume of the vector set is less than the data volume of the AI ​​sequence set. Transmitting the vector set can reduce transmission overhead compared to transmitting the AI ​​sequence set.

[0065] Optionally, the network device can also indicate to the terminal how each vector in the vector set should be processed, such as whether to normalize the vectors to improve the performance of the AI ​​model. In some embodiments, the network device can also indicate to the terminal the normalization interval.

[0066] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes: sending second information, the second information being used to instruct the AI ​​model.

[0067] In this implementation, the network device can also instruct the AI ​​model to the terminal.

[0068] Optionally, when the structure of the AI ​​model is predefined by a protocol or standard, or defined offline by network devices and terminals, the second information may only indicate the parameters of the AI ​​model, or the second information may indicate both the parameters of the AI ​​model and the identifier of the structure of the AI ​​model, in order to reduce the overhead and complexity of AI model configuration.

[0069] Optionally, when the structure and parameters of the AI ​​model are predefined by a protocol or standard, or are defined offline by network devices and terminals, the second information may only indicate the identifier of the AI ​​model.

[0070] Optionally, when the structure and / or parameters of the AI ​​model are configured by the network device, the second information can indicate the structure and / or parameters of the AI ​​model to improve the flexibility of AI model configuration.

[0071] Optionally, the AI ​​model structure can be a model structure supported by the Z4 mode with a fixed model structure in model transfer.

[0072] In some embodiments, the AI ​​model may be trained by a network device.

[0073] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes: receiving third information, the third information being used to indicate at least one of the following: supported AI model information, supported first relation, or the duration required to generate the AI ​​sequence.

[0074] In this implementation, the terminal can report its capabilities to the network device, such as whether it supports the ability to generate AI sequences through AI models. When the terminal supports the ability to generate AI sequences through AI models, it can report the AI ​​model information it supports (such as model structure and / or model parameters), and / or the first relation it supports, so that the network device can be aware of the terminal's capabilities and improve communication performance.

[0075] In some embodiments, when the terminal supports the ability to generate AI sequences through an AI model, the terminal can report the time required for the terminal to generate the AI ​​sequence to the network device, so that the network device can know the terminal's capabilities and improve communication performance.

[0076] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes: sending a sixth message, the sixth message being used to indicate the identifier of the first AI sequence.

[0077] In this implementation, the network device can indicate the identifier of the AI ​​sequence to the terminal, enabling the terminal to determine the vector input into the AI ​​model based on the identifier indicated by the network device, and thus generate the AI ​​sequence corresponding to the identifier. This implementation, where the network device indicates the identifier of the AI ​​sequence, reduces interference between AI sequences generated by different terminals compared to the terminal randomly selecting the identifier.

[0078] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes: performing any of the following operations during a first time period: receiving a first sequence, or not receiving a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0079] In this implementation, considering that the AI ​​sequence generation process will take a certain amount of time, or that there is a delay in the AI ​​sequence generation process, the behavior of network devices during this period can be agreed upon, so that network devices know what behavior can be adopted during this period of time, thereby improving communication performance.

[0080] Fourthly, embodiments of this application provide a communication method that can be applied to either a detection end or a receiving end of an AI sequence. Taking a network device as an example, the method can be applied to the network side, such as an access network device, a module (e.g., a circuit, chip, or chip system) within the access network device, or a logical node, logical module, or software capable of implementing all or part of the functions of the access network device. Taking the application of this method to a network device (such as an access network device) as an example, the method includes: sending first information, the first information indicating a plurality of sub-sequence sets, the plurality of sub-sequence sets satisfying a second relationship with an AI sequence set; receiving a first AI sequence, the first AI sequence being generated based on the first information, the first AI sequence being included in the AI ​​sequence set; wherein, the product of the number of sub-AI sequences included in each of the plurality of sub-sequence sets is equal to the number of AI sequences included in the sequence set, and the sum of the number of sub-AI sequences included in each of the sub-sequence sets is less than the number of AI sequences included in the AI ​​sequence set.

[0081] In this technical solution, after training and generating an AI sequence set, the network device can divide the AI ​​sequence set into multiple sub-sequence sets according to the hierarchical mapping method to be adopted, and send them to the terminal. Alternatively, the network device can directly train multiple sub-sequence sets during the AI ​​sequence training process, based on the hierarchical mapping method to be adopted, and send them to the terminal. It should be noted that no restrictions are placed on the hierarchical mapping method used or its specific implementation.

[0082] Taking the quotient-remainder mapping method as an example, suppose the AI ​​sequence set is represented as M×N, meaning the AI ​​sequence set includes M AI sequences, and each AI sequence has a length of N. If the number of subsequence sets is 2, then the two subsequence sets can be represented as M1×N and M2×N respectively. Here, M1×M2=M, and M1+M2<M, thus allowing the total data size of the subsequence sets to be less than the data size of the AI ​​sequence set, reducing transmission overhead. Here, M, N, M1, and M2 are all positive integers.

[0083] Optionally, the fourth piece of information can be predefined through the protocol.

[0084] In this technical solution, the terminal can determine the AI ​​sequence set based on multiple sub-sequence sets and a second relationship, and then randomly select an AI sequence from the AI ​​sequence set to send to the network device. Since the total data volume of the sub-sequence set is less than the data volume of the AI ​​sequence set, the transmission overhead is low when the network device sends the sub-sequence set to the terminal.

[0085] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the method further includes: sending fourth information, the fourth information being used to indicate the second relationship.

[0086] In this implementation, the second relationship can be configured by the network device for the terminal. For example, the network device can determine the second relationship based on real-time network conditions, load, etc., and send it to the terminal to improve communication performance.

[0087] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the method further includes: sending fifth information, the fifth information being used to indicate a third relationship, the third relationship being the relationship between the identifier of the AI ​​sequence and multiple sub-identifiers, the multiple sub-identifiers corresponding to the multiple sub-sequence sets, each of the multiple sub-identifiers being used to determine a sub-AI sequence in the corresponding sub-sequence set, and the AI ​​sequence being generated based on the multiple sub-AI sequences determined by the multiple sub-identifiers.

[0088] In this implementation, a sub-identifier can be understood as the identifier of a sub-AI sequence within a set of sub-sequences. The terminal can map the identifier of an AI sequence to multiple sub-identifiers based on a third relation, and determine the corresponding sub-AI sequence from the set of sub-sequences based on these multiple sub-identifiers, thus determining the AI ​​sequence. The AI ​​sequence can be the product of multiple sub-AI sequences.

[0089] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the method further includes: sending a sixth message, the sixth message being used to indicate the identifier of the first AI sequence.

[0090] In this implementation, the network device can indicate the identifier of the AI ​​sequence to the terminal, enabling the terminal to determine the vector input into the AI ​​model based on the identifier indicated by the network device, and thus generate the AI ​​sequence corresponding to the identifier. This implementation, where the network device indicates the identifier of the AI ​​sequence, reduces interference between AI sequences generated by different terminals compared to the terminal randomly selecting the identifier.

[0091] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the method further includes: performing any of the following operations during a first time period: receiving a first sequence, or not receiving a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0092] In this implementation, considering that the AI ​​sequence generation process will take a certain amount of time, or that there is a delay in the AI ​​sequence generation process, the behavior of network devices during this period can be agreed upon, so that network devices know what behavior can be adopted during this period of time, thereby improving communication performance.

[0093] Fifthly, this application provides a communication device that has the functions of the first aspect described above. For example, the communication device includes modules, units, or means that perform the operations involved in the first aspect. These modules, units, or means can be implemented by software, hardware, or a combination of software and hardware.

[0094] For example, the device may include a receiving module and a transmitting module. The receiving module is configured to receive first information, the first information being used to indicate a vector, the vector satisfying a first relationship with the identifier of the AI ​​sequence, the vector being used as input to an AI model, the AI ​​model being used to generate the AI ​​sequence; the transmitting module is configured to transmit a first AI sequence, the first AI sequence being generated based on the first information.

[0095] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the first information is used to indicate a vector, including: the first information is used to indicate the first relationship.

[0096] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the first information includes at least one of the following: a first formula satisfied between the identifier of the AI ​​sequence and the vector, parameters in the first formula, the length of the vector, or the processing method of the vector.

[0097] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the first information is used to indicate a vector, including: the first information is used to indicate the vector set in which the vector is located, multiple vectors in the vector set correspond one-to-one with multiple AI sequences, each of the multiple vectors satisfies the first relationship with the identifier of the corresponding AI sequence, and the multiple AI sequences belong to a subset of the sequence set in which the AI ​​sequence is located.

[0098] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the first information includes at least one of the following: the plurality of vectors, or the processing method of each of the plurality of vectors.

[0099] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the receiving module is further configured to receive second information, which is used to instruct the AI ​​model.

[0100] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the sending module is further configured to send third information, which indicates at least one of the following: supported AI model information, supported first relation, or the duration required to generate the AI ​​sequence.

[0101] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the receiving module is further configured to receive sixth information, which is used to indicate the identifier of the first AI sequence.

[0102] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the device may further include a processing module. The processing module is configured to control the device to perform any of the following operations during a first time period: sending a first sequence, or not sending a sequence; wherein the start time of the first time period is the start time for generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0103] In a sixth aspect, this application provides a communication device that has the functions of the second aspect described above. For example, the communication device includes modules, units, or means that perform the operations involved in the second aspect described above. These modules, units, or means can be implemented by software, hardware, or a combination of software and hardware.

[0104] For example, the device may include a receiving module and a transmitting module. The receiving module is configured to receive first information, which indicates a plurality of sub-sequence sets, wherein the plurality of sub-sequence sets and an AI sequence set satisfy a second relationship. The transmitting module is configured to transmit a first AI sequence, which is generated based on the first information and is included in the AI ​​sequence set. Specifically, the product of the number of sub-AI sequences included in each of the plurality of sub-sequence sets is equal to the number of AI sequences included in the AI ​​sequence set, and the sum of the number of sub-AI sequences included in each of the sub-sequence sets is less than the number of AI sequences included in the AI ​​sequence set.

