Abstract: This application provides a communication method and apparatus. The method includes: determining a first downlink signal and a second downlink signal, where the first downlink signal is a downlink signal determined based on an output of a first prediction model, an input of the first prediction model is determined based on Y pieces of downlink measurement information in X pieces of downlink measurement information, and the second downlink signal is determined based on the X pieces of downlink measurement information; and indicating the second downlink signal and the Y pieces of downlink measurement information to an access network device if the first downlink signal and the second downlink signal meet a first condition. According to the method, the first prediction model can be updated based on the foregoing information, and performance of the prediction model can be continuously improved.

Abstract: This application provides a calibration method and apparatus. The method includes: determining first channel state information and second channel state information, where the first channel state information is obtained through prediction via a first channel prediction model, and the second channel state information is determined based on channel estimation on a downlink signal from an access network device; and sending first information to the access network device if a difference metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold, where the first information indicates the second channel state information.

Abstract: A communication method and apparatus are provided. The method includes: obtaining N pieces of training data, where for one of the N pieces of training data, the training data corresponds to one piece of mark information, and the mark information indicates an attribute of the training data; and sending indication information to a terminal device, where the indication information indicates information about M artificial intelligence models, the artificial intelligence model is trained based on X pieces of training data in the N pieces of training data, and the X pieces of training data are determined based on the mark information. When the N pieces of training data are obtained, the N pieces of training data may be used for constructing training data corresponding to different mark information, to train the M artificial intelligence models or more artificial intelligence models, so that scenario-based artificial intelligence model training can be implemented.

Abstract: A random access method, a device, and a system are provided, related to the field of communications technologies.

Abstract: This application provides an artificial intelligence model download method, an apparatus, and a system. The method includes: A terminal device receives model feedback information from an access network device, where the model feedback information indicates a download mode that is of a first artificial intelligence model and that is configured for the terminal device, and the download mode that is of the first artificial intelligence model and that is configured for the terminal device is a download mode supported by the terminal device. The terminal device may download the first artificial intelligence model in the download mode indicated by the model feedback information. The terminal device can download the first artificial intelligence model supported by the terminal device, to avoid a case in which a downloaded artificial intelligence model cannot be used.

Abstract: A communication method and apparatus. A terminal receives first information from a network device. The terminal determines N pieces of training data based on the first information, and N is an integer. The terminal performs model training based on the N pieces of training data, to obtain a first AI model. The network device configures the first information used to determine the N pieces of training data for the terminal. The terminal performs model training based on the N pieces of training data autonomously. Separately configuring an AI model for the terminal is not necessary and air interface overheads are reduced.

Abstract: A model test method and an apparatus. A terminal receives first information from a network device. The terminal determines at least one test set based on the first information. The terminal tests an AI model based on the at least one test set, to obtain a test result. The terminal notifies the network device of the test result of the AI model.

Abstract: A communication method includes determining a first submodel and a second submodel. The first submodel and the second submodel are used in a matching manner. The communication method also includes sending first information. The first information indicates one or more of input data of the first submodel or output data of the first submodel.

Abstract: An access network device receives M pieces of first information of M local models from M terminal devices, where the M terminal devices one-to-one correspond to the M local models, the M local models one-to-one correspond to the M pieces of first information, and the M pieces of first information of the M local models are information received in first duration. The access network device updates a global model based on the M pieces of first information of the M local models and sends first information of the global model to K terminal devices. In the method, the first duration is set, to ensure that an update progress of local information matches an update progress of the global model and improve accuracy and performance of the global model obtained by the access network device.

Abstract: This application relates to a channel information feedback method, a channel information recovery method and an apparatus. In an example method, a terminal device receives a first reference signal from a network device. The terminal device obtains a first channel state matrix based on the first reference signal. The terminal device performs sampling on the first channel state matrix based on a first sparse pattern, to obtain a sparse channel state matrix, where the first sparse pattern is configured by the network device, and the first sparse pattern is used for performing sampling on at least one dimension of the first channel state matrix. The terminal device sends first channel information to the network device, where the first channel information indicates the sparse channel state matrix.

Abstract: In a method for determining a resource of an asynchronous physical uplink shared channel PUSCH, user equipment (UE) receives time domain resource configuration information and frequency domain resource configuration information from a network device. The UE sets, based on the time domain resource configuration information, a time domain position of a first hop of a first PUSCH and a time domain position of a second hop of the first PUSCH to be inconsecutive in the time domain. The UE then determines a frequency domain position of the first hop of the first PUSCH and a frequency domain position of the second hop of the first PUSCH based on the frequency domain resource configuration information.

