Methods, systems, and devices for wireless communications are described. A user equipment (UE) may receive, from a base station, control signaling identifying a configuration for reporting channel metrics for transport blocks that each include a plurality of code block group. The UE may receive a first transport block that includes a first plurality of code block groups. The UE may transmit a report indicating a channel metric for one or more code block groups of the first plurality of code block groups of the first transport block in accordance with the configuration for reporting channel metrics. In some cases, the UE may report a differential channel metric, based in part on a reference metric. Additionally or alternatively, the UE may report compressed feedback for a group CBGs.

This disclosure provides methods, devices and systems for radio frequency (RF) sensing in wireless communication systems. In some implementations, a transmitter device transmits sounding sequences configured for channel estimation over a wireless channel to a receiver device. The transmitter device also transmits or receives non-sounding frames associated with a channel report of the receiver device. The transmitter device transmits a frame soliciting the channel report from the receiver device. The transmitter device receives the channel report, which may include channel state information (CSI) of the wireless channel responsive to at least the sounding sequences. Some types of channel reports may take longer to generate than other types of channel reports. Transmitting or receiving the non-sounding frames during the time period may prevent other devices from accessing the wireless channel when the receiver device needs additional time to generate a certain type of channel report.

Certain aspects of the present disclosure provide techniques for qualifying machine learning model-based channel state information (CSI) predictions. An example method generally includes receiving, from a network entity, a channel state information (CSI) prediction model for quantized CSI, calculating CSI based on downlink reference signal measurements, generating a quantized CSI difference value based a quantization of a difference between the calculated CSI and CSI predicted based on a CSI prediction model, and reporting, to the network entity, the calculated CSI and the quantized CSI difference value.

Methods, systems, and devices for wireless communications are described. A user equipment (UE) may perform a measurement operation to attain multiple measurements to report to a base station. The measurements may correspond to a first number of bits if reported. The UE may compress the measurements using an encoder neural network (NN) to obtain an encoder output indicating the measurements. This encoder output may include a second number of bits that is less than the first number of bits. The UE may report the encoder output to the base station in this compressed form. At the base station, the encoder output may be decompressed according to a decoder NN. Once the base station decompresses the encoder output, the UE and base station may communicate according to the measurements determined from the decompression. In some cases, the base station may perform load redistribution based on the measurements.

Methods, systems, and devices for wireless communications are described. Feedback information associated with decoding downlink information may be reported in a first stage and a second stage. A size of the second stage may be restricted by an upper limit associated with an available set of uplink resources for transmitting the feedback information. Limiting the size of the second stage may include reducing a size of second information generated for the second stage, yielding third information for the second stage. The third information may be transmitted with first information generated for the first stage. Limiting the size of the second stage may also include generating second information having a reduced size based on determining that the upper limit may be exceeded. The reduced-size second information may be transmitted with the first information.

In accordance with an example embodiment of the present invention, a method, apparatus and software program product for partitioning channel state information into groups of different priority levels, wherein said channel state information comprises non zero coefficients and said partitioning comprises using at least one permutation function to prioritize said non-zero coefficients substantially based on an amplitude distribution of said non zero coefficients to determine mapping of said non zero coefficients to said groups, and further for omitting zero or more non-zero coefficients in accordance with said amplitude distribution based prioritizing from said groups, and for transmitting remaining non zero coefficients in said groups.

Methods, systems, and devices for wireless communication are described. Generally, the described techniques provide for efficiently reporting channel state feedback for a set of subbands. As an example, a user equipment (UE) may receive a set of reference signals on the set of subbands and determine a respective channel quality indicator (CQI) index for each of the set of subbands based on the reference signals. The UE may then transmit indications of CQI indices for different subbands of the set using variable length indications, such as indications that include different numbers of bits. Since, in some examples, a small length or number of bits may be allocated for reporting CQI indices with a high probability of being reported, overhead in a wireless communications system may be reduced.

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a user equipment (UE) may receive a single downlink control information (DCI) message scheduling a simultaneous downlink transmission and uplink transmission in an unpaired band. In some aspects, the single DCI message may include a modulation and coding scheme (MCS) field having one or more bits that indicate a first MCS for the downlink transmission. The UE may determine a second MCS for the uplink transmission based at least in part on an offset from the first MCS for the downlink transmission. Numerous other aspects are provided.

A neural network adjustment method and apparatus. The neural network adjustment method relates to a first neural network and a second neural network. The method includes: whether a currently used neural network needs to be updated is first determined; and if the currently used neural network needs to be updated, a first device may autonomously complete the update or indicate a second device to complete the update. In other words, a more appropriate neural network may be obtained by updating a neural network in an AI model, to improve precision of information transmission based on the AI model.

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a user equipment (UE) may transmit, to a base station, channel state information (CSI) feedback indicating a modulation and coding scheme (MCS) associated with a spectral efficiency below a minimum MCS index in a spectral efficiency table. The UE may receive, from the base station, a downlink transmission including a number of resource blocks (RBs), wherein the number of RBs is based at least in part on the spectral efficiency associated with the MCS indicated in the CSI feedback. Numerous other aspects are described.