[0105] In conjunction with the sixth aspect, in some implementations of the sixth aspect, the receiving module is further configured to receive fourth information, which is used to indicate the second relationship.

[0106] In conjunction with the sixth aspect, in some implementations of the sixth aspect, the receiving module is further configured to receive fifth information, which is used to indicate a third relationship, the third relationship being the relationship between the identifier of the AI ​​sequence and multiple sub-identifiers, the multiple sub-identifiers corresponding to the set of multiple sub-sequences, each of the multiple sub-identifiers being used to determine a sub-AI sequence in the corresponding set of sub-sequences, and the AI ​​sequence being generated based on the multiple sub-AI sequences determined by the multiple sub-identifiers.

[0107] In conjunction with the sixth aspect, in some implementations of the sixth aspect, the receiving module is further configured to receive sixth information, which is used to indicate the identifier of the first AI sequence.

[0108] In conjunction with the sixth aspect, in some implementations of the sixth aspect, the device may further include a processing module. The processing module is configured to control the device to perform any of the following operations during a first time period: sending a first sequence, or not sending a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0109] In a seventh aspect, this application provides a communication device that has the functions of the third aspect described above. For example, the communication device includes modules, units, or means that perform the operations involved in the third aspect. These modules, units, or means can be implemented by software, hardware, or a combination of software and hardware.

[0110] For example, the device may include a sending module and a receiving module. The sending module is used to send first information, which indicates a vector that satisfies a first relationship with the identifier of the AI ​​sequence, the vector being used as input to an AI model, and the AI ​​model being used to generate the AI ​​sequence; the receiving module is used to receive a first AI sequence, which is generated based on the first information.

[0111] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the first information is used to indicate a vector, including: the first information is used to indicate the first relationship.

[0112] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the first information includes at least one of the following: a first formula satisfied between the identifier of the AI ​​sequence and the vector, parameters in the first formula, the length of the vector, or the processing method of the vector.

[0113] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the first information is used to indicate a vector, including: the first information is used to indicate the vector set in which the vector is located, multiple vectors in the vector set correspond one-to-one with multiple AI sequences, each of the multiple vectors satisfies the first relationship with the identifier of the corresponding AI sequence, and the multiple AI sequences belong to a subset of the sequence set in which the AI ​​sequence is located.

[0114] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the first information includes at least one of the following: the plurality of vectors, or the processing method of each of the plurality of vectors.

[0115] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the sending module is also used to send second information, which is used to instruct the AI ​​model.

[0116] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the receiving module is further configured to receive third information, which is used to indicate at least one of the following: supported AI model information, supported first relation, or the duration required to generate the AI ​​sequence.

[0117] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the sending module is also used to send sixth information, which is used to indicate the identifier of the first AI sequence.

[0118] In conjunction with the seventh aspect, in some implementations of the seventh aspect, the device may further include: a processing module. The processing module is configured to control the device to perform any of the following operations during a first time period: receiving a first sequence, or not receiving a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0119] Eighthly, this application provides a communication device that has the functions of the fourth aspect above. For example, the communication device includes modules, units, or means that perform the operations involved in the fourth aspect above. These modules, units, or means can be implemented by software, hardware, or a combination of software and hardware.

[0120] For example, the device may include a sending module and a receiving module. The sending module is used to send first information, which indicates a plurality of sub-sequence sets, the plurality of sub-sequence sets satisfying a second relationship with an AI sequence set; the receiving module is used to receive a first AI sequence, the first AI sequence being generated based on the first information, and the first AI sequence being included in the AI ​​sequence set; wherein, the product of the number of sub-AI sequences included in each of the plurality of sub-sequence sets is equal to the number of AI sequences included in the sequence set, and the sum of the number of sub-AI sequences included in each of the sub-sequence sets is less than the number of AI sequences included in the AI ​​sequence set.

[0121] In conjunction with the eighth aspect, in some implementations of the eighth aspect, the sending module is also used to send fourth information, which is used to indicate the second relationship.

[0122] In conjunction with the eighth aspect, in some implementations of the eighth aspect, the sending module is further configured to send fifth information, which is used to indicate a third relationship, the third relationship being the relationship between the identifier of the AI ​​sequence and multiple sub-identifiers, the multiple sub-identifiers corresponding to the set of multiple sub-sequences, each of the multiple sub-identifiers being used to determine a sub-AI sequence in the corresponding set of sub-sequences, and the AI ​​sequence being generated based on the multiple sub-AI sequences determined by the multiple sub-identifiers.

[0123] In conjunction with the eighth aspect, in some implementations of the eighth aspect, the sending module is also used to send sixth information, which is used to indicate the identifier of the first AI sequence.

[0124] In conjunction with aspect eight, in some implementations of aspect eight, the device may further include a processing module. The processing module is configured to control the device to perform any of the following operations during a first time period: receiving a first sequence, or not receiving a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0125] Ninthly, this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of the computer program or instructions necessary to implement the functions described in the first aspect above. The one or more processors are executable to carry out the computer program or instructions, causing the communication device to implement the methods in any possible design or implementation of the first aspect above. The interface circuit is used to implement the communication functions within the communication device and / or the communication functions between the communication device and other devices or components.

[0126] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0127] In one possible design, the communication device may also include the memory.

[0128] The communication device may be a terminal, a communication module in a terminal, or a chip in a terminal that is responsible for communication functions, such as a modem chip (also known as a baseband chip) or a SoC or SIP chip that contains a modem module.

[0129] Tenthly, this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of the necessary computer program or instructions for implementing the functions described in the second aspect above. The one or more processors are executable to carry out the computer program or instructions, causing the communication device to implement the methods in any possible design or implementation of the second aspect above. The interface circuit is used to implement the communication functions within the communication device and / or the communication functions between the communication device and other devices or components.

[0130] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0131] In one possible design, the communication device may also include the memory.

[0132] The communication device may be a terminal, a communication module in a terminal, or a chip in a terminal that is responsible for communication functions, such as a modem chip (also known as a baseband chip) or a SoC or SIP chip that contains a modem module.

[0133] Eleventhly, this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of the necessary computer program or instructions for implementing the functions described in the third aspect above. The one or more processors can execute the computer program or instructions, causing the communication device to implement the methods in any possible design or implementation of the third aspect above when executed. The interface circuit is used to implement the communication functions within the communication device and / or the communication functions between the communication device and other devices or components.

[0134] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0135] In one possible design, the communication device may also include the memory.

[0136] The communication device may be a network device (such as a base station), or a module (such as a circuit, chip, or chip system) in a network device, or a logical node, logical module, or software that can realize all or part of the functions of the network device.

[0137] In a twelfth aspect, this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of the necessary computer program or instructions for implementing the functions described in the fourth aspect above. The one or more processors are executable to carry out the computer program or instructions, causing the communication device to implement the methods in any possible design or implementation of the fourth aspect above. The interface circuit is used to implement the communication functions within the communication device and / or the communication functions between the communication device and other devices or components.

[0138] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0139] In one possible design, the communication device may also include the memory.

[0140] The communication device may be a network device (such as a base station), or a module (such as a circuit, chip, or chip system) in a network device, or a logical node, logical module, or software that can realize all or part of the functions of the network device.

[0141] In a thirteenth aspect, this application provides a communication system that includes the means of the fifth or ninth aspect, as well as the means of the seventh or eleventh aspect; or the communication system includes the means of the sixth or tenth aspect, as well as the means of the eighth or twelfth aspect.

[0142] In a fourteenth aspect, this application provides a chip system including at least one processor for implementing the functions involved in any of the first to fourth aspects and any possible implementations of any of them. For example, transmitting, receiving, or processing data and / or information involved in the above methods. In one possible implementation, the chip system further includes a memory for storing program instructions and data, the memory being located within or outside the processor. The chip system may be composed of chips or may include chips and other discrete devices.

[0143] In a fifteenth aspect, this application provides a computer-readable storage medium storing computer-readable instructions that, when read and executed by a computer, cause the computer to perform the methods of any one of the first to fourth aspects and any possible implementation thereof.

[0144] In a sixteenth aspect, this application provides a computer program product that, when read and executed by a computer, causes the computer to perform the methods of any one of the first to fourth aspects and any possible implementation thereof.

[0145] The technical effects that can be achieved by any one of the above aspects from the fifth to the sixteenth, and any one of their possible implementations, are described in the same way as the technical effects that can be achieved by any one of the above aspects from the first to the fourth, and any one of their possible implementations. They will not be repeated here. Attached Figure Description

[0146] Figure 1 This is a schematic illustration of a communication system applicable to the embodiments of this application;

[0147] Figure 2 A schematic flowchart illustrating a communication method provided in one embodiment of this application;

[0148] Figure 3 A schematic flowchart illustrating a communication method provided for another embodiment of this application;

[0149] Figure 3a A schematic illustration of training a first AI model provided for one embodiment of this application;

[0150] Figure 3b A schematic flowchart illustrating the generation of a first AI sequence is provided for one embodiment of this application;

[0151] Figure 4 A schematic flowchart illustrating a communication method provided in yet another embodiment of this application;

[0152] Figure 5 A schematic flowchart illustrating a communication method provided for another embodiment of this application;

[0153] Figure 6 A schematic flowchart illustrating a communication method provided in yet another embodiment of this application;

[0154] Figure 7 A schematic illustration of determining a first AI sequence provided for one embodiment of this application;

[0155] Figure 8 This is a schematic diagram of the structure of a communication device provided in one embodiment of this application;

[0156] Figure 9 This is a schematic diagram of the structure of a communication device provided in another embodiment of this application. Detailed Implementation

[0157] In communication systems, reference signals can be generated based on predefined reference sequences, enabling communicating parties to agree on the transmitted signals. This allows the receiving end to detect the received signal based on the agreed-upon signal after receiving it from the sending end, achieving purposes such as access, synchronization, demodulation, measurement, and estimation. For example, reference signals may include, but are not limited to: preamble sequences for random access, synchronization signal sequences for synchronization, channel state information-reference signal (CSIRS) sequences and sounding reference signal (SRS) sequences for channel measurement, and demodulation reference signal (DMRS) sequences for data demodulation. Future communication systems may also include sensing reference signal sequences for sensing purposes.