Abstract: The application provides a data sending method and a communication apparatus to avoid signaling overheads caused by scheduling by a network device, and a terminal device can flexibly select a quantity of repeated transmissions based on an actual situation, so that data transmission performance is improved. The network device sends configuration information of a receive power to the terminal device. After receiving the configuration information of the receive power, the terminal device determines a quantity of repeated transmissions of target data based on the configuration information of the receive power. After determining the quantity of repeated transmissions of the target data, the terminal device sends the target data to the network device based on the quantity of repeated transmissions of the target data.

Abstract: An uplink control information transmission method and apparatus. The method includes: determining, by a terminal device, a preamble that needs to be sent; determining, by the terminal device based on the preamble, a first resource set that is in a physical uplink shared channel resource and that is used to send uplink control information, where the PUSCH resource includes a plurality of resource sets; and sending, by the terminal device, the uplink control information and uplink data on the PUSCH resource, where the uplink control information is mapped to the first resource set, the uplink data is mapped to a resource that is other than the plurality of resource sets and that is in the PUSCH resource, and no information or uplink data is mapped to a resource set that is other than the first resource set and that is in the plurality of resource sets.

Abstract: This application relates to the field of communications technologies, and discloses a DMRS port determining method and a communications apparatus, to resolve a problem that collisions occur on DMRS ports used by a large quantity of terminal devices to send uplink data and a random access effect is affected during 2-step random access. The method includes: receiving, by a terminal device, configuration information sent by a network device, where the configuration information includes DMRS configuration information and indication information of a DMRS port set, the DMRS configuration information is used to configure a plurality of DMRS ports, and the indication information of the DMRS port set is used to indicate one or more available DMRS ports in the plurality of DMRS ports; and determining, by the terminal device, a target DMRS port in the DMRS port set.

Abstract: Embodiments of this application provide a method for configuring a reference signal, to increase a rate of data transmission between an access network device and a terminal device. The method includes: An access network device determines a first reference signal pool, where the first reference signal pool includes one or more reference signals. The access network device sends first information to a first terminal device, where the first information indicates a first reference signal allocated to the first terminal device, the first reference signal is included in the first reference signal pool, and the first reference signal pool is obtained by the access network device by training a first neural network, or the first reference signal pool is obtained by another network-side node by training a first neural network and sent to the access network device.

Abstract: An uplink transmission methods and apparatuses to reduce a probability that a collision between a user equipment (UE) and another UE occurs when the UE sends uplink data, thereby improving network performance are described. A plurality of different preamble signals are sent to a network device by a user equipment (UE). A feedback signal of the network device in response to the plurality of preamble signals is received by the UE. The feedback signal in response to each preamble signal carries a timing advance (TA) corresponding to the preamble signal. A valid TA is determined by the UE from the plurality of received TAs. A parameter used for uplink transmission is determined by the UE based on a feedback signal carrying the valid TA or a preamble signal corresponding to the valid TA. Uplink data are sent according to the parameter by the UE.

Abstract: A communication method includes receiving, by a first terminal device, first information sent by a second terminal device. The first information includes information indicating a second beam selected by the second terminal device. The communication method also includes sending, by the first terminal device, (1) identification information of the second terminal device, or (2) identification information and the first information, to a network device. The first terminal device is a cluster head node, and the second terminal device is a member node of a cluster to which the first terminal device belongs.

Abstract: An uplink data transmission method and an apparatus include or relate to a terminal device which determines, based on indication information of a network device, a time-frequency resource corresponding to a random access resource, or a part of a time frequency resource in the random access resource, to directly transmit uplink data. Configuration information of a time-frequency resource for random access may be reused for configuration information of the time-frequency resource, so that signaling overheads are reduced and a resource utilization is improved.

Abstract: Embodiments of this application provide an information transmission method and a device. The method includes receiving at least one precoding configuration from an access network device, starting a timer, and then sending uplink information to the access network device on a configured grant-free time-frequency resource in valid duration of the timer by using the at least one precoding configuration.

Abstract: This application provides an asynchronous uplink transmission method, a device, and a storage medium. The method includes: receiving, by a UE, configuration information that is of a first physical uplink shared channel (PUSCH) and that is sent by a network device, where the configuration information includes a length of a cyclic prefix (CP) of the first PUSCH, and the length of the CP of the first PUSCH is less than a length of a CP of a first random access preamble; determining, by the UE based on the length of the CP of the first PUSCH, a difference between a largest stored timing advance (TA) and a smallest stored TA, and a largest delay spread, whether to send the first PUSCH; and responsive to determining to send the first PUSCH, sending, by the UE, the first PUSCH based on the configuration information of the first PUSCH.