[0158] Taking random access as an example, terminals and network devices (such as base stations) can agree on a preamble sequence. The terminal can select a preamble sequence from the agreed set and send it during a specific random access opportunity (such as a random access slot) to initiate a random access request. Correspondingly, the network device can receive uplink signals during the specific random access slot and perform correlation detection on the received uplink signals based on the agreed set of preamble sequences. This determines the correlation value between the received uplink signal and each preamble sequence in the set, thus obtaining a correlation value sequence. The correlation value sequence represents the degree of matching between the received uplink signal and each preamble sequence; a higher correlation value indicates a higher degree of matching. After determining the peak value in the correlation value sequence, the network device can determine the relationship between the peak value and a preset detection threshold. If the peak value is greater than or equal to the detection threshold, it indicates that a preamble sequence sent by the terminal exists in the received uplink signal, and the network device can provide access service to the terminal. If there are multiple peak values ​​in the correlation value sequence, it indicates that multiple terminals have sent preamble sequences.

[0159] When multiple terminals send preamble sequences, the uplink signal received by the network device is a superposition of multiple preamble sequences. However, since commonly used preamble sequences are generated based on ZC (zadoff_chu) sequences, there is a certain similarity between multiple ZC sequences, leading to interference between them. This results in low detection performance of the network device for multiple preamble sequences. Furthermore, the network device detects the received uplink signal based on a detection threshold, but the optimal detection threshold is difficult to determine, leading to low detection accuracy. Therefore, in multi-user random access scenarios, the network device exhibits low detection performance for received uplink signals, resulting in poor communication performance.

[0160] In view of this, this application proposes to employ artificial intelligence (AI) technology to improve the detection performance of network devices, thereby improving the communication performance of communication systems. For example, a dual-end AI model can be used, integrating AI models from the terminal side and the network device side as a whole to complete an AI function. This dual-end AI model can include an AI sequence from the terminal side and a corresponding AI detection model from the network device side. Since the AI ​​sequence from the terminal side and the AI ​​detection model from the network device side are jointly trained, interference between multiple AI sequences can be reduced in scenarios with large-scale concurrent access from terminals. When the network device detects the AI ​​sequence based on the corresponding AI detection model, the detection performance will also be improved accordingly. An AI sequence can be understood as a sequence determined using AI methods. The AI ​​detection model can also be called an AI detection network.

[0161] The communication system proposes a Type 1 training method for training dual-end AI models. Type 1 training involves a single node in a network device or terminal training the entire dual-end AI model and then sending the generated AI model to the peer device. For example, the network device can train an AI sequence and a corresponding AI detection model, sending the generated AI sequence to the terminal. The terminal then uses the AI ​​sequence as a reference signal for transmission, while the network device uses the corresponding AI detection model for detection. It's important to note that during the training of the AI ​​sequence and AI detection model, the AI ​​sequence is actually part of the overall dual-end AI model. In typical training methods, the AI ​​sequence is treated as a parameter within the overall model, rather than being trained as a black-box model.

[0162] However, after the AI ​​sequence training is complete, how the network device sends the generated AI sequence to the terminal becomes a technical problem that needs to be solved. One possible implementation is that the network device can directly send the generated AI sequence to the terminal. For example, in a random access scenario, the AI ​​sequence can be sent via broadcast messages, such as sending the AI ​​sequence to all terminals in the cell via a system information block (SIB). However, the data volume of the generated AI sequence is large, which may exceed the data volume range that a broadcast message can carry, making it difficult for the AI ​​sequence to be transmitted correctly, thus affecting the implementation of the entire AI function; a large data volume of the generated AI sequence will also lead to high power consumption during AI sequence transmission.

[0163] Therefore, this application proposes a communication method and a communication device. In the technical solution provided by this application, the network device can send instruction information to the terminal to indicate the generation method of the AI ​​sequence, enabling the terminal to autonomously generate the AI ​​sequence. This reduces the transmission resources and power consumption of the network device and the terminal, thereby improving the communication performance of the system.

[0164] It should be understood that the technical solution provided in this application can be applied to two-end communication scenarios where the transmitting end sends signals and the receiving end detects the signals sent by the transmitting end. For example, it can be used for the transmission and detection of preamble sequences in random access scenarios, as well as for the transmission and detection of synchronization signal sequences, CSIRS sequences, SRS sequences, DMRS sequences, etc., and this application does not limit this. For example, the network device can train AIDMRS sequences and corresponding AI channel estimation models, and send the generation method of AIDMRS sequences to the terminal, so that the terminal can autonomously generate AIDMRS sequences and send the generated AIDMRS sequences to the network device. The network device uses the AI ​​channel estimation model corresponding to the AIDMRS sequences to detect the AIDMRS sequences sent by the terminal in order to perform corresponding channel estimation. The technical solution provided in this application can be applied to the transmission of uplink and downlink signals, as well as to the transmission of sidelink signals, and this application does not limit this.

[0165] The technical solutions provided in this application can be applied to various communication systems, including but not limited to: narrowband Internet of Things (NB-IoT) systems, long term evolution (LTE) systems, LTE advanced (LTE-A) systems, LTE frequency division duplex (FDD) systems, LTE time division duplex (TDD) systems, fourth-generation (4G) mobile communication systems, fifth-generation (5G) mobile communication systems, new radio (NR) communication systems, and future communication systems, etc. This application does not impose specific limitations on these systems.

[0166] Figure 1 This is a schematic diagram illustrating a communication system to which embodiments of this application apply. Figure 1 As shown, the communication system 100 may include a radio access network (RAN) 110 and a core network (CN) 120. RAN 110 may include at least one RAN node (e.g., Figure 1 130a and 130b (collectively referred to as 130) and at least one terminal (such as Figure 1 RAN 110, denoted as 140a-140j, is collectively referred to as RAN 140. RAN 110 may also include other RAN nodes, such as wireless relay equipment and / or wireless backhaul equipment. Figure 1(Not shown in the image). Terminal 140 can be wirelessly connected to RAN node 130. RAN node 130 can be wirelessly or wired connected to core network 120. The core network equipment in core network 120 and RAN node 130 in RAN 110 can be different physical devices, or they can be the same physical device integrating core network logical functions and radio access network logical functions. In some embodiments, communication system 100 may also include Internet 150.

[0167] RAN 110 can be a cellular system related to the 3rd Generation Partnership Project (3GPP), such as 4G, 5G mobile communication systems, or future-oriented evolution systems. RAN 110 can also be an open access network (O-RAN or ORAN), a cloud radio access network (CRAN), or a WiFi system. RAN 110 can also be a communication system that integrates two or more of the above systems.

[0168] RAN node 130, sometimes also referred to as access network equipment, RAN entity, or access node, constitutes part of the communication system and is used to help terminals achieve wireless access. Multiple RAN nodes 130 in communication system 100 can be of the same type or different types. In some scenarios, the roles of RAN node 130 and terminal 140 are relative, for example... Figure 1 Network element 140i can be a helicopter or a drone, and it can be configured as a mobile base station. For terminals 140j that access RAN 110 through network element 140i, network element 140i is a base station; however, for base station 130a, network element 140i is a terminal. RAN node 110 and terminal 140 are sometimes referred to as communication devices, for example... Figure 1 Network elements 130a and 130b can be understood as communication devices with base station functions, while network elements 140a-140j can be understood as communication devices with terminal functions.

[0169] In one possible scenario, a RAN node can be a base station, an evolved NodeB (eNodeB), an access point (AP), a transmission and receiving point (TRP), a gNB, or a base station in a future mobile communication system. A RAN node can also be a macro base station (such as...). Figure 1 130a), micro base stations or indoor stations (such as Figure 1The RAN node can be a relay node or donor node (as described in section 130b), or a wireless controller in a CRAN scenario. Optionally, the RAN node can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). All or part of the functions of the RAN node in this application can also be implemented through software functions running on hardware, or through virtualization functions instantiated on a platform (e.g., a cloud platform). The RAN node can also be equipped with communication modules, circuits, or chips that perform corresponding communication functions. The RAN node can also be configured with program instructions for performing corresponding communication functions and corresponding program instructions. The RAN node in this application can also be a logical node, logical module, or software that can implement all or part of the functions of the RAN node. In the embodiments of this application, the RAN node can also be described in different ways, such as a network device. Unless otherwise specified, the term "network device" will be used in this application, which is the original term for access network equipment (such as a base station). In this embodiment, the network device may include a baseband processor. Devices used to enable AI functionality in network devices can be accelerators for AI applications in network devices, such as neural processing units (NPUs) and graphics processing units (GPUs).

[0170] In another possible scenario, multiple RAN nodes collaborate to assist the terminal in achieving wireless access, with different RAN nodes each implementing a portion of the base station's functions. For example, RAN nodes can be central units (CUs), distributed units (DUs), CU-control plane (CPs), CU-user plane (UPs), or radio units (RUs), etc. CUs and DUs can be set up separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).

[0171] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called O-CU (open CU), DU can also be called O-DU, CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called O-RU. For ease of description, this application uses CU, CU-CP, CU-UP, DU, and RU as examples. Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software and hardware modules.

[0172] A terminal can be a device or module that accesses the aforementioned communication system and has corresponding communication functions. A terminal can also be called a terminal device, user equipment (UE), mobile station, mobile terminal, etc. Terminals can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, etc. A terminal can be a mobile phone, tablet computer, computer with wireless transceiver capabilities, wearable device, vehicle, drone, helicopter, airplane, ship, robot, robotic arm, smart home device, transportation vehicle with wireless communication capabilities, communication module, etc. The embodiments of this application do not limit the device form of the terminal. A terminal typically contains a communication module, circuit, or chip that performs the corresponding communication functions. The terminal can also be configured with program instructions for performing the corresponding communication functions. In this embodiment, the terminal may include a baseband processor. Devices used to enable edge AI functionality can be edge AI application accelerators, such as NPUs and GPUs.

[0173] To facilitate understanding of the methods provided in the embodiments of this application, the following points are first explained:

[0174] First, in the embodiments of this application, "sending information" can be understood as one device sending information to another device, or it can also be understood as one logical module within a device sending information to another logical module. For example, "network device sending information" can be understood as a network device sending information to another device (such as a terminal), or it can be understood as logical module 1 in the network device sending information to logical module 2 in the network device.

[0175] Second, in the embodiments of this application, "receiving information" can be understood as one device receiving information from another device, or it can also be understood as a logical module within a device receiving information from another logical module. For example, "terminal receiving information" can be understood as the terminal receiving information from another device (such as a network device), or it can be understood as logical module 1 in the terminal receiving information from logical module 2 in the terminal.

[0176] Third, in the embodiments of this application, "sending information to a device (such as a terminal)" or the relevant illustrations in the accompanying drawings can be understood as the destination of the information being the terminal. This can include sending information directly or indirectly to the terminal. "Receiving information from a device (such as a terminal)," "receiving information from a device (such as a terminal)," or "receiving information sent by a device (such as a terminal)," or the relevant illustrations in the accompanying drawings, can be understood as the source of the information being the terminal. This can include receiving information directly or indirectly from the terminal. Information may undergo necessary processing between the source and destination, such as format changes, but the destination can understand the valid information from the source. Similar expressions in this application can be understood in a similar way, and will not be repeated here.

[0177] The following is combined Figures 2 to 9 This application describes the communication methods and devices provided in its embodiments. It is understood that this application uses network devices and terminals as examples to illustrate the interaction, but it does not limit the execution subject of the interaction. For example, the method executed by the network device in this application can also be implemented by modules (e.g., circuits, chips, or chip systems) in the network device, or by logical nodes, logical modules, or software that can implement all or part of the functions of the access network device; the method executed by the terminal in this application can also be implemented by a communication module in the terminal, or by circuits or chips (such as modem chips (also known as baseband chips), or SoC chips containing modem modules, or SIP chips) in the terminal responsible for communication functions.

[0178] Figure 2 This is a schematic flowchart illustrating a communication method provided in one embodiment of this application. Figure 2 As shown, the method may include S201 and S202.

[0179] S201, the network device sends first information, which indicates a vector. The vector and the identifier of the AI ​​sequence satisfy a first relationship. The vector is used as input to the first AI model, and the first AI model is used to generate the AI ​​sequence. Correspondingly, the terminal receives the first information.

[0180] In this embodiment, the terminal can generate an AI sequence based on a first AI model. The first AI model may or may not be located within the terminal; the terminal can invoke the first AI model, and this is not limited. Optionally, the first AI model can be predefined through a protocol, and the terminal can obtain the first AI model through the protocol. For example, the structure and / or parameters of the first AI model can be defined by the protocol, and the terminal determines the first AI model according to the model structure and / or parameters defined by the protocol.

[0181] In this embodiment, the network device can indicate the input of the first AI model to the terminal, so that the terminal can generate an AI sequence based on the first AI model and its input. For example, the network device can send first information to the terminal, which indicates the input of the first AI model. This application does not limit the specific form of the input of the first AI model. For example, the input of the first AI model can be in the form of a vector, a matrix, etc. In this embodiment, a vector is used as the input of the first AI model for explanation.

[0182] In this embodiment, the vector indicated by the first information can be used as input to the first AI model. A first relationship exists between the vector and the identifier of the AI ​​sequence, enabling the first AI model to generate AI sequences with corresponding identifiers based on the vector. That is, different input vectors to the first AI model will result in different generated AI sequences. The identifier of the AI ​​sequence can be an integer used to distinguish different AI sequences.

[0183] S202, the terminal sends a first AI sequence, which is generated based on the first information. Correspondingly, the network device receives the first AI sequence.

[0184] In this embodiment, after receiving the first information, the terminal can input the vector indicated by the first information into the first AI model to generate an AI sequence. For example, if the first information indicates a first vector, and the first vector and the identifier of the first AI sequence satisfy a first relationship, the terminal can generate the first AI sequence after inputting the first vector into the first AI model. The first AI sequence can be understood as the AI ​​sequence that the terminal determines to send. Correspondingly, the network device can receive the first AI sequence and, after receiving it, detect it. For example, the network device can detect the received signal containing the first AI sequence to determine whether the terminal has sent the first AI sequence.

[0185] In this embodiment, after generating the AI ​​sequence, the terminal can send the AI ​​sequence to the network device. Correspondingly, the network device can receive and detect the AI ​​sequence, thereby providing services to the terminal. For example, when the AI ​​sequence is a random access sequence, the network device can provide access services to the terminal after receiving and detecting the AI ​​sequence.

[0186] In this embodiment, the terminal can generate the required AI sequence based on the vector indicated by the network device and the first AI model. In this embodiment, the network device can only indicate the input to the AI ​​model, instead of sending all the trained AI sequences to the terminal, thereby reducing signaling overhead. Furthermore, the data volume of the vector is smaller than that of the trained AI sequence, and broadcast messages can carry the vector, resulting in higher transmission reliability for the vector.

[0187] In one possible implementation, the first information used to indicate a vector may include: the first information indicating a vector by indicating a first relation, or the first information indicating a vector by indicating a set of vectors. The following is in conjunction with... Figure 3 and Figure 4 The specific implementation methods of these two approaches will be explained.

[0188] Figure 3 This is a schematic flowchart illustrating a communication method provided in another embodiment of this application. Figure 3 In the method shown, the first information can indicate the vector by indicating the first relation. For example... Figure 3 As shown, the method may include S301 to S305.

[0189] S301, the terminal sends third information, which indicates whether the ability to generate AI sequences through an AI model is supported. Correspondingly, the network device receives the third information.

[0190] In this embodiment, the terminal can report its capabilities to the network device, such as whether it supports the ability to generate AI sequences through an AI model. Correspondingly, the network device can receive the capabilities reported by the terminal. For ease of description, the ability to generate AI sequences through an AI model can be referred to as the first capability.

[0191] As an example, the AI ​​sequence is a preamble sequence, or in other words, the AI ​​sequence is used for random access before the terminal has accessed the network. Therefore, it can be assumed that the terminal supports the first capability. In some embodiments, when the terminal supports the first capability, it can be considered that the terminal has at least one default first relationship, or in other words, the terminal is assumed to have at least one supported first relationship.

[0192] As an example, when the AI ​​sequence is used for sequences such as DMRS, CSIRS, and SRS, and the terminal supports the first capability, the terminal can report at least one of the following information to the network device: supported AI model information, or supported first relation. The AI ​​model information may include the AI ​​model structure and / or AI model parameters. The terminal may report the supported AI model information and / or supported first relation to the network device through third information, or through other information; there are no restrictions on this.

[0193] In some embodiments, when the terminal supports the first capability, the terminal can report to the network device the time required to generate the AI ​​sequence, or the time required for the terminal to infer the AI ​​sequence based on the AI ​​model. The terminal can report the time required to generate the AI ​​sequence to the network device using third-party information, or other information; no restrictions are placed on this.

[0194] It should be noted that S301 is an optional step, which can be performed or not.

[0195] S302, the network device sends first information, which indicates a first relationship between the vector and the identifier of the AI ​​sequence. The vector is used as input to the first AI model, and the first AI model is used to generate the AI ​​sequence. Correspondingly, the terminal receives the first information.

[0196] In this embodiment, the network device can send first information to the terminal. This first information can indicate a first relationship, which is a relationship satisfied between a vector and the identifier of an AI sequence. Correspondingly, the terminal can receive the first information. Based on the identifier of the AI ​​sequence and the first relationship, the terminal can determine the vector input to the first AI model, thereby generating an AI sequence. In some embodiments, the first relationship can be used to indicate a preprocessing operation of fixed precoding before using the first AI model to recover or generate an AI sequence. Fixed precoding can be understood as the identifier of the AI ​​sequence, and the preprocessing operation refers to mapping the identifier of the AI ​​sequence to a vector. The first relationship indicated by the first information can be a first relationship supported by the terminal.

[0197] In one possible implementation, the first relationship can be a parameterized mapping, such as a first formula between the identifier and vector of an AI sequence, where the first formula contains parameters. In this implementation, the first information can include the first formula.

[0198] As an example, the identifier of an AI sequence and its vector can satisfy a global hash mapping formula, such as: q = ((a*x+b)mod p)mod m. Here, q is an element in the vector, x is the identifier of the AI ​​sequence, a is a random integer, b is a random non-zero integer, m is the size of the hash table, or the range of values ​​after hash mapping, and p is a random prime number greater than m. The elements of a vector can be understood as the individual numerical values ​​or objects that constitute the vector. It should be understood that a vector contains at least one element.

[0199] Optionally, the first formula can be predefined via a protocol, and the parameters such as a, b, p, and m within the first formula can be indicated by the network device to improve the flexibility of the formula's use. The terminal can determine the vector input to the first AI model based on the first formula and the identifier of the AI ​​sequence. Optionally, both the first formula and its parameters can be configured by the network device.

[0200] In one feasible approach, the first relationship can be a parameterless mapping. As an example, the identifier of an AI sequence and its vector can satisfy multi-base conversion. For instance, the identifier of an AI sequence and its vector can satisfy binary conversion; the terminal can perform binary conversion on the identifier of the AI ​​sequence and use the resulting binary number as a vector. Each bit in the binary number can be used as an element in the vector, with a value of 0 or 1. The specific conversion mechanism can be predefined by the protocol or configured by the network device, and is not limited here.

[0201] Optionally, the network device can indicate the dimension k of the vector to the terminal to ensure that the vectors input to the AI ​​model maintain a uniform dimension, thereby improving the performance of the AI ​​model. k is a positive integer. As an example, when the vector is a one-dimensional vector, the dimension of the vector can be understood as the number of elements contained in the vector. After the terminal determines the vector based on the first relation, if the dimension of the vector is less than k, the terminal can expand the vector to a vector of dimension k; if the dimension of the vector is greater than k, the terminal can reduce the vector to a vector of dimension k. In some embodiments, the dimension of the vector can also be referred to as the number of bits or the length of the vector, without limitation. The network device can indicate the dimension of the vector to the terminal through the first information or through other information, without limitation.

[0202] In one possible approach, if the identifiers of the AI ​​sequence and the vector satisfy a first formula, the terminal can use the first formula k times to determine a vector of dimension k. Each use of the first formula is to determine one element in the vector; in other words, the v-th use of the first formula is to determine the v-th element in the vector, where v is a positive integer less than or equal to k.

[0203] Optionally, the network device can also indicate the vector processing method to the terminal, such as indicating whether to normalize the vector to improve the performance of the AI ​​model. In some embodiments, the network device can also indicate the normalization interval to the terminal. The network device can indicate the vector processing method to the terminal through first information, or through other information; there is no limitation on this.

[0204] Optionally, the first information can be carried by radio resource control (RRC) messages, SIB messages, or non-access stratum (NAS) messages; there is no restriction on this.

[0205] S303, the network device sends second information, which is used to instruct the first AI model. Correspondingly, the terminal receives the second information.

[0206] In this embodiment, the network device can send second information to the terminal, which indicates the first AI model. Correspondingly, the terminal can receive the second information. Optionally, the first AI model indicated by the second information can be an AI model supported by the terminal.

[0207] Optionally, when the structure of the first AI model is predefined by a protocol or standard, or defined offline by network devices and terminals, the second information may only indicate the parameters of the first AI model, or the second information may indicate both the parameters of the first AI model and the identifier of the structure of the first AI model, in order to reduce the overhead and complexity of configuring the first AI model.

[0208] Optionally, when the structure and parameters of the first AI model are predefined by a protocol or standard, or are defined offline by network devices and terminals, the second information may only indicate the identifier of the first AI model.

[0209] Optionally, when the structure and / or parameters of the first AI model are configured by the network device, the second information can be used to indicate the structure and / or parameters of the first AI model to improve the flexibility of the configuration of the first AI model.

[0210] Optionally, the structure of the first AI model can be a model structure supported by the Z4 mode with a fixed model structure in model transfer.

[0211] Optionally, the terminal or terminal manufacturer can optimize the first AI model to avoid situations where the terminal cannot directly use the first AI model.

[0212] As an example, the second information can be transmitted via air interface messages, such as RRC messages, SIB messages, or NAS messages. Air interface messages can be control plane messages or data.

[0213] As an example, the second piece of information can be transmitted via non-air interface messages, such as business data packets or application layer data transmissions. Non-air interface messages can be user plane messages or data.

[0214] As an example, the first AI model can be a deep neural network (DNN) model. DNN models are well-suited for the Z4 model transmission mode with a fixed model structure, which simplifies the transmission of model information and reduces power consumption and overhead.

[0215] Optionally, the network device can also instruct the terminal on the processing method of the AI ​​sequence output by the first AI model. The processing method of the AI ​​sequence output by the first AI model may include inverse normalization processing or quantization processing. In some embodiments, the network device can also instruct the terminal on an inverse normalization processing interval or a quantization interval, so that the terminal can perform inverse normalization processing or quantization processing based on the maximum or minimum value of the interval indicated by the network device, thereby improving the detection performance of the AI ​​sequence.

[0216] In some embodiments, the first AI model may be trained by a network device. When the network device trains the AI ​​sequence and the corresponding AI detection model, the training process for the AI ​​sequence portion may include the first AI model. As an example, the first AI model may be a deep hash embedding (DHE) model, also known as an AI compression model.

[0217] Figure 3a This is a schematic illustration of training a first AI model according to one embodiment of this application. Figure 3a It consists of two parts: the first part is the training of the first AI model, and the second part is the training of the AI ​​sequence and AI detection models.

[0218] In one possible implementation, the AI ​​sequence and AI detection model can be trained first, and then a first AI model can be trained based on the trained AI sequence. For ease of description and distinction, the trained AI sequence is referred to as the initial AI sequence. Before training the first AI model, the network device needs to determine a first relationship between a sequence identifier and a vector, and determine an initial AI model as the first AI model. The network device can randomly select one or more identifiers from the identifiers of the initial AI sequence, map the selected identifiers to vectors sequentially based on the first relationship, and input them sequentially into the first AI model to generate one or more AI sequences. The network device can determine the performance of the first AI model (such as inference accuracy or the accuracy of generating the initial AI sequence) based on the initial AI sequence corresponding to the selected identifier and the AI ​​sequence generated by the first AI model. If the performance of the first AI model does not reach the target performance, the structure and / or parameters of the first AI model can be adjusted, and the training process can be executed again until the performance of the first AI model reaches the target performance. Figure 3 As shown, identifier 1 and identifier 2 are identifiers for the initial AI sequence. The network device can map identifier 1 to vector 1 based on the first relation and input vector 1 into the first AI model to generate AI sequence 1; similarly, it can map identifier 2 to vector 2 based on the first relation and input vector 2 into the first AI model to generate AI sequence 2. The network device can determine the performance of the first AI model based on the initial AI sequence corresponding to identifier 1 and the AI ​​sequence 1 generated by the first AI model, and based on the initial AI sequence corresponding to identifier 2 and the AI ​​sequence 2 generated by the first AI model.

[0219] After the first AI model is trained, the trained AI detection model can be retrained based on the AI ​​sequences output by the first AI model to improve its detection performance. For example... Figure 3 As shown, AI sequence 1 can be transmitted through channel 1, and AI sequence 2 can be transmitted through channel 2. Noise is superimposed on the transmitted signals before being input into the AI ​​detection model. The signal received by the AI ​​detection model contains AI sequence 1 and AI sequence 2, and the AI ​​detection model can detect AI sequence 1 and AI sequence 2 from the received signal. Assuming the AI ​​detection model detects sequences Q1 and Q2 from the received signal, the network device can determine the detection loss of the AI ​​detection model based on the detected sequences Q1 and Q2, and AI sequences 1 and 2, thereby determining the detection performance of the AI ​​detection model. If the detection loss of the AI ​​detection model does not reach the target detection loss, the AI ​​detection model can be adjusted, and the detection process can be repeated until the detection loss reaches the target detection loss. For example, the detection loss of the AI ​​detection model can be determined based on the cross-entropy loss function.

[0220] Optionally, after the first AI model is trained, the loss of the AI ​​sequences output by the first AI model can be verified. For example, the loss of the AI ​​sequences output by the first AI model can be determined based on a fuzzy function to reduce interference between the various AI sequences output by the first AI model and further improve the detection performance of the AI ​​sequences. If the loss of the AI ​​sequences output by the first AI model does not reach the target sequence loss, the above steps can be repeated until the AI ​​sequence loss of the first AI model reaches the target training loss, the performance of the first AI model reaches the target performance, and the detection loss of the AI ​​detection model reaches the target detection loss.

[0221] After the first AI model is trained, the generated AI sequences can be implicitly included in the first AI model, and the network device no longer has explicit AI sequences. It should be understood that after the first AI model is trained, inputting vectors generated based on the first relation mapping into the first AI model will generate AI sequences with corresponding identifiers.

[0222] S304, the network device sends the sixth information, which indicates the identifier of the first AI sequence. Correspondingly, the terminal receives the sixth information.

[0223] In this embodiment, the network device can send a sixth piece of information to the terminal, which is used to indicate the identifier of the first AI sequence.

[0224] In one possible implementation, the sixth information may include the number M of AI sequences trained by the network device. The terminal can determine the identifier range of the AI ​​sequences based on the sixth information, thereby determining the identifier of the first AI sequence. For example, the terminal can randomly select an identifier from the identifier range of the AI ​​sequences as the identifier of the first AI sequence, improving the flexibility of selecting the identifier of the first AI sequence. In some embodiments, the network device may indicate at least one of the following information to the terminal: the length N of each AI sequence trained, or the number of channels L of each AI sequence trained. The length N and the number of channels L of each AI sequence can be indicated by the sixth information or by other information, without limitation. In this example, the size of the set of AI sequences trained is M*N*L. Each value of the AI ​​sequence can be a complex number, a real number, a floating-point number, or a fixed-point number, without limitation in this application. M, N, and L are positive integers.

[0225] In one possible implementation, the sixth piece of information may include an identifier for the first AI sequence. The terminal can directly determine the identifier of the first sequence based on the sixth piece of information, thereby improving the communication efficiency of the system.

[0226] It should be noted that step S304 is optional and can be performed or not. For example, the identification range of AI sequences can be predefined by the protocol.

[0227] In this embodiment, there is no specific restriction on the execution order of S301 to S304.

[0228] S305, the terminal sends the first AI sequence. Correspondingly, the network device receives the first AI sequence.

[0229] In this embodiment, after obtaining the identifier of the first AI sequence, the terminal can map the identifier of the first AI sequence to a first vector based on a first relationship, and input the first vector into the first AI model, thereby generating the first AI sequence, such as... Figure 3b The method shown. As... Figure 3b As shown, the first vector can be represented as [x1, x2, ..., xk]. The terminal can send the generated first AI sequence to the network device. Correspondingly, the network device can receive and detect the first AI sequence and provide corresponding network services to the terminal. In some embodiments, after the terminal maps the identifier of the first AI sequence to the first vector based on the first relationship, the first vector can be normalized, and then the normalized first vector can be input into the first AI model to generate the first AI sequence.

[0230] In some embodiments, considering that the AI ​​sequence generation process takes a certain amount of time, or that there is a delay in the AI ​​sequence generation process, when the AI ​​sequence is a transmission mode indicated by the network device, the terminal can perform any of the following operations within a first time period: send the first sequence, or not send the sequence. The start time of the first time period can be the start time of generating the AI ​​sequence, or it can be the time when the terminal selects / determines the identifier of the first AI sequence. The duration of the first time period is the duration required to generate the AI ​​sequence. Correspondingly, the network device can receive the first sequence, or not receive the sequence, within the first time period. In this embodiment, by agreeing that the network device does not expect the terminal to send the corresponding sequence within the first time period, or in other words, agreeing that the network device does not expect the terminal to send any sequence within the first time period, the behavior of the terminal and the network device within this period is standardized, allowing the terminal and the network device to know what behavior can be adopted within this period to improve communication performance. Specifically, when the AI ​​sequence is a DMRS sequence, it can be agreed that within the first time period, the terminal can ignore the corresponding uplink scheduling or use the default legacy DMRS sequence.

[0231] In this embodiment, the network device can send an indication of a first relationship to the terminal. The terminal can then map the identifier of the AI ​​sequence into a vector based on the first relationship and input it into the first AI model to generate the first AI sequence. In this embodiment, generating the corresponding AI sequence through an AI model saves air interface transmission overhead and reduces the power consumption of both the network device and the terminal compared to transmitting AI sequences generated during training. Furthermore, the method of generating AI sequences through an AI model results in higher detection accuracy for the network device, with no significant loss in detection performance.

[0232] Figure 4 This is a schematic flowchart illustrating a communication method provided in yet another embodiment of this application. Figure 4 In the method shown, the first information can be indicated by a set of indicator vectors. For example... Figure 4 As shown, the method may include S401 to S405.

[0233] S401, the terminal sends third information, which indicates whether the ability to generate AI sequences through an AI model is supported. Correspondingly, the network device receives the third information.

[0234] In this embodiment, the specific implementation of S401 can be referred to S301, and will not be repeated here.

[0235] Compared to S301, in this embodiment, when the terminal supports the first capability, the terminal can also report the set of vectors supported by the terminal to the network device. The terminal can report the set of vectors supported by the terminal to the network device through third information, or through other information; no limitation is made here.

[0236] S402, the network device sends first information, which indicates a vector set. Multiple vectors in the vector set correspond one-to-one with multiple AI sequences in the AI ​​sequence set, and each vector and its corresponding AI sequence identifier satisfy a first relationship. Correspondingly, the terminal receives the first information.

[0237] In this embodiment, the network device can determine a first relationship and map the identifiers of each AI sequence generated during training based on the first relationship, thereby generating a vector set. The device then indicates the generated vector set to the terminal via first information. Each vector in the vector set corresponds one-to-one with an AI sequence in the AI ​​sequence set. For example, when the first vector in the vector set corresponds to the first AI sequence in the AI ​​sequence set, the first relationship is satisfied between the identifier of the first vector and the first AI sequence. Accordingly, the terminal can receive the first information. It should be understood that each vector in the vector set can be used as input to the first AI model. The AI ​​sequence set may include some or all of the AI ​​sequences generated during training.

[0238] In one possible implementation, the vector set can be represented as [M, k]. Here, M represents the number of vectors in the vector set; k represents the dimension of each vector. It should be understood that the number of vectors in the vector set is the same as the number of AI sequences in the AI ​​sequence set.

[0239] Optionally, k can be less than the length N of the AI ​​sequence, so that the data size of the vector set is less than the data size of the AI ​​sequence set, thereby reducing the transmission overhead of the vector set.

[0240] Optionally, the network device can also indicate to the terminal the processing method for each of the multiple vectors, such as whether to normalize the vectors to improve the performance of the AI ​​model. In some embodiments, the network device can also indicate the normalization interval to the terminal. The network device can indicate the vector processing method to the terminal through first information, or through other information; there is no limitation on this.

[0241] Optionally, the initial information can be carried through RRC messages, SIB messages, or NAS messages; no restriction is imposed here.

[0242] S403, the network device sends second information, which is used to instruct the first AI model. Correspondingly, the terminal receives the second information.

[0243] In this embodiment, the specific implementation of S403 can be referred to S303, and will not be repeated here.

[0244] S404, the network device sends the sixth message, which indicates the identifier of the first AI sequence. Correspondingly, the terminal receives the sixth message.

[0245] In this embodiment, the specific implementation of S404 can be referred to S304, and will not be repeated here.

[0246] In this embodiment, step S404 is optional and can be performed or not. For example, the identification range of the AI ​​sequence can be determined based on the number of vectors in the vector set.

[0247] S405, the terminal sends the first AI sequence. Correspondingly, the network device receives the first AI sequence.

[0248] In this embodiment, after obtaining the identifier of the first AI sequence, the terminal can select a first vector from the vector set whose identifier is the same as that of the first AI sequence, and input the first vector into the first AI model to generate the first AI sequence, which is then sent to the network device. Correspondingly, the network device can receive and detect the first AI sequence and provide corresponding network services to the terminal. The identifier of the first AI sequence can be the same as the identifier of the first vector in the vector set.

[0249] In some embodiments, the terminal can directly select a vector randomly from the vector set and input the selected vector into the first AI model to generate an AI sequence.

[0250] In some embodiments, after determining the vector to be input to the first AI model, the terminal can normalize the vector and then input the normalized vector into the first AI model to generate an AI sequence.

[0251] In this embodiment, the behavior of the terminal and network devices during the first time period can also be defined to improve communication performance. For details, please refer to the relevant description in S305, which will not be repeated here.

[0252] In this embodiment, the network device can directly send the mapped vector set to the terminal without indicating the first relation, thereby reducing the standardization difficulty. Simultaneously, the data size of the vector set can be smaller than that of the AI ​​sequence set; therefore, compared to transmitting the AI ​​sequence set, transmitting the vector set can reduce transmission overhead. Figure 3 Compared to the method in the previous embodiment, in this embodiment, the terminal can achieve the purpose of generating the first AI sequence through the first AI model without explicitly obtaining the first relationship.

[0253] In one possible implementation, the terminal may not use an AI model to generate the AI ​​sequence, but instead uses a hierarchical mapping method to determine the AI ​​sequence. For example, the network device can hierarchically map the AI ​​sequence set into multiple sub-sequence sets and send them to the terminal. The terminal can determine the AI ​​sequence to be sent based on the multiple sub-sequence sets and the identifier of the AI ​​sequence to be sent. Since the total data volume of the multiple sub-sequence sets is less than the data volume of the AI ​​sequence set, the goal of reducing transmission overhead can also be achieved. The following section will combine... Figure 5 and Figure 6 This method will be explained.

[0254] Figure 5 This is a schematic flowchart illustrating a communication method provided in another embodiment of this application. Figure 5 In the method shown, the network device can hierarchically map the AI ​​sequence set into multiple sub-sequence sets, so that the terminal can determine the first AI sequence to be sent based on the multiple sub-sequence sets. For example... Figure 5 As shown, the method may include S501 to S504.

[0255] S501, the network device sends first information, which indicates multiple sets of sub-sequences, and the multiple sets of sub-sequences satisfy a second relationship with the AI ​​sequence set. Correspondingly, the terminal receives the first information.

[0256] As an example, after training and generating an AI sequence set, the network device can determine a set of T sub-sequences smaller than the size of the AI ​​sequence set based on the hierarchical mapping method to be adopted, and send the T sub-sequence sets to the terminal. T is a positive integer. This application embodiment does not limit the hierarchical mapping method used or its specific implementation. For example, the hierarchical mapping method can be a quotient-remainder mapping.

[0257] As an example, network devices can directly train a set of T subsequences by combining hierarchical mapping during the training process of AI sequences and AI detection models.

[0258] Taking the quotient-remainder mapping as an example, suppose the set of AI sequences is represented as M×N, meaning the set of AI sequences includes M AI sequences, and the length of each AI sequence is N. If T is 2, then the two subsequence sets can be represented as M1×N and M2×N respectively. Here, M1×M2=M, and M1+M2<M, thus making the data size of the subsequence set smaller than the data size of the AI ​​sequence set, reducing transmission overhead. M1 and M2 are positive integers.

[0259] In this embodiment, the network device can indicate the set of T hierarchically mapped sub-sequences to the terminal via first information. Correspondingly, the terminal can receive the first information. The set of T sub-sequences and the AI ​​sequence set satisfy a second relationship. This second relationship can be determined based on the hierarchical mapping method used.

[0260] Optionally, the first information can be carried by an RRC message, SIB message, or NAS message, or the set of T subsequences can be carried by an RRC message, SIB message, or NAS message.

[0261] S502, the network device sends a fourth message, which is used to indicate the second relationship.

[0262] In this embodiment, the network device can send fourth information to indicate a second relationship satisfied between the T subsequence sets and the AI ​​sequence set. Correspondingly, the terminal can receive the fourth information. After determining the second relationship, the terminal can determine the AI ​​sequence set based on the second relationship and the received T subsequence sets.

[0263] In some embodiments, considering that the second relationship can be determined based on a hierarchical mapping method, which can be a method agreed upon by the network device and the terminal, such as being predefined by a protocol or standard, or being defined by offline negotiation between the network device and the terminal, S502 is an optional step.

[0264] S503, the network device sends the sixth information, which is used to indicate the identifier of the first AI sequence. Correspondingly, the terminal receives the sixth information.

[0265] In this embodiment, the specific implementation of S503 can be referred to S304, and will not be repeated here.

[0266] Given that M1×M2=M, the terminal can determine the identifier range of the AI ​​sequence based on the received set of T sub-sequences, and randomly select one identifier as the identifier of the first AI sequence to improve the flexibility of the identifier selection for the first AI sequence. Therefore, S503 is an optional step and can be executed or not.

[0267] S504, the terminal sends a first AI sequence, which is generated based on the first information and is included in the AI ​​sequence set. Correspondingly, the network device receives the first AI sequence.

[0268] In this embodiment, after determining the identifier of the first AI sequence and the set of AI sequences, the terminal can identify the first AI sequence based on the identifier and the set of AI sequences, and then send it to the network device. Correspondingly, the network device can receive and detect the first AI sequence and provide corresponding network services to the terminal.

[0269] In one possible approach, the terminal can determine the AI ​​sequence set based on multiple sub-sequence sets and a second relation, thereby randomly selecting an AI sequence from the AI ​​sequence set and sending it to the network device.

[0270] In this embodiment, the network device can send multiple sub-sequence sets to the terminal. These sub-sequence sets and the AI ​​sequence set satisfy a second relationship. The terminal can determine the AI ​​sequence set based on the multiple sub-sequence sets, thereby determining the AI ​​sequence to be sent. In this embodiment, the total data volume of the multiple sub-sequence sets is less than the data volume of the AI ​​sequence set; therefore, transmitting multiple sub-sequence sets can reduce air interface overhead compared to transmitting the AI ​​sequence set.

[0271] In one possible implementation, when the network device indicates multiple sets of sub-sequences to the terminal, it can also indicate a third relationship to the terminal. The third relationship is the relationship between the identifier of the AI ​​sequence and the identifier of the sub-AI sequence in the sub-sequence set, so that the terminal can determine the AI ​​sequence to be sent based on multiple sets of sub-sequences and the third relationship.

[0272] Figure 6 This is a schematic flowchart illustrating a communication method provided in yet another embodiment of this application. Figure 6 As shown, the method may include S601 to S604.

[0273] S601, the network device sends first information, which indicates multiple sub-sequence sets, and the multiple sub-sequence sets and the AI ​​sequence set satisfy a second relationship. Correspondingly, the terminal receives the first information.

[0274] In this embodiment, the specific implementation of S601 can be referred to S501, and will not be repeated here.

[0275] S602, the network device sends fifth information, which indicates a third relationship. The third relationship is the relationship between the identifier of the AI ​​sequence and multiple sub-identifiers. The multiple sub-identifiers correspond to multiple sets of sub-sequences, and each sub-identifier is used to determine a sub-AI sequence in the corresponding set of sub-sequences. Correspondingly, the terminal receives the fifth information.

[0276] In this embodiment, the sub-identifier can be understood as the identifier of the sub-AI sequence in the sub-sequence set.

[0277] In this embodiment, based on the hierarchical mapping of the AI ​​sequence set into multiple sub-sequence sets, the identifier of the AI ​​sequence can be hierarchically mapped into multiple sub-identifiers, with each sub-identifier corresponding to a sub-AI sequence in a sub-sequence set.

[0278] As an example, the identifier of an AI sequence can be hierarchically mapped to multiple sub-identifiers using quotient-remainder mapping. For instance, suppose T is 2, subsequence set 1 contains M1 sub-AI sequences, and subsequence set 2 contains M2 sub-AI sequences. Then, the sub-identifier 1 of sub-AI sequence 1 in subsequence set 1 can satisfy the following relationship with the identifier of the AI ​​sequence: f1(x) = x / / M1, where / / represents integer division. The sub-identifier 2 of sub-AI sequence 2 in subsequence set 2 can satisfy the following relationship with the identifier of the AI ​​sequence: f2(x) = x mod M1, where mod represents modulo. Here, f1(x) is sub-identifier 1, x is the identifier of the AI ​​sequence, and f2(x) is sub-identifier 2.

[0279] Optionally, f1(x) and f2(x) can be predefined by a standard, and the parameters in f1(x) and f2(x) (such as x, M1) can be indicated by the network device to improve the flexibility of identity mapping.

[0280] Optionally, f1(x) and f2(x) can be an AI model. This AI model can reside in the terminal, or in other words, the terminal can invoke the AI ​​model.

[0281] Optionally, the fifth piece of information can be carried through RRC messages, SIB messages, or NAS messages, or the third relationship can be carried through RRC messages, SIB messages, or NAS messages; no restrictions are imposed here.

[0282] S603, the network device sends the sixth information, which is used to indicate the identifier of the first AI sequence. Correspondingly, the terminal receives the sixth information.

[0283] In this embodiment, the specific implementation of S603 can be referred to S503, and will not be repeated here.

[0284] S604, the terminal sends a first AI sequence, which is generated based on the first information and is included in the AI ​​sequence set. Correspondingly, the network device receives the first AI sequence.

[0285] In this embodiment, after determining the identifier of the first AI sequence, the terminal can map the identifier of the first AI sequence to multiple sub-identifiers based on a third relation, and determine multiple sub-AI sequences from multiple sub-sequence sets based on the multiple sub-identifiers, thereby determining the first AI sequence. The first AI sequence is the product of the multiple sub-AI sequences.

[0286] Figure 7 This is a schematic illustration of determining a first AI sequence according to one embodiment of this application. Figure 7 As shown, the identifier of the first AI sequence is x. The terminal can map the identifier x of the first AI sequence to two sub-identifiers based on a third relation, such as... Figure 7 The terminal can determine the sub-AI sequence Z1 from the sub-sequence set Y1 based on sub-identifier f1(x), and the sub-AI sequence Z2 from the sub-sequence set Y based on sub-identifier f2(x). After determining the sub-AI sequences Z1 and Z2, the first AI sequence can be determined. For example, the first AI sequence and the sub-AI sequences Z1 and Z2 satisfy the relationship: Z = Z1 * Z2. Z is the first AI sequence. The terminal can send the first AI sequence to the network device, and the network device can detect the first AI sequence based on the AI ​​detection model.

[0287] In this embodiment, the network device can indicate multiple sets of sub-sequences and third relations after hierarchical mapping to the terminal. The terminal can determine multiple sub-identifiers based on the identifier of the AI ​​sequence to be sent and the third relation, and determine multiple sub-AI sequences based on the multiple sub-identifiers and the multiple sets of sub-sequences, thereby determining the AI ​​sequence to be sent. In this embodiment, the amount of data in the multiple sets of sub-sequences sent by the network device is less than the amount of data in the AI ​​sequence set, thereby reducing air interface overhead to a certain extent.

[0288] It should be noted that the information sent by the network device in this application embodiment can be sent through the same message or through different messages, and this application does not impose any restrictions on this.

[0289] Figure 8This is a schematic diagram of the structure of a communication device provided in one embodiment of this application. Figure 8 The illustrated apparatus 800 can be used to implement the various steps / operations performed by a terminal or network device in the aforementioned method embodiments. For example... Figure 8 As shown, the device 800 may include a receiving module 810 and a transmitting module 820.

[0290] In one possible implementation, device 800 can achieve Figure 2 , Figure 3 , Figure 4 , Figure 5 and Figure 6 The methods shown are the individual steps or operations performed by the terminal.

[0291] As an example, device 800 is used to implement Figure 2 In the method shown, during the various steps / operations performed by the terminal, the receiving module 810 can be used to implement the operations performed by the terminal in S201; the sending module 820 can be used to implement the operations performed by the terminal in S202.

[0292] As an example, device 800 is used to implement Figure 3 In the method shown, during the various steps / operations performed by the terminal, the receiving module 810 can be used to implement the operations performed by the terminal in S302, S303 and S304; the sending module 820 can be used to implement the operations performed by the terminal in S301 and S305.

[0293] As an example, device 800 is used to implement Figure 4 In the method shown, during the various steps / operations performed by the terminal, the receiving module 810 can be used to implement the operations performed by the terminal in S402, S403 and S404; the sending module 820 can be used to implement the operations performed by the terminal in S401 and S405.

[0294] As an example, device 800 is used to implement Figure 5 In the method shown, during the various steps / operations performed by the terminal, the receiving module 810 can be used to implement the operations performed by the terminal in S501, S502 and S503; the sending module 820 can be used to implement the operations performed by the terminal in S504.

[0295] As an example, device 800 is used to implement Figure 6 In the method shown, during the various steps / operations performed by the terminal, the receiving module 810 can be used to implement the operations performed by the terminal in S601, S602 and S603; the sending module 820 can be used to implement the operations performed by the terminal in S604.

[0296] In this implementation, the device 800 may further include a processing module 830. The processing module 830 is used to control the device 800 to perform any of the following operations within a first time period: sending a first sequence, or not sending a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0297] In one possible implementation, device 800 can achieve Figure 2 , Figure 3 , Figure 4 , Figure 5 and Figure 6 The steps or operations performed by the network device in the method shown.

[0298] As an example, device 800 is used to implement Figure 2 In the method shown, during each step / operation performed by the network device, the receiving module 810 can be used to implement the operation performed by the network device in S202; the sending module 820 can be used to implement the operation performed by the network device in S201.

[0299] As an example, device 800 is used to implement Figure 3 In the method shown, during the various steps / operations performed by the network device, the receiving module 810 can be used to implement the operations performed by the network device in S301 and S305; the sending module 820 can be used to implement the operations performed by the network device in S302, S303 and S304.

[0300] As an example, device 800 is used to implement Figure 4 In the method shown, during the various steps / operations performed by the network device, the receiving module 810 can be used to implement the operations performed by the network device in S401 and S405; the sending module 820 can be used to implement the operations performed by the network device in S402, S403 and S404.

[0301] As an example, device 800 is used to implement Figure 5 In the method shown, during the various steps / operations performed by the network device, the receiving module 810 can be used to implement the operations performed by the network device in S504; the sending module 820 can be used to implement the operations performed by the network device in S501, S502 and S503.

[0302] As an example, device 800 is used to implement Figure 6 In the method shown, during the various steps / operations performed by the network device, the receiving module 810 can be used to implement the operations performed by the network device in S604; the sending module 820 can be used to implement the operations performed by the network device in S601, S602 and S603.

[0303] In this implementation, the device 800 may further include a processing module 830. The processing module 830 is used to control the device 800 to perform any of the following operations within a first time period: receiving a first sequence, or not receiving a sequence; wherein the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

[0304] Figure 9 This is a schematic diagram of the structure of a communication device provided in another embodiment of this application. Figure 9 The apparatus 900 shown can be used to implement the method executed by a terminal or network device in any of the foregoing embodiments.

[0305] like Figure 9 As shown, the device 900 of this embodiment includes a memory 910, a processor 920, a communication interface 930, and a bus 940. The memory 910, processor 920, and communication interface 930 are interconnected via the bus 940.

[0306] The memory 910 can be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 910 can store programs, and when the program stored in the memory 910 is executed by the processor 920, the processor 920 performs the execution... Figure 2 , Figure 3 , Figure 4 , Figure 5 or Figure 6 The steps in the method shown are performed by the terminal or network device.

[0307] The processor 920 may be a general-purpose central processing unit (CPU), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, used to execute relevant programs to implement the communication method shown in the embodiments of this application.

[0308] The processor 920 can also be an integrated circuit chip with signal processing capabilities. In implementation, each step of the communication method shown in the embodiments of this application can be completed by the integrated logic circuitry in the processor 920 or by software instructions.

[0309] The processor 920 described above can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor.

[0310] The steps of the method disclosed in the embodiments of this application can be directly manifested as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 910. The processor 920 reads the information in memory 910 and, in conjunction with its hardware, completes the functions required by the units included in the communication device of this application. For example, it can execute... Figure 2 , Figure 3 , Figure 4 , Figure 5 or Figure 6 The steps / functions performed by the terminal or network device in the method shown.

[0311] Alternatively, the memory 910 and the processor 920 can be integrated together.

[0312] The communication interface 930 can use, but is not limited to, transceivers to enable communication between the device 900 and other devices or apparatuses.

[0313] Bus 940 may include a pathway for transmitting information between various components of device 900 (e.g., memory 910, processor 920, communication interface 930).

[0314] Some embodiments of this application also provide a computer program product that, when run on a processor, can implement the methods shown in the foregoing embodiments. Some embodiments of this application also provide a computer-readable storage medium containing computer instructions that, when run on a processor, can implement the methods shown in the foregoing embodiments.

[0315] It should be noted that the modules or components shown in the above embodiments can be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), etc. Furthermore, when a module is implemented by a processing element calling program code, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processors capable of calling program code, such as a controller. Additionally, these modules can be integrated together to implement a system-on-a-chip (SOC).

[0316] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, software modules, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0317] The term "multiple" in this document refers to two or more. The term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Furthermore, the character " / " in this document generally indicates an "or" relationship between the preceding and following related objects; in formulas, " / " indicates a "division" relationship. Additionally, it should be understood that in the description of this application, words such as "first" and "second" are used only for descriptive purposes and should not be construed as indicating or implying relative importance or order.

[0318] It is understood that the terms "exemplary" or "for example" used herein are intended to mean as an example, illustration, or illustration. Any embodiment or design described herein as "exemplary" or "for example" should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0319] It is understood that the various numerical designations used in the embodiments of this application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of this application.

[0320] It is understood that, in the embodiments of this application, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

Claims

1. A communication method, characterized in that, The method includes: Receive first information, the first information being used to indicate a vector, the vector satisfying a first relationship with the identifier of the artificial intelligence (AI) sequence, the vector being used as input to the AI ​​model, and the AI ​​model being used to generate the AI ​​sequence; Send a first AI sequence, which is generated based on the first information.

2. The method according to claim 1, characterized in that, The first information is used to indicate the vector, including: The first information is used to indicate the first relationship.

3. The method according to claim 2, characterized in that, The first information includes at least one of the following: a first formula that the identifier of the AI ​​sequence and the vector satisfy, the parameters in the first formula, the length of the vector, or the processing method of the vector.

4. The method according to claim 1, characterized in that, The first information is used to indicate the vector, including: The first information is used to indicate the vector set in which the vector is located. Multiple vectors in the vector set correspond one-to-one with multiple AI sequences. Each vector in the multiple vectors satisfies the first relationship with the identifier of the corresponding AI sequence. The multiple AI sequences belong to a subset of the sequence set in which the AI ​​sequence is located.

5. The method according to claim 4, characterized in that, The first information includes at least one of the following: the plurality of vectors, or the processing method of each of the plurality of vectors.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: Receive second information, which is used to instruct the AI ​​model.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: Send a third message, which indicates at least one of the following: supported AI model information, supported first relation, or the duration required to generate the AI ​​sequence.

8. A communication method, characterized in that, The method includes: Receive first information, which indicates a plurality of sub-sequence sets, and the plurality of sub-sequence sets and the artificial intelligence (AI) sequence set satisfy a second relationship; Send a first AI sequence, which is generated based on the first information, and the first AI sequence is included in the AI ​​sequence set; Wherein, the product of the number of sub-AI sequences included in each of the plurality of sub-sequence sets is equal to the number of AI sequences included in the AI ​​sequence set, and the sum of the number of sub-AI sequences included in each of the sub-sequence sets is less than the number of AI sequences included in the AI ​​sequence set.

9. The method according to claim 8, characterized in that, The method further includes: Receive fourth information, which is used to indicate the second relationship.

10. The method according to claim 8 or 9, characterized in that, The method further includes: The fifth information is received, which is used to indicate a third relationship. The third relationship is the relationship between the identifier of the AI ​​sequence and multiple sub-identifiers. The multiple sub-identifiers correspond to the set of multiple sub-sequences. Each sub-identifier is used to determine a sub-AI sequence in the corresponding set of sub-sequences. The AI ​​sequence is generated based on the multiple sub-AI sequences determined by the multiple sub-identifiers.

11. The method according to any one of claims 1 to 10, characterized in that, The method further includes: Receive a sixth message, which is used to indicate the identifier of the first AI sequence.

12. The method according to any one of claims 1 to 11, characterized in that, The method further includes: During the first time period, perform either of the following operations: send the first sequence, or do not send the sequence; Wherein, the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

13. A communication method, characterized in that, The method includes: Send first information, which is used to indicate a vector, and the vector satisfies a first relationship with the identifier of the artificial intelligence (AI) sequence. The vector is used as input to the AI ​​model, and the AI ​​model is used to generate the AI ​​sequence. Receive a first AI sequence, which is generated based on the first information.

14. The method according to claim 13, characterized in that, The first information is used to indicate the vector, including: The first information is used to indicate the first relationship.

15. The method according to claim 14, characterized in that, The first information includes at least one of the following: a first formula that the identifier of the AI ​​sequence and the vector satisfy, the parameters in the first formula, the length of the vector, or the processing method of the vector.

16. The method according to claim 13, characterized in that, The first information is used to indicate the vector, including: The first information is used to indicate the vector set in which the vector is located. Multiple vectors in the vector set correspond one-to-one with multiple AI sequences. Each vector in the multiple vectors satisfies the first relationship with the identifier of the corresponding AI sequence. The multiple AI sequences belong to a subset of the sequence set in which the AI ​​sequence is located.

17. The method according to claim 16, characterized in that, The first information includes at least one of the following: the plurality of vectors, or the processing method of each of the plurality of vectors.

18. The method according to any one of claims 13 to 17, characterized in that, The method further includes: Send a second message, which is used to instruct the AI ​​model.

19. The method according to any one of claims 13 to 18, characterized in that, The method further includes: Receive third information, which indicates at least one of the following: supported AI model information, supported first relation, or the time required to generate the AI ​​sequence.

20. A communication method, characterized in that, The method includes: Send a first message, which indicates a plurality of sub-sequence sets, and the plurality of sub-sequence sets and the artificial intelligence (AI) sequence set satisfy a second relationship; Receive a first AI sequence, which is generated based on the first information, and the first AI sequence is included in the AI ​​sequence set; Wherein, the product of the number of sub-AI sequences contained in each of the plurality of sub-sequence sets is equal to the number of AI sequences contained in the sequence set, and the sum of the number of sub-AI sequences contained in each of the sub-sequence sets is less than the number of AI sequences contained in the AI ​​sequence set.

21. The method according to claim 20, characterized in that, The method further includes: Send a fourth message, which is used to indicate the second relationship.

22. The method according to claim 20 or 21, characterized in that, The method further includes: Send a fifth message, which is used to indicate a third relationship, which is the relationship between the identifier of the AI ​​sequence and multiple sub-identifiers. The multiple sub-identifiers correspond to the set of multiple sub-sequences. Each sub-identifier is used to determine a sub-AI sequence in the corresponding set of sub-sequences. The AI ​​sequence is generated based on the multiple sub-AI sequences determined by the multiple sub-identifiers.

23. The method according to any one of claims 13 to 22, characterized in that, The method further includes: A sixth message is sent, which is used to indicate the identifier of the first AI sequence.

24. The method according to any one of claims 13 to 23, characterized in that, The method further includes: During the first time period, perform one of the following operations: receive the first sequence, or do not receive the sequence; Wherein, the start time of the first time period is the start time of generating the AI ​​sequence, and the duration of the first time period is the duration required to generate the AI ​​sequence.

25. A communication device, characterized in that, It includes various functional modules for implementing the method as claimed in any one of claims 1 to 12 or any one of claims 13 to 24.

26. A communication device, characterized in that, include: A processor coupled to a memory for storing a computer program, which, when invoked by the processor, causes the apparatus to perform the method as claimed in any one of claims 1 to 12 or any one of claims 13 to 24.

27. A chip, characterized in that, include: A processor for calling and executing a computer program in memory, causing the chip to perform the method as claimed in any one of claims 1 to 12 or any one of claims 13 to 24.

28. A computer-readable medium, characterized in that, The computer-readable medium stores instructions that, when executed, cause the method as claimed in any one of claims 1 to 12 or any one of claims 13 to 24 to be implemented.

29. A computer program product, characterized in that, It includes computer program code that, when run on a computer, causes the method as claimed in any one of claims 1 to 12 or any one of claims 13 to 24 to be implemented.

Images